Skip to main content
Home Forums General Silverlight New Features in Silverlight 3 Silverlight 4 Wishlist
310 replies. Latest Post by kiener on November 6, 2009.
(3)
BigDubb
Member
65 points
77 Posts
06-17-2009 11:08 AM |
Seeing as SL3 is just around the corner might as well start the discussion around what the community would like to see in SL4.
1. Ability to define chrome on OOB applications
2. Deep linking ability/navigation in OOB applictions.
3. Printing Capabilities.
ksleung
Contributor
5198 points
993 Posts
06-17-2009 2:22 PM |
Normally I don't like laundry list but since the SL team tracked the SL3 wishlist let me name a few things I wanted (and what I want is quite realistic :D)
1. Printing support
2. Better shader support... GPU-rendering of pixel shader effects, better HLSL support (move at least to HLSL 2.0b)
3. Fix WriteableBitmap (in addition to the security issues) -- allow modification in background thread, better API (e.g. array copy), improve rendering performance
4. Need better LocalMessage APIs -- discovery, stream-based or array-based message (rather than just strings), multicast
5. Better File IO support -- ability to specify filename in SaveFileDialog, grant read/write access to whole directory on user approval
06-17-2009 2:28 PM |
Also the Silverlight team would be wise to reconsider the OOB model. The original model is nice in theory but in reality, it seems so crippled compared to Adobe AIR. It is true that AIR is perhaps a big security liability -- it is not very different from downloading and installing full-privilege software -- users don't seem to mind at least judging by the success of many Twitter clients. I like security done right, and Microsoft has already learned the lessons, but at some point we need to offer a different, more AIR like, OOB model for developers to choose. Of course, we must message this correctly, like, telling in your face that OOB-full-privilege is the same as downloading and installing an executable.
TomGiam
706 points
204 Posts
06-17-2009 3:59 PM |
I agree, the 64 instruction slot limit in pixel shader 2.0 is very limiting (In 2.0b it's 512), especially since in SL3 you can only apply one effect per element (i.e. no Effects group) so to work around that you need to combine multiple effects into one.
Tom
06-17-2009 4:40 PM |
1. By far my number one feature need is PRINTING.
2. After that it's the ability to export any UIElement as a bitmap to both the server and client
3. Reading and writing from/to the local machine is also very useful, including reading all files (mainly images) from a directory.
wackyphill
37 points
115 Posts
06-17-2009 10:14 PM |
Better mouse support. (Scroll wheel, doubleClick, right-CLick)
Printing
Being able to host HTML content within the SL app would be nice.
tfmillet
10 points
5 Posts
06-19-2009 10:06 AM |
1. HTML Viewer Control
2. Right-click Context Menus
3. Printing
Zachary D.
76 points
187 Posts
06-19-2009 10:26 AM |
1) Printing
2) Host html inside of silverlight control
3) Right mouse button support
Consider giving the business developers some love.
tanmoy.r
3580 points
708 Posts
06-19-2009 10:53 AM |
Thanks for starting this thread early. :) Now nobody can say community didn't startyearly enough to get them implemented.
My lists
1. Mouse wheel support
2. Full 3D support as in WPF with GPU acceleration.
3. Print, Microphone and WebCam support
4. Linux Support. (I dont want to see SL 2 is supported at linux where SL 4 is the latest version out there)
5. Some more Killer features like DeepZoom, PixelEffects etc.
6. Support for other bindings in WCF.
7. Complete source code for controls.
8. More integrated support for Unit Testing.
06-19-2009 5:41 PM |
Also having the StringFormat option for Binding like WPF got in 3.5SP1 would be very handy so Value converters are not always needed.
firsname...
2 points
1 Posts
06-20-2009 7:57 AM |
MUST HAVE RTL - BiDI - RIGHT TO LEFT!
I can't understand how can it be that rendering text right to left stead' of left to right is so incredibly complex to develop
that it's worth simply out front discarding 280 mil Arabic speakers, Israel - Hebrew, Pakistan - Urdu, Farsi...
I just don't get it.
I'm also beginning to be somewhat insulted. I just can't see how this can somehow be a business driven decision.
Being a Microsoft (PC for all) developer i am disappointed by this attitude...
Version 4? :-(
alok572
334 points
63 Posts
06-22-2009 12:23 AM |
There are many small things that are previously not present in silverlight like, no flexible table control that will help to merge columns or rows.
No support for mobile applications..... etc
The team should consider these small issues.
davesmits
624 points
198 Posts
06-22-2009 1:47 PM |
- Webcam
- Audio recording and streaming
and better mouse support (mousewheel i.e.) are the things i really missing
06-22-2009 6:18 PM |
Being able to select, right-click and copy text from a textblock should be there as well.
Flash does this fine. It would also be nice to have gif support.
GeminiYe...
6 points
12 Posts
06-25-2009 5:41 AM |
MouseRightButton
Dropdown Menu
Alessand...
62 points
52 Posts
06-25-2009 6:13 AM |
Line, Ellipse, Rectangle, .... commands on WriteableBitmap.
andulvar
143 points
101 Posts
06-25-2009 11:05 AM |
Number 1 for me is still to fix the direct TCP/IP socket capabilities to match, and to be compatible with, Adobe Flash:
- Use the same policy file port number (port 843)
- Use the same policy file format
- Allow socket connections on any port
- Allow a server to emit its policy file on the same socket as the data connection, skipping the dedicated policy port altogether
Current capability is too restrictive to be useful for non-toy applications.
Psychlis...
6035 points
973 Posts
06-28-2009 1:22 PM |
Hi All
This list is great. However, please understand that the Silverlight team works primarily off of scenarios rather than bullet points. Concrete scenarios help them target and justify exactly what they need to do.
So, instead of saying "Printing" say how you would use it. That will help clear up whether you need a report-writer, print-screen capability, ability to write strings out to a printer, print all the contents of a grid in a paginated way etc.
Same thing with the other feature requests.
Keep them coming, just please put some context around it :)
Pete
06-28-2009 7:01 PM |
Psychlist1972:Hi All This list is great. However, please understand that the Silverlight team works primarily off of scenarios rather than bullet points. Concrete scenarios help them target and justify exactly what they need to do. So, instead of saying "Printing" say how you would use it. That will help clear up whether you need a report-writer, print-screen capability, ability to write strings out to a printer, print all the contents of a grid in a paginated way etc. Same thing with the other feature requests. Keep them coming, just please put some context around it :) Pete
Pete,
Good to hear that the Silverlight team is listening. I agree that a more constructive way to provide input is warranteed. I can't speak for the whole development community -- individual who champions for their own need should speak up -- but the following is the list that I have (excluding printing, which I hope others with stronger opinion would speak to that):
Pixel shader is a great addition to Silverlight 3 beta, but it is *really* limiting. The killer is that only the very primitive version HLSL 2.0, is supported. Other than the most basic shader codes, usable only as "pretty printing" of images, it is way too limiting for any interesting effects. I know, 'coz I have written a few effects and I know the limits. There are quite a few interesting Adobe Pixel Bender effects that cannot be ported to Silverlight just for this reason. So given HLSL 2.0b has the same instruction set as HLSL 2.0, why not give us at least that much? I understand that perhaps the reason is really that the SL team wants to support HLSL effects GPU-native in the future. Well, can you say something about that? If not, why don't the SL team gives the user at least the equivalence of Adobe Pixel Bender effects, especially if currently it is CPU-rendered? Suppose GPU-native is in the plan, you can also do it such that HLSL 2.0 is GPU-rendered, and HLSL 2.0b is CPU-rendered. In this scenario, at least the developer is given the choice.
What are the effects useful for? For starter, interesting image editing applications... I can say more about that if the SL team is interested.
This has been talked about many times in the SL game forum. The performances of WriteableBitmap sucks. I think there are several shortfalls that can be easily addressed without changing the fundamentals of the architecture. I think others can speak to other requirements but mine are rather simple: (1) ability to work on WriteableBitmap in non-UI thread when locked. (2) provide an API to use Array.Copy() to copy bits, especially if (1) is not feasible for architecture reason.
Another important features is the Silverlight WriteableBitmap security issue. This also has been talked about to death and in fact it was mentioned by at least two SL team members that it will be fixed in SL3 RTW. However, other than that, there is no details about it.
Why important? This is obvious. For make Silverlight a decent game, graphing, imaging platform, this is a must.
There are other things that are good to have -- blend mode, direct drawing APIs -- but I think if there is such thing as the lowest possible hanging fruit, WriteableBitmap is it.
Currently, we cannot find out the names of the other LocalMessageListener. In sample code, it is usually that hardcoded listener and sender names are used. However, how about the scenario where multiple instances of SL applications are created and destroyed at different times? My app has that scenario -- I can open several instances of the apps and I want them to talk to each other... But right now I have to use try/catch block to create instance names. And instead of multicasting I'll need to try to send messages to all possible names. I think it'd be helpful to have more meaningful discovery APIs.
dumb_dumb
73 points
35 Posts
06-28-2009 9:57 PM |
1. Printing --> report writer is the most mandatory however print screen or content of the grid is also important ^^
2. Better mouse support --> native mouse wheel support,
I guess my wish list is still exactly the same with my wish list for SL 3.0 previously :P
06-29-2009 9:50 AM |
Some context for my requests:
1. HTML Viewer/Browser control. We are building an application that displays archived information that is in HTML format. Displaying the HTML is not a problem with traditional web apps or with desktop apps such as WinForms or WPF, since a WebBrowser control exists which can render the HTML. We would like to move the old ASP app to SIlverlight, but in order to do so, we need the ability to display HTML content. We have looked at 3rd party controls, but they are limited. The HTML pane controls require windowless mode, and do not work in Safari. They also prevent overlaying Child Windows on the HTML content. 3rd-party RichTextBox controls that support HTML are limited as well -- they only support HTML if it is HTML 4.0, and formatted in a specific way. Our archived content often uses the older versions HTML, which iss not HTML 4.0 compliant. In short, we need the equivalent of the WinForms and WPF WebBrowser control in Silverlight.
2. Right-click context menus -- We are trying to build an enterprise business application -- users are too accustomed to right-click options to take that away. We don't want to limit our UI design based on this limitation.
billsm
Participant
1301 points
472 Posts
06-29-2009 4:16 PM |
1. A silverlight -sql reports and printing support for it.
2 Let's face it data is the most important item out there in wpf silverlight and asp.net and getting it right means everything. The videos and training is too easy. The rest wcf application is a nice start but it really needs to expand beyond the baby level. I would like to see videos and training using beginner medium advanced and the use of wcf and sql using design patterns in various applications. example using service factory host in deploying a web site. ..
3 Get a educational poll and then act on it!!! Technology is great however their needs to be higher learning curve!!!! Please help us grow!!
Thanks
planetma...
60 points
20 Posts
06-29-2009 4:51 PM |
I suppose this is going to be fairly low down on the priority list, but surely this could make the 3.0 release?
Expose the PixelShader.SetStreamSource() method, which is available in Wpf.
Without this it is not possible to dynamically load generated shaders. I was feeling rather pleased with myself having written a custom shader compiler with Silverlight and F# only to discover I couldn't actually load the bytecode.
Andrew.
Krasshirsch
1029 points
295 Posts
06-30-2009 3:22 AM |
The overall rendering performance must be improved.
http://silverlight.net/forums/t/104774.aspx
snap254
2 Posts
06-30-2009 11:18 AM |
To re-iterate some of the other postings with some "scenarios"
Looking forward to it!
thx
pelister
14 points
13 Posts
06-30-2009 2:48 PM |
WSHttpBinding for wcf and right click support for TextBlock and TextBox controls
ernestasju
07-01-2009 6:35 AM |
Some of my suggestion might be stuped... sory for that...
1. 3d in silverlight would be very good. It would be nice to make menu similiar to http://crytek.com menu made with Flash.
2. Animated images would be nice too.
3. That's just idea, but... In addition to out of browser mode add jump out mode (make a popup window and copy silverlight app to it).
4. Silverlight windowless mode is still to slow!!! Improve!!!
5. **** Silverlight animation file format (SAFF). **** Sometimes we don't need that UI and all that "advanced features", we want to make a simple vector graphics based animation (stuped suggestion...).
6. Install of Silverlight should be not do downloaded as exe file, but installed like NORMAL activeX control!!! Users might think that that is more one crazy virus and at all activex installs easier.
7. Better compression of application. Users hate [???.? %] view.
8. If you trying make Silverlight more secure just make brik wall view control and strech it onto all other controls. Dump suggestion...
9. I wanna make my own Open/save file dialogs. My styled. Looking like i want!!!
10. Look at 1st-9th suggestions and make SL10 compactible with DirectX 10. Playing Crysis at highest setting in BROWSER would by superb!!!!
11. And more one suggestion: Think why Silverlight can't be used on Linux, ever? (==> .xuniL no elbaliava ton si krowemarF TEN. <==).
Happy improving...
Look at http://cid-5ec2e68efd6234b0.skydrive.live.com/self.aspx/SL/SlSugedo.png
bye...
07-01-2009 1:42 PM |
Ability to launch SL application directly in OOB mode.
This would essentially be like a click once application, but cross platform.
Jalfrezi
31 points
23 Posts
07-01-2009 2:45 PM |
Turn Silverlight into a (basic) audio platform.
Audio recording:
Audio playback:
russau
07-02-2009 12:50 AM |
Herthoren
85 points
78 Posts
07-02-2009 7:13 AM |
1. Enhanced low-level audio-creation support.
2. Mouse wheel support, FULL right-click support (that silverlight menu can be hacked out anyway, writing seperate code for every browser, so allowing us to actually use right-click for whatever we want (even if it's not displaying a menu) would only make our job (especially testing) a LOT-LOT easier, and would not make Silverlight a tiny bit more insecure).
3. Full-screen WITH full-keyboard support.
4. Yeah i'm gonna get mocked for this, but I can't stop requesting UDP SUPPORT, comeon, java applets support it! :o)
5. Arbitrary (ok, at least much wider) port support for socket communications.
6. Server capabalities (or peer-to-peer communication support if you find that sound so much better)
4352. Not so very important for me, but hardware acceleration for pixel shaders would be nice :P
Even if hardware-acceleration would be OS dependent (e.g. would only be available on certain versions of windows, maybe even only for certain (popular) GPU drivers...), I still think it wouldn't hurt that bad, since it's only a performance issue and systems with very slow GPU would get just as poor performance anyway, so applications would have to be prepared for systems without hardware acceleration either way.
3456344563. Microphone (and maybe even webcam) input handling would also be cool, after all Silverlight is clearly behind flash in this one.
steveli
127 points
19 Posts
07-02-2009 2:48 PM |
1. Support for WebCam, Microphone, Audio (with ability to record)
2. Printing, in other words being able to send graphics and text to print
3. Better pixel shader support, btw is it possible to build shadow pixel shader in SL3 inside using fx file?
07-02-2009 3:54 PM |
steveli:1. Support for WebCam, Microphone, Audio (with ability to record)2. Printing, in other words being able to send graphics and text to print3. Better pixel shader support, btw is it possible to build shadow pixel shader in SL3 inside using fx file?
Regarding (3), SL3 beta already supports Pixel Shader. You need to compile an fx file (HLSL 2.0 only) to ps using DirectX SDK or other means. It is pretty straight-forward to include the ps into a Silverlight project. If you are talking about dynamically generated fx/ps files, no solution exists at this point.
07-02-2009 4:12 PM |
Yes I know that Silverlight3 supports pixel shaders, however I'd like to see any sample on how you can build shadow shader using HLSL in Silveright3.
I did reflection shader, you can read more on my blog:
http://www.cellbi.com/blog.aspx
But what I can't understand is how you can build a shader,
which will know bounds of UIElement it is applied to?
E.g. if you apply DropShadow pixel shader to text block then shadow is nicely added for each char in the text.
I'd like to be able to build glow shader using the same approach. Do you know of any way to do this?
07-02-2009 5:18 PM |
Steveli,
To implement glow, you probably want to start with an understand of how blur works. Gaussian blur is actually pretty versatile...
First, let's take a look at DropShadow, which is nothing more than a special version of blur. You take your picture, convert it to grayscale, blur it, and put it under the control with a slight offset. In the case of text, since the background is transparent (important!), the shadow will appear to be behind the text.
A simple way to perform Gaussian blur is to do it in two passes, taking advantage of the fact that Gaussian kernel is "separable". First pass is to blur horizontally, for each pixel computing the weighted average of 2R + 1 pixels, where R is the radius and the weights form a Gaussian distribution. Second pass is to blur the horizontally-blurred image vertically, using the same technique. This can be done as two pixel shader effects. See an example of a 1D blur in http://www.gamedev.net/community/forums/topic.asp?topic_id=427696 (I just did a simple search). You can't directly do it in SL3 beta since it doesn't support multiple effects. However, it is easy to hack... Just add a Border element containing the image, and put an VGaussian effect on the border and a HGaussian effect on the image.
I suspect this is the approach taken by the SL team, especially since the effect only supports up to a radius of 20. This approach has a complexity of O(R * N), where N is the number of pixels, so imposing the maximum value of R = 20 seems reasonable. However, it is possible to implement Gaussian Blur in O(N) times using a recursive Gaussian approach, still parallelizable but beyond what HLSL 2.0 can do, but I don't believe this is being done by the SL team, because if they do, they won't have to impose a 20-pixel limit. This also means that we shouldn't go too crazy with blurring, since it could get expensive.
Also don't forget to check out the WPF shader effects library (effects usable by Silverlight, with minor C# code change).
07-03-2009 12:46 PM |
Thanks for the explanation
07-04-2009 5:00 AM |
1. Make Silverlight crossplatform. (Silverlight is part of .Net. (Windows only), but Flash is NOT part of any framework (so universal).
2. Make Silverlight fastest.
3. Make Silverlight smallest.
4. Make Silverlight BEST!!!
07-04-2009 5:07 AM |
ernestasju:*. Silverlight is part of .Net. (Windows only), but Flash is NOT part of any framework (so universal).
07-04-2009 5:19 AM |
You're right that ported Silverlight versions is available on other operating systems (Moonlight on Linux, luckily Silverlight on Mac, but what's about Solaris and others?), but Flash is Flash on almost every operating system and Flash doesn't requie .Net or Mono or other frameworks. That's why Flash is so universal. And more one thing, Moonlight on Linux is so slow, then Flash runing good. But still i'am not Flash fan (developing on Flash is terrible (C# vs Action Script. C# wins....) ).
07-04-2009 5:35 AM |
ernestasju:Flash doesn't requie .Net or Mono or other frameworks
Silverlight does not need any of those Frameworks either.Silverlight has it's own framework.Flash infact has two, one that supports ActionScript 1&2 and a complete new onw for object orientated ActionScript 3.The OS support has nothing to do with underlying technology, but with choice.
ernestasju:Moonlight on Linux is so slow
It has still some issues, but the speed is the same.
ernestasju:what's about Solaris and others
Flash does not work on Solaris either, at least not stable, the issue and bug list is as long as an aircraft carrier.
http://labs.adobe.com/technologies/flashplayer10/releasenotes.html
It would be stupid to delay development, possibly compromise speed and support for Silverlight just to support another 0,000001 % operating systems.
For that matter, Silverlight doesn't run on DOS either and no one is complaining. Even Sun kicked Solaris out of the window cause it was so bad, they are now selling Windows Servers. Dead operating system died for a reason and should stay dead.
jimevans
9 Posts
07-05-2009 11:00 AM |
Bill Locke
8 points
8 Posts
07-05-2009 5:34 PM |
I apologize if I might list some things that might be available in 3.0. Most of our experience is in SL 2.0 and I have just converted a SL 2.0 application to 3.0 and am moving some of the EF stuff to RIA with a custom DAL. The original program that we are taking to SL is a fairly mature win32 program.
1) Printing Support – within our application we use Crystal Reports for the major portion of our printing. We can continue to handle this using the hosted viewers of Crystal. However there are several options that we give to users that are more like a quick and dirty print option than a report, like printing the contents of the current grid or list view on the screen to the printer. This is more than a screen grab, because the lists can be quite large. We have .net code to do the printing, it would be nice to port it.
2) Source code – having source code to the controls is key to our ability to customize them to our needs. We very seldom get away with using stock controls.
3) Combobox – this control could use some fixing. I don’t know if the SL 3.0 control is better, I know our changes to this control are still working in 3.0. A) Ability to bind to data using the integer version of the foreign key rather than to entities produced by the EF. We have managed to convince the EF to do this the way we want and have modified the combobox to support this. Perhaps it is a lack of understanding on our part, but we just couldn’t make things work right any other way. B)Better mouse support – scroll wheel in particular. C)Keyboard support – for navigating to items in the list through the letter typed on the keyboard. We managed to support this ourselves as well. D) Better integration with loading of data. The lists in the combobox takes some time (we’re working with screens that may have 4 or 5 tab pages, 100+ controls and as many as 10-20 comboboxes). Because of the async nature of binding , if the user interacts with the combo before it has the list loaded, then the combo would never drop correctly. We wrote an integration between our data control and the combo control that keeps the combo disabled until the list is finished loading. The data control handles all the async stuff and notifies the controls when done.
4) OOB - This next item, I don’t know if it’s the WCF team of EF team or Silverlight team… Anyway, from what I understand of the network stack to support EF, WCF, (probably RIA), it’s written on top of the browser. This means that those applications won’t run OOB. This means that most LOB apps aren’t going to run OOB, probably the one class of apps that would benefit the most from that. This also means that there are subtle (and not so subtle) differences between browsers (and OS) in terms of the network stack. I can’t get our program using the EF to work on a MAC while it runs great on a PC. I’m hoping RIA solves that for us.
5) DataSet - This item, I’m sure I’ll take some graf for, but I’m going to say it anyway. Please provide a loosely typed object for data binding client side (like the DataSet, DataTable, whatever). Generic types are great (I guess), but we do not want to have to regenerate client side code every time we change the database. I understand that most people will probably write XAML and use strongly typed objects and so what difference does it make, the client side code will change anyway. But, we don’t use the screen designer in .NET or SL or WPF or the next new technology. We use our own designer and translate the definitions for whatever engine will deliver the UI. We deliver all of our screen definitions through Sql and build the controls on the fly, within code (well %80 of them anyway). In the win32 program we do not change any code to deliver changes to existing screens or new screens (there are exceptions). With Silverlight, I can’t replicate that code (without going a lot deeper than a small team is able). We are using our screen definitions in SL (translated from the win32 program), but have to use the EF and build the EF models and use T4 templating on the client side to build everything we need there to update those models. It looks like RIA will help, by building the client code for us and let us use models on the server side we build ourselves (maybe), but it’s still a lot messier than what we’ve been doing.
07-05-2009 9:09 PM |
@Bill
1. Printing is high on the request list for Silverlight 4.
2. The toolkit includes source code. I'm not sure what the plans are for the SDK and core controls.
3. Try with Silverlight 3 RTW and let me know how it works for you. I'm curious if there's anything there that may help you out.
Would you still want scroll wheel support if it didn't work cross-platform? For example, the browsers on the mac do not expose the appropriate events/notifications to allow the scroll wheel to work in a plug-in.
4. OOB has very good networking support. Start a thread on anything that isn't working.
As for your EF app, can you explain what's happening to make it not work on the mac? RIA isn't going to do anything magical to enable a cross-platform application - Silverlight does that in its core. If you're running into problems getting it to work on the mac, please start a thread with the issues. Keep in mind that Silverlight 2+ only works on Intel-based macs, so if you're running into issues with a powerpc mac, that is expected -- the platform is not supported.
5. We're doing a very large data-driven application here. DataTable/whatever isn't going to be your savior no matter what -- it doesn't contain enough metadata. What you can do, and this is something my colleage on the team did, is write a custom IValueConverter that takes in the field name for an item in a dictionary. Behind the scenes it does the binding for you and it all pretty much works as expected.
If you want to regenerate entities to take advantage of RIA Services, Silverlight 3 validation and annotation, then you'll need to run a server-side process to generate those entities when the metadata changes.(We're doing this for our actual entities, not our screen definitions). As much as you can, try and exist within the Silverlight ecosystem. The more you stray from standards in binding and controls, the less likely you'll be able to take advantage of the advances in RIA and business app dev in Silverlight.
Hope that helps.
halo2red...
07-05-2009 9:18 PM |
I would Love a Translate built in it. So when someone writes something in german, It would Translate into English or whatever your lang is. I hate that the most about translating sites.
07-06-2009 11:17 AM |
Thanks for the reply Pete.
1. Good
2. Hopefully they'll decide to release that code, that's where most of the action is.
3. So far it hasn't helped much, but I haven't dropped back from our customized control yet.Scroll Wheel - probably not, I can still see my testing group glaring at me though. They don't care about the mac much.
4. OK, when I get there, I was basing my OOB comment on some other threads I'd read, not experience.
As for the MAC (that was based on experience), here is a thread I started on that. I've remote debugged the application from VS 2008 (good work there) and ran it through Fiddler with the PC as a proxy. It just doesn't work. http://silverlight.net/forums/t/103417.aspx
5. I would really like more information about that if you have it (I may have seen that article if he wrote one, sounds familiar, problem was updating data or something). Getting to the same level in this technology from "regular" .net has been a bit of struggle. I can understand that we should be able to write the code to do what you are describing, I just haven't gotten the foothold I need to get it done. Understand we have to support the full CRUD cycle.
I'm doing my best to stay within that ecosystem. My problem is probably as much not understanding how to handle this technology in a "generic" way. We NEVER write a "normal" data access program. Everything is driven from the definitions in the database. In a way that stuff is metadata, even if we aren't handling it that way. The application code is dumb, it never knows what the screen looks like or even what the data looks like (there are exceptions obviously), only the defining data knows that. Even the validation works that way for us. We don't decide what fields are required, that is configured along with security within the application. We also use regular expressions for validation, so what we're doing blends very well with what is happening in RIA (BTW, my experience with RIA runs more like 2 days, but I am retreiving the screen definitions that way now). We've developed our own ways of doing things and we have to decide what will fit with the Silverlight way and what we'll have to write. Hopefully our misunderstanding won't lead us down the wrong road. My request may stem as much from that misunderstanding as anything else.
We're just 3 guys and only been at this for 6 weeks. We're delivering our first SL app in a client testing environment this week (SL 2, Entity Framework). It should make production sometime in August. We're very excited about this platform.
07-06-2009 1:34 PM |
Once again, need better WriteableBitmap APIs. The following is some comments I made on another thread:
Good to hear that you made some progress.
One comment on WriteableBitmap.
I don't think the SL team will provide access to the WriteableBitmap in background thread. The reason is that SL must have full access to all UI assets whenever it is running the UI thread, so it really cannot afford to let you muck with it in the meantime.
I do hope that the SL team can provide better WriteableBitmap APIs. Currently the only pixel access is the [] operator, which I image has tons of checks (security and bound) on a per-pixel basis. This is highly inefficient especially we all want to use WriteableBitmap in a very heavy way. Why not provide wholesale array copy API? For example,
WriteableBitmap.CopyFrom(int[] Src, int SrcWidth, int SrcHeight, Rect Region, int TranslateX, int TranslateY);
WriteableBitmap.CopyTo(int[] Dest, int DestWidth, int DestHeight, Rect Region, int TranslateX, int TranslateY);
This way, whatever check only needs to be done once.
It is still okay to require that these APIs be called from the UI thead. But at least this approach would allow us to write decent code to perform all the maths in the background thread, and make one async call to push the result to the WriteableBitmap.
StefanOlson
215 points
90 Posts
07-06-2009 7:23 PM |
Psychlist1972: Would you still want scroll wheel support if it didn't work cross-platform? For example, the browsers on the mac do not expose the appropriate events/notifications to allow the scroll wheel to work in a plug-in.
Yes! Absolutely! Why penalize 90% of users because 10% of users don't have that functionality available. Customers running on Windows are going to expect mouse scroll wheel support
...Stefan
07-07-2009 12:01 AM |
StefanOlson: Psychlist1972: Would you still want scroll wheel support if it didn't work cross-platform? For example, the browsers on the mac do not expose the appropriate events/notifications to allow the scroll wheel to work in a plug-in. Yes! Absolutely! Why penalize 90% of users because 10% of users don't have that functionality available. Customers running on Windows are going to expect mouse scroll wheel support ...Stefan
I think I am sick of the "lowest denominator" argument as well... Just because a platform -- Mac in particular -- doesn't support this almost universal, everybody-loves-it, feature (scroll wheel) doesn't mean such supporting such feature is of zero value. Yeah it may make some Mac owner unhappy, but so what? When are they happy with Microsoft products anyway :D.
Also, if the "lowest denominator' argument holds, why on earth is Silverlight supporting multitouch in SL3 RTW?
Apple, ironically, completely ignores the mantra to be the lowest denominator. They do whatever they want, sometimes to their detriment and sometimes to their wild success. You can say that's why they are NOT as successful as Microsoft in the old days. But I would say that's also why they ARE successful in modern days... with innovative designs as iPod and iPhone that break all rules and make no apology. What? Smartphone with no physical keyboard?
So how to deal with platforms that don't have scroll wheels? Well, I'd say leave it to the developers. If I am developing an SL game for iPhone (for the sake of argument let's say such a thing is possible!), I sure as hell won't assume that the scroll wheel is available. But if I am developing a heavy business app that benefits tremendously with scroll wheel, who is anyone to argue that I *DON'T* need a scroll wheel?
One last argument... some may say that the Javascript hack is good enough. My answer is a four-letter three-letter word: OOB
07-07-2009 12:06 AM |
FWIW, the mousewheel mention was not an official opinion of any sort; I was just curious as to the community take on it.
I personally believe the subset of things that work only on platform X need to be kept to a very bare minimum. That's not to say something can't be in there, just that it needs to be carefully considered. The bigger the gap, the less likely Silverlight is to be taken seriously.
07-07-2009 12:41 AM |
Psychlist1972:FWIW, the mousewheel mention was not an official opinion of any sort; I was just curious as to the community take on it. I personally believe the subset of things that work only on platform X need to be kept to a very bare minimum. That's not to say something can't be in there, just that it needs to be carefully considered. The bigger the gap, the less likely Silverlight is to be taken seriously. Pete
I agree with your position in general. The subset needs to be kept to a very bare minimum. Scroll wheel happens to be one of the few exceptions, IMHO.
07-07-2009 10:43 AM |
Wow, didn't think the mouse wheel would be that hot a topic. In reality it causes me as much greif when it is supported. The wheel doesn't always do what the user expects and of course expectations vary with the user. When you click on a scrollbar, you know what is going to happen. When you use the wheel, some programs scroll what the mouse is over, others scroll whatever has focus. .Net by default reacts in the later way. The scroll wheel seems like a simple thing, but it really becomes a pain to support it a way that will make the majority of users happy.
I wouldn't cry about it not being supported, but like I said, that's not my QA groups opinion, it's one of the first things they mentioned about the SL app.
ColinBlair
5623 points
1,090 Posts
07-07-2009 5:42 PM |
Psychlist1972: FWIW, the mousewheel mention was not an official opinion of any sort; I was just curious as to the community take on it. I personally believe the subset of things that work only on platform X need to be kept to a very bare minimum. That's not to say something can't be in there, just that it needs to be carefully considered. The bigger the gap, the less likely Silverlight is to be taken seriously. Pete
I think a built in mousewheel implementation which is limited to supporting controls that inherit from RangeBase (or some other similar restriction) would be fine. Something that enforces the fact that the mousewheel is optional much shouldn't be mandatory for your UI to function.
Kbalz
213 points
185 Posts
07-08-2009 9:31 AM |
Mouse wheel is a must. 85% of users are on a PC and use mouse wheel everyday.. so when they are trying to use the mouse on my datagrid that is in a scroll viewer, they have a confused look on their face. Sorry that Safari hasn't brought their browser into this century, but we need to get some work done.
rshelby
273 points
64 Posts
07-08-2009 12:57 PM |
Extend 3D capabilities for example:
Creating a 3d cube by simply taking a rectangle and setting a thickness / depth property. Or doing the same thing to a circle in order to make a 3d cylinder.
07-08-2009 1:08 PM |
Double-click support. For example, in the DataGrid, allow a user to double-click an item, so that a Child Window can be displayed, or other action performed. Users look at Silverlight apps as so "Windows-like", that they expect to be able to double click on items and have default actions occur.
Zack123
4 points
4 Posts
07-09-2009 12:14 PM |
Currently the proxy generation marks Begin and End operations as private.
So now I need to edit the reference.cs everytime I refresh a service reference, or do a yucky macro to find replace, which is always iffy.
This is not great if you have a service caller helper that abstracts away the service calling process, where you need to pass around functions.
It is great that you want enforce async, but the best way to deal with that is to implement functional handlers, and then you make it difficult by not exposing the functions for use.
Zack
Helper<Request, Response> handler = Helper.CreateAsyncResult<Request, Response>(proxy.BeginMethod, proxy.EndMethod);
handler(request, (response, async) =>{ lock(this) { Dispather.BeginInvoke(() => {//Code goes here to act on response if UI bound }); }});
07-09-2009 1:43 PM |
ksleung: StefanOlson: Psychlist1972: Would you still want scroll wheel support if it didn't work cross-platform? For example, the browsers on the mac do not expose the appropriate events/notifications to allow the scroll wheel to work in a plug-in. Yes! Absolutely! Why penalize 90% of users because 10% of users don't have that functionality available. Customers running on Windows are going to expect mouse scroll wheel support ...Stefan I think I am sick of the "lowest denominator" argument as well... Just because a platform -- Mac in particular -- doesn't support this almost universal, everybody-loves-it, feature (scroll wheel) doesn't mean such supporting such feature is of zero value. Yeah it may make some Mac owner unhappy, but so what? When are they happy with Microsoft products anyway :D.Also, if the "lowest denominator' argument holds, why on earth is Silverlight supporting multitouch in SL3 RTW?Apple, ironically, completely ignores the mantra to be the lowest denominator. They do whatever they want, sometimes to their detriment and sometimes to their wild success. You can say that's why they are NOT as successful as Microsoft in the old days. But I would say that's also why they ARE successful in modern days... with innovative designs as iPod and iPhone that break all rules and make no apology. What? Smartphone with no physical keyboard?So how to deal with platforms that don't have scroll wheels? Well, I'd say leave it to the developers. If I am developing an SL game for iPhone (for the sake of argument let's say such a thing is possible!), I sure as hell won't assume that the scroll wheel is available. But if I am developing a heavy business app that benefits tremendously with scroll wheel, who is anyone to argue that I *DON'T* need a scroll wheel?One last argument... some may say that the Javascript hack is good enough. My answer is a four-letter three-letter word: OOB
Look like my wish is granted... *NATIVE* scroll wheel support in SL3 RTW!
07-09-2009 2:41 PM |
Hi Zack,
For the proxy also excited an interface ythat contains the methods you want. The implemation implements that interface explicit. So now you do:
Proxy p = new Proxy();
but you could allso use
IProxy p = new Proxy();
Now you have direct access to the begin and end methods
fgoulet
07-09-2009 2:44 PM |
A bridge to SQL Server compact edition.
Would be awesome to have a SQL Server compact database file in isolated storage and been able to retrieve and update data via linq.
A must for offline or occassionaly connected lob application I think.
nyhetsgr...
1365 points
300 Posts
07-09-2009 3:25 PM |
Right click support is the number 1 feature I am waiting for.
07-10-2009 5:26 AM |
Exactly what I need - just a few changes to my proxy wrapper! Thanks davesmits.
ted7
13 points
07-10-2009 8:46 AM |
I'd like to see: 1) built in PhotoSynth type things. 2) A component for doing virtual reality panoramas and seamlessly linking them, 3) an easy way of doing lip synching with models like what MS did with MS Agent technology a few years ago.
Sledge70
4970 points
899 Posts
07-10-2009 9:32 AM |
A: True right mouse click support (Not using windowless mode).B: Double Click built in
Come on, we're trying to write LOB apps here whereby we know what our target browsers are.
Frank Wu
94 Posts
07-10-2009 9:39 AM |
Support mobile OS and real cross-platform
nosuchth...
53 points
51 Posts
07-10-2009 11:58 AM |
Mobile support!!
fool
7 points
30 Posts
07-11-2009 7:58 PM |
My bias is that I'm mainly interested in Silverlight as a platform for games, hence my wishlist:
1. RenderOptions BitmapScalingMode NearestNeighbor, as in WPF, that can be applied to an Image to control the magnification filter used.
2. Mobile platform support - including iPhone!
3. Better out of browser support, in particular being able to completely control the window size/resize/style/state(min/max/topmost) etc including changing these on the fly.
4. Binding StringFormat & RelativeSource FindAncestor conveniences - and whatever else can help close the portability gap with WPF.
My emulator proving ground, http://virtu.codeplex.com/Wiki/View.aspx?title=Silverlight, was reasonably successful so I think Silverlight has a lot of potential for games...
mabra
15 points
07-12-2009 10:02 AM |
Hi !
- Right mouse clicks [context menu] is an absolute MUST
- Number of maximum open TCP port should be enlarged, say 20
07-12-2009 11:10 AM |
right-click, right-click, RIGHT-CLICK! at least textboxes should have a minimal cut-copy-paste menu like their html counterparts. even Flash does have it for textboxes!
07-12-2009 11:28 AM |
pelister:right-click, right-click, RIGHT-CLICK
Especially on a mobile device or touch pad it will enhance the experience greatly ;)
07-12-2009 11:48 AM |
forgot to mention; WSHttpBinding was an important loss in sl3. the sessionless web services suck!
07-12-2009 2:42 PM |
pelister: forgot to mention; WSHttpBinding was an important loss in sl3. the sessionless web services suck!
I don't believe SL2 ever had support for wshttpbinding. It was basichttpbinding only.
07-12-2009 5:20 PM |
Psychlist1972: pelister: forgot to mention; WSHttpBinding was an important loss in sl3. the sessionless web services suck! I don't believe SL2 ever had support for wshttpbinding. It was basichttpbinding only. Pete
that's why i was complaining:) wshttpbinding is in my wishlist for sl4.
bridenou
33 Posts
07-13-2009 4:15 AM |
wshttpbinding is on my list as well......
I also have a couple of rather esoteric requests.
1) Ive been doing an app that requires dynamic assembly creation, however, it appears that once you create an assembly using Reflection.emit, you cant save it to a database.... would really be nice. (some way to get the bytes out) Maybe a way to compile code with the CodeDom classes would be nice as well.....
2) Also, it seems to be impossible to create a silverlgiht enable wcf interface that can be used from both the silverlight client, but also on the server side. meaning, my application needs to cnnect to various servers, all of them should be exposing the same service, so I would like to create the interface, create a bunch of server side services that expose it, and use in the client as well. However, because Silverlight does not use the same System.ServiceModel.dll as does normal .NET, you cant share these interfaces this way. Seems a fairly big flaw, unless Im misunderstanding somehting here. The only way Ive found around it is, create the interface on the server side, create a service reference on the client, then whenver you create a client object, repoint it s endpointAddress to the alternate servers. Seems a little cludgy.
Cheers
-brent
Montago
265 points
190 Posts
07-13-2009 7:27 AM |
1. MediaElement.SpectrumPulse event aswell as MediaElement.Equalizer[]
It would be cool to make visualizer effects for music playback in SL - the SpectrumPulse could pulse 30fps or something similar to CompositionTarget.Rendering... Where the developer may update the visualizer control using the values from the eventargs... i don't have a recipie for what i want in that event... but enough to visualize the music :-) (maybe look at the Winamp visualizer api's...? or just make something simple)
a way to control the Equalizer in the MediaElement - to boost bass, treble or whatever the user wants.
2. Click, Doubleclick, MouseScroll and RightClick event on all usercontrols that have MouseDown/Up/Enter/Leave events
I really miss a genuine Click event (mousedown -> mouseup on the same element) instead of having to code Click myself, its not impossible, just anoying to do it myself...
DoubleClick applies to the same issue as Click... possible to do it yourself, but anoying to do so....
MouseScroll just like we know it from HTML would be great.... although i've read that the Javascript-helper method work better then native support...
RightClick is a MUST !... WindowLess is not a solution, its a hack !!! ... Native Rightclick event is needed for business/game apps !!!
3. Keyboard Override
When making Business/Game apps, its crucial to support keyboard shortcuts !!
I want full support for Event cancelling, so that i may control what happens when the user press F1, F2, F3, F4, F5 etc... and i want to be able to make CTRL + [key] shortcuts !!! without having the browser to take my event. - its possible in Javascript, why not Silverlight ??
4. Clipboard support
I want to copy text/pictures to/from the clipboard in a UI-Userevent just like flash has it...
This would enable me to take screenshots with the PrintScreen key, and post it on a Silverlight site.
5. Mouse-Drop files
Would be VERY usefull to enable drop files using the mouse... for file-upload or editing...
dovepanda
07-13-2009 9:20 AM |
In a business app, I found these useful feature missing and wish for it in SL4 very much!(except printing, mousewheel, which is already well-known)
1. Binding.NotifyOnSource/TargetUpdated
WPF has it. It creates better user experience in scenerios, and it is hard and ugly to work around. With this achieved, the missing yet another key feature, SL Trigger feature would be easier to work around. It's of high priority to us.
2. Trigger, DataTrigger, MultiTrigger, Setter can Set Custom Property. Or At Least Custom Visual States
Trigger series is one of the most excellent WPF features. I still don't understand why it's missing. With triggers we can do a lot more than just simple visual states. And now SL is really limitted, in complicated scenerios. Also high priority.
3. MultiBinding
4. OOB Automatic Download Update and Start Newest Version
5. Style Type="{x:Type }"
WPF has it. Once defined a Style in a ResourceDictionary, it becomes global to all that references the ResourceDictionary. It saves a lot redundant coding, and maks code more elegant. Would be nice if it comes in SL4, but ok if it doesn't.
K2P2
983 points
319 Posts
07-13-2009 9:39 AM |
Z-Index bigger than Int16.MaxValue.
I feel like I'm back in the 80's. (8 bit processors.)
07-13-2009 9:42 AM |
Workflow (subset maybe) on the client side.
07-13-2009 10:05 AM |
K2P2: Z-Index bigger than Int16.MaxValue. I feel like I'm back in the 80's. (8 bit processors.)
Except that 8bit processors would limit you to a max value of 255 (if unsigned)
What's the use case for a larger Z-Index? Do you really have more than 32k items to order, or is it just convenience?
07-13-2009 10:17 AM |
This may be a strange request, but this is a wishlist, isn't it ?
No right click support !!
http://silverlight.net/forums/p/109982/248484.aspx#248484
miantosca
88 Posts
07-13-2009 10:54 AM |
Real performance profiling - like the full .NET framework has
07-13-2009 10:56 AM |
Printing support built in
Commands - like WPF
07-13-2009 11:10 AM |
Krasshirsch: This may be a strange request, but this is a wishlist, isn't it ? No right click support !! http://silverlight.net/forums/p/109982/248484.aspx#248484
i can understand your point. but first of all, it's not ok to concentrate that much on mobile devices. sl is not a technology for the mobile yet, even after this much years it still has even no beta for any mobile device. and i doubt it will have for the mobile for the near future.
besides, why do you insist on taking right-click as a flaw? i read your post on your new thread, and i found it so much extraordinary taking right-click as some sort of platform dependency. remember that .net now has a successful linux version, mono, even though microsoft had carefully injected into the bcl dozens of dependencies on win32:))))) besides, flash does have a minimal cut-copy-paste right-click menu and it has a mac version for years. its iphone version will be soon around.
when i say right-click, i dont mean a full blown "context menu", but we should be able to copy-paste text on the text-related controls. trying teach copying some text by ctrl-c is a pain in the a** for the average joe when he is ignorant of what a browser is!
07-13-2009 12:05 PM |
Hey pelister,
thnx 4 takin an interest in this, thought I comment here, but It would be better to leave the landry list out of discussions ;)
pelister:sl is not a technology for the mobile yet, even after this much years it still has even no beta for any mobile device. and i doubt it will have for the mobile for the near future.
I believe this is incredibly short sighted of you, . When developing an application of this magnitude you can't just plan for today or tomorrow, you have to look far ahead and anticipate potential platforms, mobile support is no pipedream anymore but planned and if there is one thing we know about microsoft, they finish their plans, not always on time and with varying success, but they finish.
It is true that the development of the mobile plugin is slow and it will probably take some more years, but there will be a plugin for mobile devices eventually, of this I am sure
pelister:besides, why do you insist on taking right-click as a flaw?
Because it is, it complicates usability and encourages UI developers to create bad UI's.
pelister:besides, flash does have a minimal cut-copy-paste right-click menu and it has a mac version for years. its iphone version will be soon around.
Yes and do you know how long Adobe is already plotting and "workarounding" to get flash running on a phone ?And look what the result is http://www.adobe.com/products/flashlite/version/
Action Script 1 & 2 ? Even if, and i gotta stress the if, your flash ap runs on the phone it will behave different as expected or do you think apple will invent a right click to open this menu to copy and paste stuff ? No they won't ..It's those small things that plague adobe and make flash lite a very, very poor second class citizen.So I ask you, why not trying to do it the right way from the start ? You don't need right click to open a menu, there are dozents of different possibilities.
MRichard...
16 points
07-13-2009 12:24 PM |
1. Printing - The ability to take a report and print it is invaluable, and a huge missing element from SL that will prevent me from building business applications 100% in Silverlight.
2. Wheel and Right mouse button support. To make a modern day application that has the usability people are used to, fully supporting the mouse is very important. Maybe for those mac users out there, you implement it the way it is in WINMO with a click and hold.
3. Web cam support - the ability to capture video from a web camera, and save it, upload it, stream it would be very beneficial.
4. Upload Control - i know there are a lot of options out there, but why not just include good upload tools out of the box.
Thanks,
Mark
07-13-2009 1:25 PM |
Yeah totally forgot: Managed profiler for SL apps!!!.
(although I don't think this is actually version dependent)
To the anti right click people:
One of the key concepts you're missing is that right click can already be handled arbitrarily with little workarounds.
If a project needs right click, developers will do these workarounds for IE and Firefox, which will already be pretty expensive to test-through and keep up with updates, so they may just leave out the less popular brwosers / OS.
I think it's pretty clear that not getting built in support for right click will rather decrease overall platform independence of SL applications, than extending it.
If you give someone the possibility to screw up, he will ! ? Well then why even let developers do custom controls... Actually why don't just give developers some built-in templates for their apps and take away the (resource-limited) turing-complete programming language from them, since it allows for so many screw ups. Yeah, I think that would be best direction for future Silverlight releases.
When I say right-click, I dont mean a full blown "context menu" neither. I mean arbitrarily handling of the right click as in WPF.
Even if not for every platform Silverlight supports, it would still be much wider than what small teams can test for if they decide to use right clicks now.
Right click will be used with or without official support for it, you can count on that one.
It's just that without official support, applications will require more work to achieve the same thing, and they will be less stable.
07-13-2009 2:06 PM |
Hey Herthoren,
Herthoren: One of the key concepts you're missing is that right click can already be handled arbitrarily with little workarounds. If a project needs right click, developers will do these workarounds for IE and Firefox, which will already be pretty expensive to test-through and keep up with updates, so they may just leave out the less popular brwosers / OS. I think it's pretty clear that not getting built in support for right click will rather decrease overall platform independence of SL applications, than extending it.
Very true and I wouldn't mind if this bug would be fixed in SL 4 to conform with the OOB version.Especially in a business RIA these two worlds "browser / and out of browser" should behave equally.
If you give someone the possibility to screw up, he will !Perhaps I should rephraze this. Taking away the possiblity to include right click, keeps one from doing potential fatal business decisions.So you probably ask youself, what the hell does he know about my business ?Nothing, but I know one thing ... No one will refrain buying or using an application because it uses no right click.But platform dependence, even on such small scale, can be a reason to take another product, that is truly independent.
Herthoren: Well then why even let developers do custom controls... Actually why don't just give developers some built-in templates for their apps and take away the (resource-limited) turing-complete programming language from them, since it allows for so many screw ups. Yeah, I think that would be best direction for future Silverlight releases.
Well then why even let developers do custom controls... Actually why don't just give developers some built-in templates for their apps and take away the (resource-limited) turing-complete programming language from them, since it allows for so many screw ups. Yeah, I think that would be best direction for future Silverlight releases.
This was the common strategy in Winforms, Java Swing and tons of other UI widgets, but exactly creating custom UI's is what Silverlight is all about.Without this feature Silverlight would have no reason for existence.
But even in this "presumably free" environment rules and protocols exist which are enforced as much as possible, that's why classes are sealed, that' why the security restrictions exist and why the API surface is kept as small as possible to ensure safety and to take as many possibilities to "screw up" away as possible. You cannot ensure perfection, but it's always risky to break these restrictions in favor for a single platform. Perhaps the right click will create problems, perhaps not, but in the end, if you've favored one, why not favor more ? If one is not carefully, SL will end up like flash running on all plattforms ... but different.Perhaps not today or tomorrow, but at some point a company will get into problems, first turning to SL for it's independence ending up creating different versions for different platforms, because of such small issues.I know this is all very theoretical and seems far fetched and one might say, what do I care about Silverlights future I need a contextmenu today.Perhaps ... but this is not about company XY and there needs, this is about the future of this, dare I say "Damn Fine" peace of work called Silverlight.
Up till now Silverlight is remarkably similar on all platforms, the "can and can't on Platform XY" list is really small compared to the one of Adobe's flash, I'd like to see it staying this way.
07-13-2009 3:25 PM |
I forgot to add HTML rendering - lots of content in HTML and easy to author - would be nice to display HTML in SL without having to play the windowsless mode / div overlay tricks that I had to do in the past
07-13-2009 3:42 PM |
Krasshirsch: SL will end up like flash running on all plattforms ... but different. Up till now Silverlight is remarkably similar on all platforms, the "can and can't on Platform XY" list is really small compared to the one of Adobe's flash, I'd like to see it staying this way.
SL will end up like flash running on all plattforms ... but different.
If Silverlight will force developers to make the same UI for 30" monitors with keyboard and mouse input, that they do for 2" cellphones, Silverlight is going to fail without a question over platforms that allow developers to take the best advantages of each system they're developing for.
I could go much deeper into this, but it's just this simple. You can make UIs that work on all systems, but users will be more performant using UIs that takes advantage of the current system they are using it on.
Yes, platform independence is a great thing about Silverlight, but it's hell not the only important feature why people want to use it for.
If it allows to use the advantages of different platforms, developers can still make applications that are totally platform independent, but they also can make them competitive with other apps on certain systems.
07-13-2009 3:47 PM |
miantosca: I forgot to add HTML rendering - lots of content in HTML and easy to author - would be nice to display HTML in SL without having to play the windowsless mode / div overlay tricks that I had to do in the past
How much HTML rendering do you want? What are the use-cases? I know where I'd use it, but I'm curious what others think.
Keep in mind that to have full HTML rendering, you're recreating a browser, inside a plug-in, inside the browser.
If the renderer would run across tags it doesn't support, what should happen? Toss them or display them? What about external stylesheets? Javascript? Object tags?
Not being difficult, I'm just wondering how much support is needed for folks to consider it usable.
07-13-2009 3:49 PM |
Oh, and let's say for a moment that there's some way to just host the underlying browser rendering engine without recreating it. Would the differences in how things are rendered be worth it? Keep in mind that until that, Silverlight has had an excellent x-plat story in terms of what renders in one browser/os is the same as in the others.
07-13-2009 3:51 PM |
Yes object tags please! We could include the silverlight app recursivly into itself, until the CPU burns in flames ;)
07-13-2009 3:55 PM |
For me, I wouldn't need external style sheets, javascript, or object tags. Just the ability to render HTML text, including images, and supporting HTML 2.0 through 4.x and beyond. We would use it more as a "Rich Text" rendering, rather than interactive HTML. We have a lot of data (Exchange email), which has been archived and converted to HTML. We want to render the HTML of the original message correctly.
07-13-2009 3:57 PM |
Psychlist1972:Oh, and let's say for a moment that there's some way to just host the underlying browser rendering engine without recreating it. Would the differences in how things are rendered be worth it? Keep in mind that until that, Silverlight has had an excellent x-plat story in terms of what renders in one browser/os is the same as in the others. Pete
That would be okay, if it wasn't displayed in "windowless mode". We need the ability to have the content "flow" with other Framework element objects, and also needs to be displayed within such controls such as a ChildWindow.
07-13-2009 4:04 PM |
tfmillet: For me, I wouldn't need external style sheets, javascript, or object tags. Just the ability to render HTML text, including images, and supporting HTML 2.0 through 4.x and beyond. We would use it more as a "Rich Text" rendering, rather than interactive HTML. We have a lot of data (Exchange email), which has been archived and converted to HTML. We want to render the HTML of the original message correctly.
For this behaviour there is a really simple way to accomplish this, you could write a simple parser that does nothing more than convert HTML to XAML,
you know <div></div> -> <Grid></Grid> and so on, when you done, rush it through the XamlReader and voilá, you got yourself an asskick html renderer.
07-13-2009 4:27 PM |
Psychlist1972: Oh, and let's say for a moment that there's some way to just host the underlying browser rendering engine without recreating it. Would the differences in how things are rendered be worth it? Keep in mind that until that, Silverlight has had an excellent x-plat story in terms of what renders in one browser/os is the same as in the others. Pete
How do you wish to accomplish this feat in OOB ?
not a lot of rendering - really just enough to allow an app to format user entered content using simple tags like bold, italic, ul and ol - basically rich text type formatting.
Krasshirsch suggested creating a simple parser to convert html to xaml - I think that could easily be done and will have to do it for the project I am working on right now - would be nice to have something in the toolkit that everyone could use and not have to re-invent the wheel (and test the heck out of that wheel)
- Michael
07-13-2009 4:31 PM |
Krasshirsch: How do you wish to accomplish this feat in OOB ?
Reminder: I don't work for Microsoft (in case there's any confusion)
Silverlight Out of Browser is still hosted by the browser's rendering engine. On Windows, that's shdocvw and related stuff. On Mac, it's the rendering engine bits from webkit.
So, it's all still there.
07-13-2009 5:01 PM |
Psychlist1972:Reminder: I don't work for Microsoft (in case there's any confusion)
Psychlist1972: Silverlight Out of Browser is still hosted by the browser's rendering engine. On Windows, that's shdocvw and related stuff. On Mac, it's the rendering engine bits from webkit.
You learn new things all the time :) So no browser installed, no silverlight huh ? a pitty...
07-13-2009 5:11 PM |
Krasshirsch: So no browser installed, no silverlight huh ? a pitty...
So no browser installed, no silverlight huh ? a pitty...
I'm not so sure it's as cut and dry as that. shdocvw in windows is pretty core. It's probably even present in the "no browser" EU builds. MS wasn't just pulling a marketing stunt when they said IE was core to the windows operating system, they were serious. They had made a decision to base a number of things on it.
I can't speak to what happens on the mac, but I believe the core safari bits are there much like the core IE bits on windows.That's a gut understanding, not based on actually doing any testing or verification.
07-13-2009 5:25 PM |
Opinions are my own. Oh, and I don't work for Microsoft.
Great, now I got you to update your signature ... man thats not my day ...
07-13-2009 5:30 PM |
Krasshirsch: For this behaviour there is a really simple way to accomplish this, you could write a simple parser that does nothing more than convert HTML to XAML, you know <div></div> -> <Grid></Grid> and so on, when you done, rush it through the XamlReader and voilá, you got yourself an asskick html renderer.
Except for the fact that in a text block it's impossible to provide hyperlinks with surrounding text unless you use some very hacky workarounds which all have major disadvantages. One of the key issues for Silverlight 4 is to provide support for flow documents which are available in WPF and would allow rich text content. Then an HTML to flow document converter would not be that difficult (and there is actually ready one for WPF).
07-13-2009 5:36 PM |
Montago: 3. Keyboard Override When making Business/Game apps, its crucial to support keyboard shortcuts !! I want full support for Event cancelling, so that i may control what happens when the user press F1, F2, F3, F4, F5 etc... and i want to be able to make CTRL + [key] shortcuts !!! without having the browser to take my event. - its possible in Javascript, why not Silverlight ??
I absolutely agree with this, any line of business application needs to be able to support Alt+[Key] shortcuts to provide easy access around the screen. People working with data entry applications for example do not want to be using the mouse all the time.
PomAH4uK
07-13-2009 5:47 PM |
Fix Silverlight WCF Duplex.
Web camera
Microphone.
UDP :)
Opporunity save Bitmap to image.
Jamebo
07-13-2009 8:51 PM |
1. LicenseProvider
2. Better supports for Eastern language Input Method Editor (TSF) for Japan, China
07-13-2009 11:43 PM |
PomAH4uK: Fix Silverlight WCF Duplex. Opporunity save Bitmap to image.
What do you want changed with WCF Duplex?
On the bitmap/image question, are you asking for something with a jpeg/png codec? You can already save the bits out to a file, and you can use the png encoder written during the SL2 timeframe (probably with some minor modification) to save out a PNG. a lossless bitmap should be even easier.
jsneeringer
07-14-2009 12:46 AM |
borismod...
07-14-2009 11:29 AM |
Our main wishes are improvements of integration between WCF and Silverlight: 1. Error handling support: exception thrown from WCF can be caught in Silverlight 2. Compression of serialized data 3. Better localization support: SL would detect user's language and load dynamically required Resource file
07-14-2009 11:56 AM |
This is more of a Visual Studio request.
It would be very handy if when refactoring a control name in code behind it also changes the name of the control in XAML and vice versa.
Same for event handlers in both directions.
estern
404 points
122 Posts
07-14-2009 12:14 PM |
The ability to edit the code while debugging.
07-14-2009 12:41 PM |
estern: The ability to edit the code while debugging.
Boy, do I miss that ability!
07-14-2009 6:03 PM |
this would absolutely make my life as a developer so much better! I would also like to see the ability to change XAML code while debugging,
NigelLee...
32 points
22 Posts
07-15-2009 5:40 AM |
1. LayoutTransform.
RenderTransform doesn't move related items, so I end up with text over images or even worse, under images when I resize an image with RenderTransform.
2. Credentials in WebClient
I want to call a web service with username & password.
3. tilt delta on the mousewheel, so I can scroll right & left as well as up and down.
4. Proper inheritance for XAML created classes without any fumbling around.
5. Better resolution of actualheight/actualwidth - it is too often 0. I dont want to have to Measure() everything. The real problem is that I cannot understand how and when the sizes of things get set. Can you make it much easier please, I am not young enough to have sufficient remaining years to investigate such things.
6. Ability to create uielements in separate thread from UI thread
7. A move away from xaml and back to code. It is easy in xaml to add an image or a colour, but in code, you have to create intermediate steps like bitmapimage for the source of an image and solidcolorbrush for the backgroudn of a whatever. Can code compilers not be as clever as XAML compilers when it comes to understanding the programmer's wishes. Personally I am finding XAML less an less useful, especially when it comes to debugging.
8. crossdomain.xml and clientpolicy.xml - don't want them, don't understand why I want them.
9. I have a feeling that creating uielements in silverlight is slower than in WPF. I haven't tested it, but will one day get around to it.
10. Can I have a project with two outputs in Visual Studio, one for WPF and one for silverlight. I want to see Silverlight's limitations highlighted in the code, so I can work around them like #if silverlight ... or #if wpf .... This would allow me to create a proper windows wpf application and a silverlight one, so I can give full power to a desktop user, and a cross platform application to a casual user without having to manually maintain two projects. I have a feeling that running a silverlight application out of browser on a windows box is a pointless and restricting activity.
All in all, I am very pleased with Silverlight, although you may not think so from the above comments.
07-15-2009 6:21 AM |
NigelLeeming: 5. Better resolution of actualheight/actualwidth - it is too often 0. I dont want to have to Measure() everything. The real problem is that I cannot understand how and when the sizes of things get set. Can you make it much easier please, I am not young enough to have sufficient remaining years to investigate such things.
in every UIEvent you can read those properties... (atleast thats how i do it)
NigelLeeming: 6. Ability to create uielements in separate thread from UI thread
i agree... UIElements/FrameworkElements who's parent are NULL shouldn't complain about thread exceptions... hence they arent drawn...
NigelLeeming: 7. A move away from xaml and back to code. It is easy in xaml to add an image or a colour, but in code, you have to create intermediate steps like bitmapimage for the source of an image and solidcolorbrush for the backgroudn of a whatever. Can code compilers not be as clever as XAML compilers when it comes to understanding the programmer's wishes. Personally I am finding XAML less an less useful, especially when it comes to debugging.
i totally agree !... why are there 120 colours to choose from in XAML, but only 15 in C# ???
NigelLeeming: 8. crossdomain.xml and clientpolicy.xml - don't want them, don't understand why I want them.
its a security issue (preventing Silverlight Spam-bots/Botnets ) get used to it - its for the greater good !
NigelLeeming: 10. Can I have a project with two outputs in Visual Studio, one for WPF and one for silverlight. I want to see Silverlight's limitations highlighted in the code, so I can work around them like #if silverlight ... or #if wpf .... This would allow me to create a proper windows wpf application and a silverlight one, so I can give full power to a desktop user, and a cross platform application to a casual user without having to manually maintain two projects. I have a feeling that running a silverlight application out of browser on a windows box is a pointless and restricting activity.
can't be done... SL is meant to run i a browser, with async WCF calls and HTMLPage interactions and events only available in a browser... there are too many differences to SL from WPF to make this possible ...
07-15-2009 10:51 AM |
Oh, I totally forgot, Silverlight for mobile phones. I work for a company that provides video services for companies, and I have a real need to be able to build a silverlight app for mobile that would allow for uploading and managing video and images. Of course, I'd like it for WinMo, Android, Blackberry, Pre, and iPhone (good luck on getting that one!).
enantiom...
39 points
59 Posts
07-17-2009 10:54 PM |
I would like DataTemplate to have a DataType property so that silverlight can automatically discover the default template to use. Good for MVVM pattern:
<DataTemplate DataType="{x:Type vm:FooViewModel}"> <v:FooView /></DataTemplate>
Gimme!
pkr2000
1219 points
377 Posts
07-20-2009 7:18 PM |
For me, better File IO support, the ability to use FileAccess.ReadWrite is a must-have for me. Ability to use the Browsers network stack for OOB, it's a real pain to have .net authentication services suddenly ripped from you...ok then I just want to be able to continue using .net authentication in OOB, I don't care about the network stack. Also I don't want to see DataSet/DataTable, keep 'em out ;)
07-21-2009 2:10 PM |
Definitely better wcf support. Error handling in wcf is a nightmare. Could we also get a set of templates for setting up wcf services in different. The one that I run into all the time is the scenario where I need to transmit a large amount of data with a large object graph. Setting up the config file every time to do this correctly is very painful. Only occasionally do I get it right on the first try, and I usually spend the afternoon trying to figure out why the limits I set are not working correctly. Maybe when you create a silverlight enabled service object, it gives you the choice to step through a wizard to answer a few questions and customizes the config for you.....
mbrookfield
30 points
17 Posts
07-21-2009 4:37 PM |
Agree with tanmoy.r on Full 3D support. Perspective 3D??? Really??? More like tricky bit of horse dung. Ok guys, for the record, FLIPBOOK != true 3D rendering!!!!
Given that KIT 3D has been available on Codeplex for some time and Microsoft had plenty of time to implement REAL 3D rendering, I can't understand why it wasn't made available in SL3. Extremely disappointing to say the least. OK, so given that, for all you folks out there who require 3D tools for grownups: http://www.codeplex.com/Kit3D.
angelok
07-21-2009 4:47 PM |
Must-have's preventing my company's adoption of Silverlgiht: 1. Printing 2. FlowDocument Fixed Document & XPS Others: 3. Implement the rest of the WPF object model and controls 4. 3D
07-21-2009 4:50 PM |
i dont know what the new matrix projection does for 3D... but im pretty sure that Microsoft is working on a really smart an sleek 3D modelling language for Silverlight... why would they give us D3D/OpenGL-like language when they could do something LINQ'y or Lambda'tastic ?? i bet there will come 3D to SL soon, beating the crap out of Flash and whatever's out there..
jducoeur...
07-21-2009 5:11 PM |
I'm currently working on a serious web-based business app, built on an all-.NET back-end infrastructure, with users who are mostly IE-based. Our first client was written in Flash; right at the moment, I'm trying to understand whether Silverlight would be a good alternative going forward.
I don't know Silverlight at all well yet (I do know both Flash and .NET quite well), but from the previous comments, I suspect the big desires for us would be:
07-21-2009 5:35 PM |
Agree with you Montago but 3D class libraries already implemented in WPF. In fact, I used the WPF 3D documentation when working with Kit3D. While I think that going in another direction might provide some great opportunities to rethink 3D in Silverlight from the ground up, their best (and most expeditious) move would be to port what they already have from WPF over to Silverlight (WPF/E). Keeping the programming interfaces between the two as consistent as possible would be a big win for all of us.
07-21-2009 5:38 PM |
On the other hand, why the heck didn't they do that with SL3?????? (Rolls eyes ... smoke rising from head .... grrrrrr!!!!)
07-22-2009 5:25 PM |
Oooooo!!!! Hay I have an idea, how about making an install program that actually works!!!! It must be hard coming out with such an awesome application just to have it all undermined by the idiots that created the install. People are having issues with this going back to version 2 and you guys still haven’t corrected it. SOMEBODY WAKE THE HECK UP BACK THERE!!!!!!! On the outside chance that someone from the Borg actually see this and give a crap, be my guest and Google (yeah Google and not Bing. I need at least some components in my developer’s bag of tricks that actually works): “Silverlight 3” and (0x80091007 or “failed with error: 12180” or “does not have sufficient access rights).
Jeesh!!!
andrewjw
14 Posts
07-22-2009 7:37 PM |
My wish list:
Jac_P
1044 points
189 Posts
07-23-2009 2:26 AM |
The only thing I wish for is the following :
I create an application in Silverlight (or WPF), Create another one WPF (or Silverlight) project, add the files from the first project to the latter as link, and the solution work without any further changes.
Essentially, the code should be interoperable without any change at all.
Looking forward to MS to give a clean solution this time, unlike the previous versions of SL where everything breaks when you migrate an application from a previous version.
But based on the PREVIOUS BITTER experiences, this should not be expected. (Many things that used to work in SL2 stopped working in SL3).
They just come up with new features for the geeks and basically peanuts for the professionals.
Jac
07-23-2009 6:01 PM |
mbrookfield: Agree with you Montago but 3D class libraries already implemented in WPF. In fact, I used the WPF 3D documentation when working with Kit3D. While I think that going in another direction might provide some great opportunities to rethink 3D in Silverlight from the ground up, their best (and most expeditious) move would be to port what they already have from WPF over to Silverlight (WPF/E). Keeping the programming interfaces between the two as consistent as possible would be a big win for all of us.
You all point this out like the SL team could get this done overnight :)
WPF is platform-specific, and relies very heavily on its integration with DirectX. Silverlight is cross-platform. While hardware-accelerated 3D will probably be in there some day, it's not a simple or quick thing to do. It has to be prioritized with everything else.
Remember earlier I said the team needs use-cases, not bullet points? This is a prime example. The perspective transform satisfies the vast majority of use-cases that were presented to the team. If it doesn't satisfy yours, speak up, and be specific :)
Remember that SIlverlight 2 RTWd in October. Silverlight 3 has released just 9 months later. I don't know about you guys, but for 9 months of work, the release is pretty impressive to me.
07-24-2009 4:03 AM |
Psychlist1972:While I think that going in another direction might provide some great opportunities to rethink 3D in Silverlight from the ground up, their best (and most expeditious) move would be to port what they already have from WPF over to Silverlight (WPF/E).
I do not think it is wise to implement any WPF 3D features into Silverlight.
Most users, me included, jump onto the "3D" train to be able to develop games with it. Although WPF satisfies basic 3D needs it is not suited for game development whatsoever.A 3D Environment requires heavy work in both graphical visualization as well as pure floating point calculation speed.WPF models are extreme heavyweight objects to fit into the XAML presentation framework. Property access is severely slowed down due to a lot of boxing and unboxing arisen through the DependencyProperty environment.
Games of a descent size are not developed using XAML, they can't be, a game environment is far to dynamic to be statically declared.Therefor the whole XAML integration will fall on deaf ears. WPF allows you to create descent and really nice UI's, but thats it, no games.Sacrificing a huge amount of speed just to be able to declare my viewport in XAML is no viable solution for me and games need one thing above all ... Speed!This "small" overhead is not important when dealing with a handful of objects, but the more you have the worse it'll get.
Furthermore full 3D support requires a huge amount of new objects, which would bloat the runtime.
The best bet is to speed up overall rendering, which would allow third parties to create viable 3D environments as it is already happening, unfortunately current implementations fail to live up to expectations, due to slow rendering.You can also enable and disable GPU acceleration in Flash. If you enable it, everything you draw, no matter what or where or when, is done so by the GPU, as it should be in Silverlight. The BitmapCache was a nice idea, unfortunately it creates more problems than it solves.
No, WPF integration is a big no no ... please don't!Silverlight could bring games to the next generation, but not with half baked tools.Ever wondered why there are no 3D browser games like Quake or Unreal or Warcraft3 (and no, Quake Live is no browser game) ?Up till now there was no possibility, since Flash just isnt up for the task. Developing a complex system with Flash would take twice as much developers and a tenfold of time compared to using the .NET environmentnot to mention flash's lack of calculation speed, my lil' brother can multiply floats faster using his fingers.
For the first time there's a tool out there, which could be up to the task. Development tools for Silverlight are excellent, in fact the best this world has to offer, calculation speed is excellent, we have more than enough speed here.
So, whats the catch ?
To be completely neutral and to continue my sarcastic statements, my lil' brother can paint faster using pen and paper than silverlight currently does ;)So, that's the problem, there will be no 3D without better rendering performance, no matter how much we wish, no matter what will be included.
I hope the SL team reads this, for I'll drop some knowledge passed along to me from my university professor for artificial intelligence."Boy ... do or don't, theres no room for maybes here".
If you decide to include full 3D support please do it the correct way or else leave it be.
my 2 cents
07-24-2009 6:19 AM |
Krasshirsch:Ever wondered why there are no 3D browser games like Quake or Unreal or Warcraft3 (and no, Quake Live is no browser game) ?
Quake [Silverlight 3] -> http://www.innoveware.com/quakelight.html
Unreal'a'like [Shockwave] -> http://www2.rasterwerks.com/game/phosphor/beta1.asp
footnick
33 points
62 Posts
07-24-2009 6:52 AM |
Krasshirsch:Most users, me included, jump onto the "3D" train to be able to develop games with it. Although WPF satisfies basic 3D needs it is not suited for game development whatsoever.A 3D Environment requires heavy work in both graphical visualization as well as pure floating point calculation speed.
Hi Krasshirsch
I don't think you are correct to say that 3D is primarily a tool for developing games. I used to work for a company developing a 3D toolkit that supported 100's of applications in none games environment such as design, marketing, training, simulation, etc all of these were growing rapidly and many would benefit from 3D support in web client applications.
Floating point calculation speed or bloated silverlight engines shouldn't be an issue as I had assumed that any implementation would make use of graphics hardware rather than implementing 3D rendering directly. Given that Silverlight already apparently supports hardware graphics extending this to 3D shouldn't be a huge overhead.
I agree that Xaml isn't appropriate for defining larger data sets, but Silverlight isn't all about Xaml.
Nick
07-24-2009 7:38 AM |
I agree. We use 3D for medical visualization. Implying 3D is just for games simply doesn't take into account what business apps really do. I'd like an easy port, but would gladly rewrite using writeable bitmap in order to get the other benefits on silerlight (i.e. easier deployments, etc). The real deal-breakers for us are printing and flow/fixed document support and xps viewer. The bloated engine statement is putting the cart before the horse. We're not architecting yet, just offering requirements. The engineers will figure out how to implement later, and perhaps they can find ways to accomplish the task.
07-24-2009 10:15 AM |
Montago:Quake [Silverlight 3] -> http://www.innoveware.com/quakelight.html Unreal'a'like [Shockwave] -> http://www2.rasterwerks.com/game/phosphor/beta1.asp
I'm sorry, I should have mentioned state of the art 3D Games. 40 FPS, no lighting, 320x200 resolution for "Quake 1" on a 2,6 Ghz Quad Core is not really what I had in mind, especially when the original Requirements were 66 Mhz and a 1MB SVGA video card ..., so no ... it's a nice experiment but it doesn't count.
I can't say much for Shockwave, I dislike using it, but still ...
if you squeeze ten more of those "Wannabe" 3D games it's still nothing compared to the vast amount of client games.
07-24-2009 10:50 AM |
footnick:Hi Krasshirsch I don't think you are correct to say that 3D is primarily a tool for developing games.
I don't think you are correct to say that 3D is primarily a tool for developing games.
This is a very subjective view, I never said it is the only, but we could argue for days 'bout that ...
Nevertheless, if you want to venture beyond displaying a handful of static 3D entities and perhaps rotating a camera a bit a WPF like environment will just not do.
Even if your assumptions were true and 3D would be primarly used in design, marketing, training, simulation, many of these usages would operate in a game like environment.
You have user input, you generate data on top of this input and then you display the result.
Let's take the simple fact of displaying a car engine in all details using 3D visualization. We habe thousands of unique parts (screws, platings, pipes, ...) all tied together in this huge scene graph.You should be able to view those parts seperately, take them apart, put them together.
Let's take CAD systems, you add/remove and swap items all the time until you created whatever it was you wanted to create, you will not do this in XAMLyou will do this in code.
I'm not saying this is not possible using WPF, but it's just not suited for complex dynamic scenes with lot's of interactions.#
Sure, for the 3D piano with 5 Buttons and perhaps a handful of visual states, which is shipped as an example, it's enough, but c'mon ...
footnick:Floating point calculation speed or bloated silverlight engines shouldn't be an issue as I had assumed that any implementation would make use of graphics hardware rather than implementing 3D rendering directly.
This is because you appearently have no clue how a 3D engine operates, even rather simple well designed engines need a minimum of a couple dozen extra classes and structs. The more complex engines have hundreds of objects in addition to the DirectX and OpenGL Sdk's. There is a reason why the Ogre3D engine binaries are larger than the whole SL plugin itself.
footnick:I agree that Xaml isn't appropriate for defining larger data sets, but Silverlight isn't all about Xaml.
Unfortunately when it comes to visualization of any kind it is, the SL team will not add any objects to the visualization tree that cannot be integrated or merged with other obejcts using XAML. For XAML integration and interaction all objects must derive from DependencyObjects which unfortunately bloats them and makes them useless in large quantities.In 3D we leave the 1 - 10 world and enter the 1000 - 10000000 universe, even with more or less small and uncomplex objects, and no, a plane is not a complex object.
07-24-2009 11:38 AM |
Krasshirsch:This is because you appearently have no clue how a 3D engine operates
Not sure if that counts as 'abuse'
Suffice to say that we have an excellent 3D application that uses WPF to produce some really cool results and more importantly exceeds the original requirements in terms of both quality and performance. Our real world experience doesn't match your theoretical view point. Unfortunately the app isn't out to market yet so I can't demonstrate it to you.
If we wanted a games engine then we'd use one. What we want is a WPF level API to 3D hardware graphics within Silverlight.
07-24-2009 12:20 PM |
I'm sorry, I didn't want to offend anybody. My apologies if it sounded that way.
The point is: A WPF 3D Environment may suite you, but it doesn't suite the game sector at all. So here's the question, implement a new feature that helps one side a lot and leave the rest standing on the shores, or implement a feature were all can benefit from ?
Therefor I do not want to see any WPF like 3D implementations, I'd rather see GPU accelerated rendering and leave the hard part to third party companies.Kit3D is modeled 1:1 after it's WPF counterpart, even the namespaces are identical, you can nearly copy paste your 3D code and it'll work, it's just currently pretty useless, since there is no GPU accelerated drawing, so a handful of FPS @ most.
Balder is a 3D "game" engine build with SL 3, it's really nice and fast, since it doesn't use dependencyobjects, but it too is pretty useless since there is no GPU accelerated drawing yet.
You see, all those engines are already out there only the fuel is missing ;)
07-24-2009 1:11 PM |
LMAOROF!!!! Awesome Montago!!! Good job dude.
As for Silverlight and gaming my humble (heh) opinion would be that like Flash, and regardless of hardware acceleration, there will be little usefulness in Silverlight for such applications for some time. In my mind, the biggest mistake Macromedia made in marketing Flash was taking the easy way and allowing artists and designers dictate the market potential for such a tool. They got lazy and instead of creating a robust programming environment for developers to create something truly original and out of the box, they slapped in together as an afterthought and as a result Flash became mainly a cute little IDE for creating banner ads and maybe some rudimentary games. They missed the opportunity to create a real alternative to HTML based web interfaces; true business RIA. There were a few amazing demonstrations of what could actually be accomplished with such a tool such as Marketrac: http://marketrac.nyse.com/mt/, but by then it was too late. Flash had already been pigeonholed into a world of tunnel-visioned idiocy. Trying to make Silverlight into a game development platform would run the risk of doing the same. I see Silverlight as providing the tools to invent and create the next generation of user interface for web based business applications: RIA!!!!! Sure it’s fun to create game like environments for Silverlight, and indeed that may provide some of the forward vantage point that will lead to the aforementioned next gen web based GUI. However, I truly believe it’s better to let Silverlight become something greater.
As for "bloaty-ness" of introducing some or all of the WPF 3D engine, I guess I must be missing something because it seems a simple to me as referencing libraries, or not. If you want a 3D light, include that. If you want a full implementation, include that. Why should everyone’s ability to apply their ideas be preempted because some just don’t "get it"? Only on the backs of new tools can the next vistas become visible. As B. L. Whorf once said: "Language shapes the way we think, and determines what we can think about." In this case Silverlight is the current state of the language. Don't screw it up.
PS: There is a trend that XAML should be the end all for creating Silverlight applications. This would be a HUGE mistake. In enterprise level applications, much must be generated at runtime. For instance, the Entity Frameworks apparent insistence that a data object be defined in advanced ("strongly typed"), is just plain silly. (Back to Comp Sci 101 guys!!! Who the heck knows 100% of the time what your data structure is going to look like in advance???) This unfortunately renders the .NET RIA Services pretty much useless except in a few applications. It's like the 80-/20 rule ... in reverse. My point is that not everything should be restricted to a purely ZAML perspective.
07-24-2009 1:24 PM |
PS: Using Kit3D and some quaternion rotation math, I was able to render a set of up to 4000 cubes (6 polygons each), in a slow albeit reasonably useful manner. This, without hardware acceleration. Again, it was very slow; way to slow for gaming, but plenty useful enough to provide some amazing capabilities to a next gen web based GUI for serious business applications.
07-24-2009 1:42 PM |
mbrookfield:Flash had already been pigeonholed into a world of tunnel-visioned idiocy. Trying to make Silverlight into a game development platform would run the risk of doing the same.
Currently Silverlight's only purpose besides creating RIA's is flickering banner ads and pacman games just like flash, since there is no capability to support something visually more demanding. Why not try to skip this pathetic part and move into the real entertainment market ?And I don't mean ports of 1000000 year old games that barely run on todays high end machine's.
mbrookfield:As for "bloaty-ness" of introducing some or all of the WPF 3D engine, I guess I must be missing something because it seems a simple to me as referencing libraries, or not.
07-24-2009 2:05 PM |
Krasshirsch:Currently Silverlight's only purpose besides creating RIA's is flickering banner ads and pacman games just like flash, since there is no capability to support something visually more demanding.
What a shame if that were true. Luckily its not. Period!! The things I'm working on fly directly in the face of that belief. I think you'll find a lot of other voices out there opposing your view.
Krasshirsch:If it just were that simple ..., but true, if the SL team could completely isolate the WPF 3D part it would be great, inside WPF it's not.
Sigh (rolls eyes): http://msdn.microsoft.com/en-us/library/system.windows.media.media3d.aspx
Small or large parts of which could be broken out as needed.
07-24-2009 2:13 PM |
MBrookfield:Sigh (rolls eyes)
MBrookfield:The things I'm working on fly directly in the face of that belief.
07-24-2009 2:18 PM |
I just tried this one http://www2.rasterwerks.com/game/phosphor/beta1.asp Really nice, it would be really awesome to be able to create something like this for the web without having to rely on archaic scripting languages and debuggers.
07-24-2009 2:40 PM |
Krasshirsch:if it can be completely isolated, great
I think that’s what MS should be focusing on. Maybe this dialog will contribute to that happening.
Krasshirsch:Like what ?
Without going into details as its proprietary, the ability of the user to say:
Click on a data grid entry representing some actual physical or representation object and that object being located and identified within a 3D space wherein the space and or object might be flipped in three axis’s, 360 degrees. Likewise, another object might be located within the given space clicking on which would bring up details of that object in a data grid. The objects and construction of the space are completely data driven and created at runtime ... for instance.
07-24-2009 5:24 PM |
one of my friends and his team made this game in 3 weeks:
http://www.amazingmonk.com/
powered by the Unity3D engine: http://unity3d.com/gallery/
its amazing what it can do, and how simple it is... there is a demo of a FarCry island... really amazing !
http://unity3d.com/gallery/
http://unity3d.com/gallery/live-demos/tropical-paradise <-- farcry in the browser...
danielkward
7 Posts
07-24-2009 5:42 PM |
PRINTING!
07-25-2009 8:28 AM |
Krasshirsch:it would be really awesome to be able to create something like this for the web without having to rely on archaic scripting languages and debuggers.
Several years ago Microsoft joined teams with Dassault Systems [the creator of CATIA Software] to develop a standard and format called "3DXML". The nice thing about it was you could create any object you wanted using the CATIA 3D CAD Software, and then export it to a 3D XML file. In other words, if you know CAD [like I do], then you had the ability to create these amazing 3D worlds with little effort, and then export it to a 3D XML file that can be viewed in their player online.
When I first heard of Silverlight, I thought, okay now things are making sense. It thought, this was the reason why Microsoft helped Dassault come up with a 3D XML standard that could be viewed on the web. Back in 2006, I was experimenting with CATIA to make these amazing 3D objects with little effort. So this was why I was a bit disappointed with the 3D in Silverlight [because of my assumptions]. Don't get me wrong, Silverlight is great and I'm excited about its future. It's just I was hoping for something similiar to my experiences with 3D XML.
click http://www.3ds.com/company/3d-experiences/ to see some of the 3D experiences and games.
Gallery at http://www.3ds.com/products/3dvia/3d-xml/gallery/
If you haven't seen 3D XML yet, it's worth taking the time to download their player and check it out. Their player is a lot larger, so it takes a little longer to download. The 3D format and specification is also there.
http://www.3ds.com/products/3dvia/3d-xml/overview/#vid1
CarlBrui...
07-25-2009 8:32 AM |
Interesting.
I dabble with Flash every now and then. I'm a .NET developer and have started using Silverlight more and more.
The killer features that Silverlight lacks is;
Someone mentioned a killer feature, imagine this;
Whilst Wii controllers have been done, it would be seriously cool if users could use their Xbox360 Natal controller to manipulate 3D in Silverlight. Consider the user base of Xbox360 users, that would give MS the user numbers to finally boast that it's taking chunks out of Flash.
Carl[::TILL THE END::]
07-25-2009 10:25 AM |
rshelby:you could create any object you wanted using the CATIA 3D CAD Software, and then export it to a 3D XML file
XAML supports true 3D just fine. What is required are tird party tools that translate various formats back and forth. ZAM3D for instance not only provides a basic 3D creation environment but imports from 3DS Max, DXF, etc. When you have what you want, ZAM3D kicks out pure XDAML that can be used directly in Blend. What is missing is the support in Silverlight to consume that output. Part of it can be consumed by employing KIT3D, but only a small subset of functionality. Also, Sketchup does a pretty good job of it as well and seems to support more formats. This functionality has been available since Silverlight/Blend 2.
ZAM3D:
http://www.erain.com/Products/ZAM3D/DefaultPDC.asp
Sketchup:
http://sketchup.google.com/
KIT3D:
http://www.codeplex.com/Kit3D
Hope this helps :)
07-25-2009 10:46 AM |
CarlBruiners:Whilst Wii controllers have been done
For those who may not know what Carl is talking about:
http://www.youtube.com/watch?v=QgKCrGvShZs
CarlBruiners:3D data format,
Xaml is quickly become one of the many standard formats available out there. See ZAM3D by Electric Rain for instance. (Since you’re a flash developer you may remember the name from Swift3D)
07-25-2009 12:04 PM |
quick fixes for serious regression errors
http://silverlight.net/forums/t/103424.aspx
marcoenxuto
07-27-2009 9:07 AM |
Good afternoon gentlemen,
the Only think i would like to see, is a very important thing which i havent seen no one report, i guess... The connection to sql databases, like *.vb os *.cs does.
ADO.Net
just like that, get datasource from SQL Server, and workaround of that.
Thats all!
Regards
staceywi...
3 Posts
07-27-2009 12:32 PM |
"Creating a 3d cube by simply taking a rectangle and setting a thickness / depth property. Or doing the same thing to a circle in order to make a 3d cylinder."
1) Totally. Also be able to extrude text (embos, bevel, etc) and have shadow effect respect the thickness and a light position. Is this 2.5D? Even lift the 2.5D elements off a Z-axis. Also be able to easily place and "stick" images on cubes and balls so they rotate with the cube and not have to program the perspective for each image seperatly- it would just work as expected. Then be able to set the view-port to *inside the cube to create easy carosel effect with images stuck inside the cube (instead of outside). Being able to "wrap" a stack panel around a ball or cylinder would be interesting also to create circular queue effects without having to deal with end-point logic.
2) Print. One cool idea would be to expose a system PrintCanvas. You can just add controls to it like a normal canvas using code or xaml strings - then have a Print() method after adding is complete. As async is the model, you could also just keep adding to it as you get data (say from a server or sql report) and it will page and print as data is produced. The user could control this behavior from the silverlight Properties dialog. Default would just queue xps documents with some kind of default DOS limits. Policy would allow user to select which apps can queue and/or print to which printers. Being xps based, user could then right-click on queued documents and view or print or email or delete, etc.
3) In the same way, policy could be used to control external devices such as sound and webcams. Maybe in a VirtualPC fashion. Where local SL devices are virtual and map to real local or network devices. That way, even devices become "composable" and pipeline-able. For example, a speaker Line device may be mapped to a remote internet mic, that can be piped into your mixer control to add background music which could be piped into live sound feed control of a new channel, that could then be Tee'd off to a recording device, etc. Ultimately allow all kinds of ways to hook things up.
the.boarder
07-27-2009 5:51 PM |
jredfield
20 points
10 Posts
07-29-2009 11:16 AM |
As a SL developer, I just want better error messaging. Xaml parsing errors are still difficult to debug sometimes. The features mentioned above are certainly important, but faster development time, due to better tools and error messaging, debugging, etc. is really important. The faster we can crank out those totally cool SL apps, the faster it will catch on universally.
Other than that, PRINTING for sure!
Meanwhile, one item on the SL3 wish list that was done, and we don't have to worry about, is considerably better font rendering. SL3 is now a legit LOB tool!!
puckoff7337
70 points
53 Posts
07-29-2009 11:28 PM |
Most important for me is to make it possible to do more minor designer changes.... a display that is also movable would be nice (similar to what we have in C#) I know there is expression blends but it would be really nice to have this possible w/o leaving VS2008
Also more closely matching classes like the sockets class to C# would be nice, the sockets class now still leaves a little to be desired.
Better MDI support would also be a huge plus (though the Code Plex toolkit does a good job making up for this)
Some other good ones I saw others mention were:
1) Host html inside of Silverlight control
2) Full support of all mouse buttons
3) Full Linux support
4) Better integration of the various components that are now in the CodePlex silverlight toolkit natively
07-30-2009 5:30 AM |
Our focus is on property marketing and the features that we would most like to see include ....
1. 3D rendering with HW acceleration (we had been excpecting this at SL3 - in our view the simple perspective offered at SL3 is just a gimmick whilst proper 3D offers real functionality for business applications of the future - hopefully the simple perspective was just a stop gap because 3D missed the cut?)
2. XPS and printing
3. Support for Mobile devices.
4. Mic support and speech recognition (ideally on mobile devices)
Given that adding new features will increase the size of Silverlight we would prefer some granularity with Silverlight applications including optional modules as required (such as the Silverlight Toolkit). We would be happy to disable functionality for platforms that don't support optional modules.
Whilst some of our applications are used by the general public and require the sort of public trust and system penetration offered by Flash and hopefully Silverlight in the future, other applications could support an installation process which could perhaps allow applications to control security issues such as access to trusted native DLL's, more control of file open and save, and easier management of local storage. Currently we use WPF for some of our applications but we would much prefer to manage a single development environment with wider platform support such Silverlight.
chaitany...
250 points
56 Posts
07-30-2009 12:50 PM |
I wish to see some more killer features, to win against FLASH, like
1. More timeline animation control functions
2. Printing/Webcam/Audio Input support
3. Lesser file size (more easy to load)
usamakhan86
07-31-2009 12:04 AM |
Map Contorl for gis
Silverlight 4 Wishlist