- From: Philipp Serafin <phil127@gmail.com>
- Date: Wed, 5 Nov 2008 21:47:36 +0100
On Wed, Nov 5, 2008 at 4:00 PM, Leons Petrazickis <leons.petrazickis at gmail.com> wrote: > It matters in the sense that web browsers would have to implement both > approaches for backwards compatibility. > This depends what you mean when talking about "implementing" a tag. Browsers already load all tags and attributes they encounter into the page DOM today , regardless if they "know" them or not. This is also the behavior that HTML5 demands, if I'm not mistaken. Interactive elements and "special" elements like "script", "head", "table", etc do need further special treatment, but so far browsers seem to do little more with semantic elements than apply a default style to them. And sice CSS selectors have already widespread support as well, this behavior would be trivially repeatable for new tags by web authors themselves. So there is not that much of an implementation burden at the moment. >Standards that have tried to make changes like that -- XHTML2 comes to >mind -- have not been as successful as HTML4 [...] We can't really make any statements about how successful XHTML2 would be on the public web. It's not yet a recommendation (though this would probably not change much) and no browser implemented it, so there was never opportunity to find out. >Adoption and usability are the twin goals -- not purity or >consistency. Of course I fully agree that purity must not the only goal itself. The W3C definitely went overboard with this in their latest proposals. However, sometimes it does look as if the WHATWG takes it to the opposite extreme and treat purity and consistency as explicit non-goals. If consistency hampers usabillity, usabillity is of course to be preferred. However, I think, often the two are interrelated instead of opposing goals. After all, each inconsistency is a new rule that implementors have to implement and authors have to learn. This makes the language harder to understand and actually slows down adaption. Look at XML for the opposite story. Many people have frowned at the mass of XML-based languages, formats and specifications that popped up. But then again, why DID so many people create XML formats so quickly? Because the concept of "nested tags with attributes" is really simple to understand. Of course HTML5 has to cope with really horrible legacy content and many inconsistencies are just there. But we should still try not to introduce more. And yes, I do believe <reference class="abbreviation" ...> would be easier, because then web authors wouldn't need to remember which semantic content can be described by tags and which needs custom classes. But anyway, this discussion is moot, since many of those tags can't be changed due to backwards compatibility. Maybe someone should clarify though how agents are supposed to use those semantic tags, especially semantic, user-defined classes. What actual benefit do I have if I use class="heading" instead of class="style15"?
Received on Wednesday, 5 November 2008 12:47:36 UTC