Lists Home |
Date Index |
> Yup. The havoc that CDATA sections and external entities play with tree
> data models is becoming obvious,
I recognized this when DOM L2 was moving through draft and a long thread on
www-dom about how TreeWalker interacted with entity ref nodes illustrated how
brittle everything had become.
It was also a long thread on www-dom that made me understand how mad the idea
of live node iterators was.
I must say, though, was that all it took was the first, cursory reading of the
L2 namespaces section to understand how deeply broken it was. No thread
> and the notion that they HAVE to be resolved
> or thrown away before the InfoSet-based specs see them is becoming widely
> accepted. That seemed like "cheating" in DOM Level 1 days.
> In retrospect, the 80/20 solution seems obvious, but recall that two years
> ago, the notion that XML 1.0 was more complex than it really ought to be was
> quasi-heretical even on this list.
I (and Mike Olson) learned started implementing DOM almost immediately after
learning XML back in early '88. It's dangerous for me to try stretching my
memory back that far, but I think most of the choices made in DOM seemed
perfectly sensible back then. So I should probably not blame anyone for those
L1 decisions. However, I agree with you that a re-start would be nice.
I should make the point, though, that Megginson's choices in SAX 1.0 do seem
remarkably prescient. It came after DOM L1, but only a little bit afterwards.
Also, IIRC, Megginson also waved off the SML crowd, so it's interesting that
SAX got things so rightly.
Uche Ogbuji Fourthought, Inc.
http://uche.ogbuji.net http://4Suite.org http://fourthought.com
Track chair, XML/Web Services One Boston: http://www.xmlconference.com/
The many heads of XML modeling - http://adtmag.com/article.asp?id=6393
Will XML live up to its promise? - http://www-106.ibm.com/developerworks/xml/li