ctceismc : the export widget isn't very farfetched. But the "import HTML" isn't going to happen soon. For that, you need to parse HTML elements, and overlay them in front of C2's canvas, because C2, is, basically, a canvas only library (even when some plugins are writing HTML elements on top of that). A plugin maker can try that, but it's a huge job !

bjadams : what do you mean by "printing" ?

It's from LostGarden

http://www.lostgarden.com/2007/05/dancs-miraculously-flexible-game.html

He made two tilesets at the time. One vector (XAML and the like), and in the comments you can get PNG render.

It's a CC attribution, if I remember well...

Savvy001: the thing you are asking for isn't doable in JS/HTML only. You need to send a request to the server (Ajax, CGI or something else), to store your data. The server need to know what to do with that. So you also need to code for the server, something which isn't tractable by Construct.

Edited the first post to give the url for the capx.

Ashley : yes, I was thinking about the memory consumption. One possibility is to have an option for the user to delete the array used to store the images once the coloring is done. It would allow you, for example to colorize some Sprites at the beginning, and then dump what you don't need. The minus is on the fact that you can't force the GC to delete it immediatly, but at least you are sure that you are going to get memory back in a few cycle...

Edit : I have another last trick to use, and if it works, we can have image coloring while keeping low memory use <img src="smileys/smiley2.gif" border="0" align="middle" />

For the moment, and for what I saw on blogs about the media queries, the iPad3 report a "virtual resolution" of 1024x768. No WebgL for UIWebView for the moment, though...

Since I'm working on the realtime effect plugins for Sprite (using WebGL), I was asking myself if there wasn't a possibility or an hack to do that without WebGL. Javascript is slow, parsing Bitmap matrices is slow, but there's a trick I finally implemented in a Behavior, from there : http://www.playmycode.com/blog/2011/06/realtime-image-tinting-on-html5-canvas/

And here's the demo, using Construct 2 :

<img src="http://dl.dropbox.com/u/1412774/RecolorDemo/demo.png" border="0">

http://dl.dropbox.com/u/1412774/RecolorDemo/index.html

The idea is to load the image one time, parse its bitmap array, and split it in 4 canvas (so yes, it can be slow when loading the image at the beginning, and it's memory consuming when you load large images). After that, it's just an operation of pasting those canvas with the global effect flag of the HTML5 canvas set at the right value. Starting from there's, it's near realtime if you use coloring effects. (As the demo shows, if you try to do that all the time, when checking the 'realtime' box and moving the sliders, it's going to be slower).

In a few days, the behavior will be available for everybody !

StephenC : before buying the next version, did you already sold some creation made with C2 ? You know, the important thing is the marketing, not the "complexity" of the game.

To add something to Ashley's defense, when you're head in the code all day, sometime coding something more "easy" or "light" can give you some relief... <img src="smileys/smiley2.gif" border="0" align="middle" />

AarongamerX : thanks for the design doc ! HAving those is very helpfull, it makes you do less non-useful code, and go directly to the right way of doing things.

PixelMonkey : Ashley is right. In HTML5, images are stored as matrices, one per color (R,G,B), and Alpha. There's no concept of "indexed paletted image" (i.E. one byte value translating to one entry in the palette).

On that demo, the coder used image from an old Amiga and then DOS code he made. Each image is sent as JSON to the browser, with an index for each pixel translating to an entry in the palette.