Lists Home |
Date Index |
Been there. Note, for example,
the problems of subsetting, billion laughs,
recursion, and so on. One says, we can
solve these in the code. Another says,
declarative means can be used. It seems
that the SGML Declaration functionality
or something like it just won't go away.
As we go, we seem to rediscover more and
more of SGML. I doubt we are done with
the core XML specification but I do think
we must be very careful because there are
definitely opposing views about what should
be included moving forward. Will the goal
of reduced complexity in the core lead to
more complexity in the applications?
Markup is bigger than SGML though. Markup just
doesn't have a definition stronger than
"text with a namespace and reserved
characters" unless one tosses in structure
and structure leads to the problems. That
doesn't mean it isn't useful; just that
it starts the path toward externally declared
schemata and that toward 'types'. But
otherwise, markup is only a namespace
above delimited ASCII, arguably the most
used form of text-based interchange.
All of these, including the infosets, are
demonstrably useful in proper contexts. What
we should be strict about is proper contexts.
I sometimes despair that the often sensible
advice to "dare to do less" for each specification
can in aggregate, lead to a loss of rigor nd
that to ambiguity and complexity.
There just ain't no free lunch.
From: Simon St.Laurent [mailto:firstname.lastname@example.org]
email@example.com (Bullard, Claude L (Len)) writes:
>The "essence"? Disapprovingly? No. I
>simply don't like to see the clarity of
>the XML 1.0 specification muddied by attempts
>to use XML as a brand name for platforms.
That's a lot of why I'm talking less and less about XML and more and
more about markup.
I've found that it's a much freer generic universe, one where developers
can focus on the work once again without concern for all the stories
they've heard about the XML brand. It makes it much easier to get past
various flavors of propaganda.
Mike quoted Martin Gudgin:
>My objective is very much like Tim's -- to maximize the value of the
>XML "brand name". That requires careful consideration of what is of
As much as I like Martin, his refining of the "essence" corresponds
alarmingly well with my understanding of "dilution". Strange world