RE: [xml-dev] How to handle "newline" characters in an XML parser.

["Len Bullard" <cbullard@hiwaay.net>]

All true.  And not the problem at hand.

If you are renting a flat from someone and it doesn't have hot water in the
bathroom and you are expected to heat it yourself, do you expect to read
that in the ad for that apartment or to be told that on making an offer?

There may not be a law for that.  The truth in advertising problem has been
with us for a long time.

If on the other hand, I get a specification for a proper subset of SGML, I
expect it to conform to all of the rules of SGML for those features which it
applies.  I don't expect an XML parser ad to say that.  I have to know it.
If I get a piece of software for a proper subset of XML, I expect it to
conform to XML; that is, I expect all XML software to be able to process it.
If I buy software and it says it is an XML parser but does not process all
of the features available for XML, I'd call that false advertising but I may
want to check.  That is why there are legal marks for such systems.

False advertising is different from the problem of not having a means to
declare a proper subset except for natural language, and also, no
conformance testing.  We don't usually explain the inefficiencies of XML
when we sell it to customers, do we?  Do we expect others to do that?

What *we* have to be careful of is widespread pronouncements that such a
subset would have no value to some subset of XML users.  That way lies the
death of innovation as if XML were the final word on markup technology and
its application to real world problems.  It isn't just as SGML wasn't.
Otherwise, XML and its constraints become the constants of inefficiency
applied to any and all applications on the WWW.  Because inefficiency is a
cost figured locally, that would be madness.

len


From: Michael Kay [mailto:mike@saxonica.com] 

> He would definitely want to control the distribution and 
> contracts, squeeze down the network a bit, yes. 

And make sure his users are absolutely clear that they are not using XML,
will not be able to use any XML tools with their data, and can expect short
shrift from suppliers of XML tools if they complain.

Sadly, however, this will do no good. We'll still see the messages on this
and other lists: "I need to generate XML that doesn't use
minimized-empty-tags because my client can't accept them".