Cart
[Home] [By Thread] [By Date] [Recent Entries]
>Interesting point that Tim raises. The real-world >interoperability problems that I hear about seem to >mainly involve encodings, whitespace handling, >entities of various types, namespaces, and of >course valid/wellformed. A good number of them are >"RTFM" issues not conformance per se, e.g. >whitespace handling, and the numerous ways that >people manage to get confused by namespace declarations >and prefixes. Right. It's not the syntax, it's the usage thereof, things outside XML 1.0, and the things that fall outside the 80% mark. >My suspicion is the same as Don's -- people who don't >have interoperability problems are probably not pushing >the XML 1.0 envelope, either by accident or design. Well, the 80/20 rule of standards conformance comes partly from implementation cost, and partly from user expectations. If users aren't using some features, they're less likely to be strenuously tested. I think standards are useful for setting the "high tide" mark. Communities (in the large) naturally tend to adopt something less (the larger the community, the more likely that mediocrity be the norm).
|
