4 ms·

This post exhibits everything that is wrong with standards bodies.

alexchamberlain 14y ago
Why?
batista 14y ago
Nothing of the sort actually. On the contrary, it shows how by breaking things into smaller modules instead of a monolithic one, they got things moving along faster.
- pbhjpbhj 14y ago
  >they got things moving along faster // We can't really be sure of that can we. What we can be sure of is that there's no way to summarise the CSS standards compliance of a browser or to indicate concisely the browser level requirements of a website for a human. This I fear makes designing for the web a case of breaking down and designing for lots of individual UA that meet a list of hundreds of different point versions for different abilities instead of being able to simply design to a standard HTML5+CSS3 and readily see that the top X browsers will support (with some issues of course) the elements of those standards. Instead of one current standard there are now hundreds of standards?!
  - batista 14y ago
    >We can't really be sure of that can we. No, but we can be sure that conceptually it's far better than the monolithic practice, which has been shown in the past to slow down the standardising process. >What we can be sure of is that there's no way to summarise the CSS standards compliance of a browser or to indicate concisely the browser level requirements of a website for a human. That was always true. I mean, we might have had a "way to summarise the CSS standard compliance", as in "this browser is CSS 2 compatible", but we never had a 100% compatible browser for any previous version. I doubt we still have a full CSS2 compatible browser engine, so so much for summarizing. Now, it could be easier: we could have a browser that is 100% compatible with A, B, C, D CSS modules and lacks in X, Y, Z. It's not like developers care about the overall compatibility. What they do care is "can I use this or that feature?". To say "this engine is 90% CSS3 compatible doesn't answer that". Only 100% compatibility would answer those kinds of questions ("yes, you can use ANY feature"). So, they way to summarise compliance will be, as it is the practice anyway, with some list of CSS features and Yes/No tick marks. And for some things, with the knowledge that for that module the engine has 100% compliance (easier to achieve than the ellusive 100% compliance over ALL modules).
    - pbhjpbhj 14y ago
      >Now, it could be easier: we could have a browser that is 100% compatible with A, B, C, D CSS modules and lacks in X, Y, Z. // In CSS2 there are about 500 or so different things that you'd need to list (http://www.w3.org/TR/CSS2/indexlist.html http://www.w3.org/TR/CSS2/indexlist.html). I dispute that people care about 100% compatibility. >So, they way to summarise compliance will be [...] some list of CSS features and Yes/No tick marks. // I think you misunderstand the meaning of "summarise", "this engine is 90% CSS3 compatible" is a summary. Quirksmodes tables - http://www.quirksmode.org/css/contents.html http://www.quirksmode.org/css/contents.html - whilst useful during development aren't really something you can put before a user as an equivalent to an "HTML5" badge.

Why?

Nothing of the sort actually. On the contrary, it shows how by breaking things into smaller modules instead of a monolithic one, they got things moving along faster.

pbhjpbhj 14y ago
>they got things moving along faster // We can't really be sure of that can we. What we can be sure of is that there's no way to summarise the CSS standards compliance of a browser or to indicate concisely the browser level requirements of a website for a human. This I fear makes designing for the web a case of breaking down and designing for lots of individual UA that meet a list of hundreds of different point versions for different abilities instead of being able to simply design to a standard HTML5+CSS3 and readily see that the top X browsers will support (with some issues of course) the elements of those standards. Instead of one current standard there are now hundreds of standards?!
- batista 14y ago
  >We can't really be sure of that can we. No, but we can be sure that conceptually it's far better than the monolithic practice, which has been shown in the past to slow down the standardising process. >What we can be sure of is that there's no way to summarise the CSS standards compliance of a browser or to indicate concisely the browser level requirements of a website for a human. That was always true. I mean, we might have had a "way to summarise the CSS standard compliance", as in "this browser is CSS 2 compatible", but we never had a 100% compatible browser for any previous version. I doubt we still have a full CSS2 compatible browser engine, so so much for summarizing. Now, it could be easier: we could have a browser that is 100% compatible with A, B, C, D CSS modules and lacks in X, Y, Z. It's not like developers care about the overall compatibility. What they do care is "can I use this or that feature?". To say "this engine is 90% CSS3 compatible doesn't answer that". Only 100% compatibility would answer those kinds of questions ("yes, you can use ANY feature"). So, they way to summarise compliance will be, as it is the practice anyway, with some list of CSS features and Yes/No tick marks. And for some things, with the knowledge that for that module the engine has 100% compliance (easier to achieve than the ellusive 100% compliance over ALL modules).
  - pbhjpbhj 14y ago
    >Now, it could be easier: we could have a browser that is 100% compatible with A, B, C, D CSS modules and lacks in X, Y, Z. // In CSS2 there are about 500 or so different things that you'd need to list (http://www.w3.org/TR/CSS2/indexlist.html http://www.w3.org/TR/CSS2/indexlist.html). I dispute that people care about 100% compatibility. >So, they way to summarise compliance will be [...] some list of CSS features and Yes/No tick marks. // I think you misunderstand the meaning of "summarise", "this engine is 90% CSS3 compatible" is a summary. Quirksmodes tables - http://www.quirksmode.org/css/contents.html http://www.quirksmode.org/css/contents.html - whilst useful during development aren't really something you can put before a user as an equivalent to an "HTML5" badge.