To XSLT or not to XSLT

IlyaM on 2003-03-27T09:45:59

Just found by accident quite interesting post on usefulness of XSLT.

On related note I recall somebody who used XSLT to generate (sic!) CSV files from XML.


CSV from XML

ajt on 2003-03-27T11:29:20

While one may find the idea of generating a CSV file from an XML file strange, it's often one of the examples to follow "Hello World" that most XSLT books cover in their introductory chapter....

As you say an interesting article. I like the point about costly development tools, I think XML Spy is very powerful, but it isn't cheap. You can edit XML/XSLT in a simple text editor, but you really need a proper UTF-8/XML aware one to do any thing serious without a lot of pain. You could argue that expensive incompatible tools killed SGML, one hopes the same fate does not lie ahead for XML...

Re:CSV from XML

ziggy on 2003-03-27T20:16:54

You could argue that expensive incompatible tools killed SGML, one hopes the same fate does not lie ahead for XML...
An expensive and ambiguous formal definition that was overly hyperlinked (on paper), difficult to read, and impossible for a human to parse was more responsible for killing SGML. If that wasn't enough, the near complete inability for anyone to create a fully compliant SGML parser that interoperated with all of the other fully compliant SGML parsers was the final nail in the coffin.

(ISTR that the SGML industry produced exactly two SGML parsers that closely approximated the SGML standard: James Clark's nsgmls and another proprietary one by some SGML tools vendor. They handled some edge cases differently, yet both of them strongly asserted that their SGML parser was correct according to the SGML standard, and the other implementation didn't implement some aspect properly. For all practical purposes, SGML was so complex that only James could implement it, and some other standards (like DSSSL) were so much more complex that he implemented a working subset and lost interest.)