HTML5 hackers wanted

mmp

United States2130 Posts

November 22 2011 05:14 GMT

I made a post in the Big Programming Thread but I'll post here too:

On November 22 2011 14:09 mmp wrote:
Are there any Javascript/Canvas/HTML5 hackers out there? I've been working on a framework for Canvas 2D Context for a while now and things are starting to come together (it's about 95% tested and 90% documented but there are some known holes, so be gentle).

The project is hosted at: http://gitorious.org/meta2d/core/trees/master/

If your browser is awesome (tested in FF7, Chrome14 -- Opera won't work right now, I assume IE will fail) you can try a demo of HTMLMouseEvent triggers here: http://mikemp.mit.edu/meta2d/demo/mouse.html

Unlike some other frameworks out there, this one is aimed at scalability of interactive applications without obfuscating the Context2D API. So you can port your existing canvas code over easily and use the extras that this framework provides to improve performance as needed.

It's a powerful utility if you want to write an HTML5 game, among other applications.

I'd appreciate some feedback & if you're interested in getting involved or have any questions feel free to pm me.

Gummy

United States2180 Posts

November 22 2011 05:22 GMT

This is pretty cool. I don't know how the browser handles the rendering/event detection of elements vs. drawn canvas shapes. iirc, you have to create the bounding boxes yourself, and do an expensive user-code loop of the mouse's current location against all the drawn shapes. Is your project relevant to this issue?

mmp

United States2130 Posts

November 22 2011 05:29 GMT

On November 22 2011 14:22 Gummy wrote:
This is pretty cool. I don't know how the browser handles the rendering/event detection of elements vs. drawn canvas shapes. iirc, you have to create the bounding boxes yourself, and do an expensive user-code loop of the mouse's current location against all the drawn shapes. Is your project relevant to this issue?

Yes, the project addresses this performance issue (among others).

When you define an entity's repaint behavior, you can draw on a canvas with a normal-looking Context2D, apply whatever transformations you want, and then pass that same Context over to the mouse collision detector for, say, alpha-masking.

It does mouse collisions in O(logN) expected runtime for N entities. I can run that demo with 1000 shapes on my netbook before getting significant lag.

Hidden_MotiveS

Canada2562 Posts

November 22 2011 17:23 GMT

So does the mouseevent trigger use any html5?

mmp

United States2130 Posts

November 23 2011 06:41 GMT

On November 23 2011 02:23 Hidden_MotiveS wrote:
So does the mouseevent trigger use any html5?

The canvas is the HTML5. Normal HTMLMouseEvents are propagated through entities inside the canvas as though they were elements -- just set onevent to your function(Event) callback.

Enderbantoo

United States465 Posts

November 23 2011 08:33 GMT

hey that's exactly what i'm doing. Im writing a canvas workflow program: http://www.enderbantoo.github.com/LONI-1

and i experienced all of what gummy was talking about, expensive bounding functions client side

Enderbantoo

United States465 Posts

November 23 2011 08:36 GMT

On November 23 2011 02:23 Hidden_MotiveS wrote:
So does the mouseevent trigger use any html5?

much more simple

canvas.onmousedown = function(e);

handle eah different clicks as e.button

Hidden_MotiveS

Canada2562 Posts

November 24 2011 02:42 GMT

On November 23 2011 15:41 mmp wrote:

Show nested quote +

The canvas is the HTML5. Normal HTMLMouseEvents are propagated through entities inside the canvas as though they were elements -- just set onevent to your function(Event) callback.

Ah I see now. Thanks.

Please or register to reply.

HTML5 hackers wanted

Completed

Ongoing

Upcoming