[whatwg] on codecs in a 'video' tag.

[Gervase Markham]

Ian Hickson wrote:
> On Thu, 29 Mar 2007, Gervase Markham wrote:
>> Dave Singer wrote:
>>> No, writing it into the HTML specification is not a commercial reason.
>> Assuming you have commercial reasons for supporting HTML 5 (which I 
>> suspect you do, otherwise you wouldn't be here) then having Ogg 
>> specified gives you a commercial reason to support it.
>>
>> If that's not a commercial reason, then what would be a commercial 
>> reason? If everyone else implemented it?
> 
> A _commercial_ reason would be "our customers demand it". Customers in 
> this case would be users, and users would demand it if a big video site 
> started using Ogg Theora.

Which they will only do if browsers support it. Chicken and egg. So this 
again boils down to my characterisation of Apple's position as "if 
everyone else supports it, we will". Which seems a lot more like "free 
formats only if unavoidable" rather than "we have a strong commitment to 
interoperability".

I look forward to hearing Dave's suggestion for how we promote 
interoperability among <video> implementations.

>> Why don't we all just go away and implement what we think is best for 
>> HTML 5, and then put a spec together after the fact? Then we wouldn't be 
>> forcing any issues, and there would be no "fiat". But we all know how 
>> well this approach to standards works.
> 
> Actually that's pretty much exactly what we're doing with HTML5.

HTML 5 seems to be a mixture of the two approaches. For pragmatic 
reasons we are (holding our noses, in some cases) standardising existing 
things, often because of the market dominance of the existing 
implementation. But in many other cases, we are speccing things first, 
trying them out and then updating based on feedback. Which I would say 
is the right approach.

>>> No matter what the spec. says, if broad support became a reality, then 
>>> yes, it may be in our commercial interests.
>> So Apple's strategy is to wait and see what codec everyone else 
>> implements, and then implement that one?
> 
> Everyone's strategy should be to implement what they need to implement. 
> Implementing random stuff without good reason ends up simply bloating your 
> product.

Straw man. That's not what I said. Apple's "if broad support became a 
reality, then yes, it may be in our commercial interests" seems to 
translate pretty safely to "if everyone else supports it, we will". If 
not, how are they different?

>>> anyone *can* implement the codecs we implement;  they may choose not 
>>> to, for commercial reasons (e.g. they don't like the license)
>> Oh c'mon, that's a ridiculous definition of the word "can". How exactly 
>> "can" the KDE project implement a codec in Konqueror which requires 
>> royalties? How "can" the Mozilla project implement such a codec without 
>> removing the redistributability of Firefox?
> 
> In both cases, by negotiating appropriate licenses with the IP owners.

Given than giving an appropriate license compatible with free software 
would basically mean the IP owners saying "yes, we are happy to get no 
more revenue from this, ever", I consider that somewhat unlikely, and 
not really falling within a definition of "can". But perhaps you know 
different. Is Google planning to buy out the MPEG patent portfolio and 
put it under a royalty-free license? That would be a wonderful 
contribution to interoperable video standards.

Gerv

(Again, as in this entire thread, speaking only for myself.)