3 ms·

I know I'm taking to a wall for most of you, but anyone with an open mind and more than 3 years of experience knows: The reason we moved to H.264 is because it

by radley 16y ago

I know I'm taking to a wall for most of you, but anyone with an open mind and more than 3 years of experience knows: The reason we moved to H.264 is because it's a standard in the video world. It unified online video really well - everyone uses it and we were all tired of the fragmentation. Flash is also a standard, particularly for multimedia. You can't have 99.7% penetration and not be a standard. With all due respect to web standards: a standard is something that everyone uses. I think this is the best and most ironic line in the article: "HTML5 is an emerging standard and anyone dealing with it has to expect change constantly."

pohl 16y ago
With all due respect to web standards: a standard is something that everyone uses. It depends upon your priorities. Others would say that a standard is something that is completely & unambiguously specified so that interoperable implementations can be made. The terms under which such implementations are created is yet another concern. We probably need 3 orthogonal axes to plot various standards: ubiquity, specification, licensing terms. A "defacto standard" would score high in ubiquity but might have no specification (let alone one of quality) and might have bad or undefined licensing terms. Other things may have excellent specifications but niche existance. Something like H.264 has a specification, high penetration, but licensing terms that require certain participants to pay. Something like WebM is shooting for very liberal licensing terms, they are currently a niche but are angling for penetration — although I'm not sure if they have/intend-to-have a specification upon which interoperable implementations can be based. Edit: there are also those who seems to put the greatest importance upon who controls the specification. They wouldn't accept PostScript as a standard, for example, because it isn't controlled by a standards bureaucracy —despite the clarity of its specification. I personally believe that unless we can openly acknowledge all of these components of the concept of "standardization" we'll have a lot of heat & no light when we're discussing it. My personal leanings: I value a complete & ambiguous spec above all else, and loathe de facto standards that thrive in absence of one. I don't feel standards bodies are a must-have, but acknowledge that they can be nice. I also acknowledge that standards bodies can give the illusion of community control when in fact they provide no such guarantee. I'd rather not have to pay to implement a spec, but in some problem domains (such as codecs) I can understand why some would want to be rewarded for their work.
- radley 16y ago
  I say tomato, you say Nix v. Hedden... =)

by radley 16y ago

pohl 16y ago

With all due respect to web standards: a standard is something that everyone uses. It depends upon your priorities. Others would say that a standard is something that is completely & unambiguously specified so that interoperable implementations can be made. The terms under which such implementations are created is yet another concern. We probably need 3 orthogonal axes to plot various standards: ubiquity, specification, licensing terms. A "defacto standard" would score high in ubiquity but might have no specification (let alone one of quality) and might have bad or undefined licensing terms. Other things may have excellent specifications but niche existance. Something like H.264 has a specification, high penetration, but licensing terms that require certain participants to pay. Something like WebM is shooting for very liberal licensing terms, they are currently a niche but are angling for penetration — although I'm not sure if they have/intend-to-have a specification upon which interoperable implementations can be based. Edit: there are also those who seems to put the greatest importance upon who controls the specification. They wouldn't accept PostScript as a standard, for example, because it isn't controlled by a standards bureaucracy —despite the clarity of its specification. I personally believe that unless we can openly acknowledge all of these components of the concept of "standardization" we'll have a lot of heat & no light when we're discussing it. My personal leanings: I value a complete & ambiguous spec above all else, and loathe de facto standards that thrive in absence of one. I don't feel standards bodies are a must-have, but acknowledge that they can be nice. I also acknowledge that standards bodies can give the illusion of community control when in fact they provide no such guarantee. I'd rather not have to pay to implement a spec, but in some problem domains (such as codecs) I can understand why some would want to be rewarded for their work.

radley 16y ago
I say tomato, you say Nix v. Hedden... =)