[
Lists Home |
Date Index |
Thread Index
]
From: "Dare Obasanjo" <dareo@microsoft.com>
> What purpose does posting such comments serve besides acting as
> an advertisement of your biases?
Without going into the merits, I guess the writer was referring
to the added requirements made by an early MS XML parser
notoriously including that namespaces had to be #FIXED. [1]
So people had to alter their DTDs to work with MS XML.
Still, I don't see that only Microsoft developers should take it as a salutory
plea not to embrace and [restrict|extend] (i.e. to make the default operation
of a parser less or more than the spec says, perhaps out of entirely non-sinister
motives), because it is a good thing for all developers to keep in mind.
(For example, IIRC James Clark's expat could also be criticized on
similar grounds: it did not handle parameter entities: however this was
clearly documented and expat was exploratory. The cumulative
effect of expat and MSXML was to cripple the effectiveness of DTDs:
I think this initial unreliability held XML innovation back a year or two
--and created the RELAX NG-sized hole that WXS sat in--because it
prevented the basic software engineering requirement of testable data integrity:-- validation. It is ironic that at the same time as unit testing was becoming
popular, the leading XML parsers did not provide reliable validation. )
Cheers
Rick Jelliffe
[1] http://www.xml.com/pub/a/1999/11/parser/?page=2
|
