- From: Chris Wilson <cwilso@MICROSOFT.com>
- Date: Thu, 30 Sep 1999 09:29:14 -0700
- To: "'jelks@jelks.nu'" <jelks@jelks.nu>, Daniel.Glazman@der.edf.fr
- Cc: www-style@w3.org
Jelks Cabaniss [mailto:jelks@jelks.nu] wrote: >One can go to the bathroom, and one can go for a drive. Aggregating the two >because they both involve "going" would over time make for a rather unpleasant >car. ... The appearance of bathroom examples in any discussion is a sure sign the discussion is not headed in a positive direction. >What's wrong with using a CAS (or whatever you want to call it) as a *separate* >linking mechanism? Is it because bringing in a new MIME type would be too much >trouble? Sneaking behaviors into the back door (via CSS) seems to be a quick >hack to avoid some kind of problem like this. But what are the long-term >ramifications of doing this? And most importantly, what does behavior have to >with style -- other than that you *could* share the same declarative syntax? There are quite a few problems with using BECSS as a separate linking mechanism. The first, of course, is that there is NO standard provision for linking them into the document. We would, in essence, end up having to use the LINK REL=STYLESHEET and embedded STYLE elements to link these in - or come up with some new elements, which is a difficult thing to do at this stage with any hope of interoperability. Creating a new MIME type is the least of those worries. In addition, implementation experience with how designers and authors use CSS teaches us that web sites are commonly developed with separate content and presentation - and part of that presentation is interactivity, and designers would like to design their presentation at one pass, not two. The approach we've taken does not exclude those who wish these to be separate (such as yourself) from making them so. >Yes, right now we LINK in a stylesheet into (for example) an HTML document. >BECSS would be an even *further* level of indirection: HTCs would be linked into >the CSS that's linked into the HTML doc... The recent focus is on packaging, >where each of these stand alone and unlinked, but "wrapped" in another resource, >yet BECSS seems to be going in the opposite direction. If you choose to use HTCs, that's true - but you'll also note that HTCs are fairly heavy-weight, and intended for implementing presentational components, which are really best if they can be componentized and kept in independent modules. HTCs are totally document-independent, reusable modules - the BECSS stuff is not necessarily document-independent at all (although it can be - just like a CSS stylesheet can be, but often isn't). The event handling part of BECSS is lighter-weight, and is embedded directly. Come on, you really can't use the argument that you want style and interactivity in separate documents, and then say we're making the wrong decision by componentizing. >It sure seems that the CSS part of BECSS is a short term hack that sounds OK on >the surface, but will prove as bad an idea as FONT was to markup. Now that's a cry of "Wolf! Wolf!" if I ever heard one. -Chris Wilson
Received on Thursday, 30 September 1999 13:14:56 UTC