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

Re: [Public WebGL] ETC texture compression.

>From the app developer's point of view, there are really three separate
situations with extensions.

1) Sometimes they are used to add features to OpenGL that lots of
hardware supports - but which the standard has yet to encompass.
2) Sometimes they are used when one particular implementation has a sexy
new feature that the others don't - and which is therefore unlikely to
make it into the spec.
3) Sometimes they are used to cover an annoying situation where some
feature is widely supported but which is somehow legally encumbered and
cannot be a part of the standard.

In the case of (1), you can probably just assume the existence of the
feature - perhaps some ancient or odd-ball hardware doesn't support it -
but that's life.   I prefer that old hardware doesn't advertise the
feature if using it causes a fallback on software emulation of the
feature - so this is OK.

In the case of (2), it's rare that I'll use the extension.  Even though
the feature might provide some massive speedup, it's likely that it does
that on platforms that are already on the cutting edge of performance
where I don't NEED a massive speedup.   About the only time I'll do
something here is when nVidia have implemented some feature one way and
ATI implemented more or less the same feature in a slightly different
way.  Since both major brands implement more or less the same thing, I
can use the extension mechanism to trigger special case code to cover
the subtle differences.

In the case of (3), you swear and curse at the stupidity of software
patents and probably use the feature without a fallback.  S3TC texture
compression is a good example of that.

Oliver Hunt wrote:
> Which is basically why I opposed them some time ago (I still do, but I
> have no new arguments and merely repeating them is a waste of
> everyones time)
> --Oliver
> On Aug 24, 2010, at 11:28 AM, Cedric Vivier wrote:
>> On Wed, Aug 25, 2010 at 02:24, Gregg Tavares (wrk) <gman@google.com
>> <mailto:gman@google.com>> wrote:
>>     Sure pages often have to check for a few incompatibilities in
>>     browsers but those are arguably failures of the browser makers to
>>     agree rather than by design such as with OpenGL extensions.
>> Yeah but by this argument why we do allow and plan for WebGL
>> extensions then? By definition they won't be available on all
>> implementations either (thus needing special care by its consumers
>> not to break on non-supported platforms...).
>> Regards,

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