[whatwg] Endianness of typed arrays

[Charles Pritchard]

On 3/28/2012 12:57 AM, Boris Zbarsky wrote:
> On 3/28/12 12:49 AM, Robert O'Callahan wrote:
>> On Wed, Mar 28, 2012 at 5:33 PM, Charles Pritchard<chuck at jumis.com>  
>> wrote:
>>
>>> It does seem like a bigger "warning" in the spec may benefit 
>>> developers.
>>> A warning and an example of how to check for big-endian results.
>>>
>>
>> Asking developers to write extra code paths for users that don't 
>> exist is
>> futile. Even if you could persuade them to do it, since there's no 
>> way to
>> test those code paths, most of them will get it wrong.
>
> Indeed.  And note that developers don't read specs.  They just do 
> whatever tech evangelists tell them too.  And tech evangelists are 
> telling them to use typed arrays + XHR to load non-byte data of 
> various sorts "more efficiently".  I can guarantee that several 
> hundred people were present at the multiple SXSW talks where I heard 
> this promoted from the podium, and that very very few of them, if any, 
> have any plans to read the spec.

We've been here, too.

Most developers work with multiple programming languages. They read 
manuals, they read blogs, they try things out, and then they often code 
more in other languages than in JS.

Yes, many of them do not know the ins-and-outs of GL. It's true, they 
don't. But those that endeavor to make  quality games and demos do try 
to consume as much as they can.

I don't want to be in this kind of conversation with you nor Robert. 
Your statements of fact are fine. Your claim that "developers ... just 
do whatever" is toxic. It's not healthy for readers of the mailing list, 
it's not healthy for your relationship with developers. You and Robert 
are veteran developers of a software with hundreds of millions of users. 
That puts you both in a position of authority. When developers lash out 
as you (as they, or I, are sometimes prone to do), that's a shame, but 
it's not particularly harmful. When you or Robert disparage developers 
as a class, that is harmful.

We've seen the arguments; I agree that testing endianness sucks; testing 
WebGL is outright frightful, both APIs are error-prone, and many 
developers are unaware of their nuances. Many tech evangelists exploring 
HTML5 game development are still learning the ropes.

-Charles