Squeak as Metaverse reminds me of something concrete...

Richard A. O'Keefe ok at atlas.otago.ac.nz
Mon Jul 24 00:33:43 UTC 2000


Henrik Gedenryd <Henrik.Gedenryd at lucs.lu.se> wrote:
	In the light of the many statements of what XML is supposed to be, perhaps
	you could paraphrase another member of this list:
	
	XML is a great step forward for the HTML community.
	
Since design is of great importance in OOP, and since XML raises a number
of extremely interesting design questions, perhaps I may be indulged if I
answer this.

Perhaps the single biggest contribution XML makes to the HTML community is
that it isn't _quite_ backwards compatible with HTML, so maybe, just MAYBE,
the companies that make HTML editors will finally take the trouble to make
products that generate syntactically valid XHTML.  They have *NOT* for
the most part bothered in the slightest with generating syntactically valid
HTML.

There is a major difference between HTML and XML.  HTML does have rather
fuzzy layout semantics associated with it.  No two browsers display the
same page the same way, but there is a family resemblance.  There is just
enough layout semantics in HTML to do useful things.  But XML has no
layout semantics associated with it whatsoever.  You _cannot_ usefully
browse XML without stylesheets.

So maybe, just MAYBE, the outfits that bring us browsers will finally see
fit to finish implementing CSS2.  On the other hand, maybe they'll decide
to toss CSS2 on the junk heap without ever giving it a fair go, and go
straight to XSL.  On the gripping hand, if they had a hard time with CSS2,
when can we expect to get working XSL processors in browsers (other than
James Clark's, which I hear he has stopped maintaining...)

And when you add in XBase, Namespaces, XPath, XLink, XInclude, XSchema,
XSL, ... you have to say "XML?  what flavour?"

XML has the potential to take the Web a big step towards high quality
type; it also has the potential to fragment the web.  (See the current
disagreement over how to interpret Namespaces.)  The industry would have
to be short-sighted and stupid to fumble this ball, which is why I fear
it may happen.

The only browser I have access to which understands MathML does such a
terrible job of math layout that I prefer GIFs.





More information about the Squeak-dev mailing list