[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Public WebGL] Typed arrays and network data



Related to WebGL typed arrays, there was an interesting discussion
about binary data at public-script-coord@w3.org and
es-discuss@mozilla.org last year; it would be great these discussions
will be converged:

http://lists.w3.org/Archives/Public/public-script-coord/2009OctDec/0093.html

-- snip --
I'd rather not propose APIs like that in the initial version (though I
think eventually we may want a way to copy sequences of 16-, 32-bit or
64-bit values swapping from network byte order to host byte order or
vice versa to make it practical to interpret popular binary formats.
Maciej

   http://lists.w3.org/Archives/Public/public-script-coord/2009OctDec/0105.html
---------

In addition to this, Web IDL very prematurely defines the array types
for array host objects like org.w3c.dom.OctetArray,
org.w3c.dom.FloatArray, etc. represented as octet[], float[].
Eventually es5 typed arrays should be associated with Web IDL arrays
in some form, and reflected back to Web IDL Java binding, too. The
outcome *could* be something like this:

interface FileReader {

  void readAsBinaryArray(in Blob fileBlob);
  readonly attribute octet[] binaryResult;

};

interface WebGLRenderingContext {

  void uniform1fv(WebGLUniformLocation location, float[] v)
raises(DOMException);
  void uniform1iv(WebGLUniformLocation location, long[] v) raises(DOMException);

}:

Regards,

 - Shiki

On Thu, Feb 4, 2010 at 10:17 AM, Arun Ranganathan <arun@mozilla.com> wrote:
> On 2/3/10 3:51 PM, Vladimir Vukicevic wrote:
>>
>> On 2/3/2010 3:44 PM, Oliver Hunt wrote:
>>>
>>> On Feb 3, 2010, at 3:40 PM, Kenneth Russell wrote:
>>>
>>>> On Wed, Feb 3, 2010 at 3:37 PM, Oliver Hunt<oliver@apple.com>  wrote:
>>>>>
>>>>> Things like this should really go to the webapps or possibly the main
>>>>> whatwg mailing list as this api really shouldn't be webgl specific, and more
>>>>> importantly we need input from the wider html5 community for these APIs.  I
>>>>> don't really like the way in which we're moving further and further from
>>>>> purely webgl related APIs and trying to specify APIs that really need input
>>>>> from non-webgl people.
>>>>
>>>> On what mailing list is the typed array proposal primarily being
>>>> discussed?
>>>
>>> Typed arrays i think are probably being discussed on the es5-discuss
>>> mailing list, networking related APIs probably belong on whatwg or some such
>>> -- it seems that this binary data concept should really be tied into the
>>> Blob interface which is part of the FileAPI proposal
>>
>> Blob, ideally, can just become ArrayBuffer if there's agreement on
>> ArrayBuffer and the rest of the typed array bits.
>
> By "just become" do you mean Blob should have an ArrayBuffer property?  In
> general, I like the idea of linking the two.
>
>> es-discuss was suggested by Arun and Brendan, but I don't think there's
>> actually much interest in the typed array proposal there (there isn't active
>> opposition, just sort of a "that's interesting" silence).
>
> Well, es-discuss is the place where we can have *this* discussion, for one.
>  Some of the folks on this listserv are likely to also read es-discuss (I
> know olliej and vlad do).  I picked the ECMAScript listserv for this because
> I really think that we need new language-specific primitives.
>
> You an subscribe/unsubscribe here:
> https://mail.mozilla.org/listinfo/es-discuss
>
> I know that *yet another listserv* is a daunting thought, but our discussion
> does influence the rest of the platform.
>
> Vlad's "strawperson" is here:
>
> http://wiki.ecmascript.org/doku.php?id=strawman:typed_arrays
>
>> So I'll look to put it elsewhere, or just develop it within the webgl
>> group as a separate standalone specification.  That's not ideal, because as
>> Oliver says that it's not really a purely webgl thing, but we have a need
>> for getting it done in finite time...
>
> We should perhaps lead with our own discussion, and then ping the es-discuss
> listserv.   FWIW, we can also ask the Chair of TC-39 (the group basically
> overseeing next versions of the ECMAScript language) to put this on the
> March meeting agenda.  Essentially, I'd like to see a successor edition of
> ECMAScript specify this.  If that isn't an ideal way to go, we should try
> something like W3C's WebApps.
>
> -- A*
>>
>>    - Vlad
>>
>> -----------------------------------------------------------
>> You are currently subscribe to public_webgl@khronos.org.
>> To unsubscribe, send an email to majordomo@khronos.org with
>> the following command in the body of your email:
>>
>
> -----------------------------------------------------------
> You are currently subscribe to public_webgl@khronos.org.
> To unsubscribe, send an email to majordomo@khronos.org with
> the following command in the body of your email:
>
>

-----------------------------------------------------------
You are currently subscribe to public_webgl@khronos.org.
To unsubscribe, send an email to majordomo@khronos.org with
the following command in the body of your email: