Weekend Discussion

Friday March 24th, 2000

Time for another weekend discussion. The question this weekend:

What do you feel is an appropriate level of integration between a browser and an Operating System? And at the other end of the spectrum, what are your feelings about Mozilla's cross-platform approach?

#70 IE errata

by ttfkam

Monday March 27th, 2000 2:33 PM

You are replying to this message

The comparison of XUL with DHTML behaviors gave me a nice chuckle. Thanks. Work kinda sucked this morning. No seriously, DHTML behaviors is a good implementation of scripted HTML. This brings up some of my gripes with HTML: presentation and content in one piece.

An HTML "div" tag means symantically a page division, but it also has a layout meaning with states that it should leave an open carriage return after the closing tag. On the other hand, XUL defines the component only. Yes, you can take away that layout behavior from the "div" tag, but then you're left with a open XML tag container... just like XUL only you had to work to get it there. There is a strict demarcation between the logic (usually found in a separate, linked script), presentation (CSS) and the component itself (XUL).

This separation is commonly referred to as an MVC (Model-View-Controller) architecture. In other words, modification of one component does not usually affect the functionality/presence of the other two. DHTML behaviors tend to blur this distinction. MVC is kinda like object-oriented programming. It starts with folks bemoaning the apparent loss of their preferred mode of development/use, follows with an excited "Oh! I get it!", and finishes with a resigned "I can't believe I used to do it the other way."

I agree with you about Visual Studio and its depth, but a GUI development platform based on DOMs and stylesheets is not that terrible to create. XPCOM development tools will probably lag as you say, but using/reusing existing objects will be a relative breeze.

IE4's XML parsers and DOM support was a shoddy exercise to say the least. After all, why do you think that Microsoft changed the API so dramatically from v4 to v5? The Java parser sucked arse then and still sucks arse now.

With regards to instantiating a DOM via JavaScript, you are correct up to a point. Instantiation of an external DOM is a DOM level 2 item which the coders have not tackled en masse yet. The funny thing about the Mozilla group is that they seem to be trying to finish one spec and implement it fully before working in earnest on its successor. But you are correct, Mozilla is behind in this area.

Schemas, XSLT, XPath? Have you even glanced in the general direction of the W3C draft for Schemas? Check out the MS implementation and compare with the spec. Yes, I know, the Microsoft.XMLDOM came out early on in the spec, but do you see any work being done on it in the new technology previews?

XSLT is in fact in development and folks are coding away on a workable implementation to be included into the main branch. It isn't there yet, but it's not like they're ignoring it either.

Honestly I don't know the status of XPath support. I'll concede this one unless anyone steps up to correct me.