Re: (Repeat) Decision: C.4 (Predefined entities)

At 11:19 AM 11/11/96, Joe English wrote:
>bosak@atlantic-83.eng.sun.com (Jon Bosak) wrote:

>There is no problem with predefined entities.  There is a big problem
>with canonizing a set of predefined entities that *cannot be redefined
>by an XML document or application*.

  I agree with this, and would strongly support overridable predefined entities.

  Of course, that means that a DTD-less instance with redefined entities
will produce different results. I expect people could learn to live with
this, since those who create XML DTDs will have to learn how, and that
learning will lilkely include the details about overriding. I am sure that
the issue of DTD-free processing is what drove the committee to the
un-redefinability condition.

>Why?  There are no predefined entity names in SGML, let alone
>reserved, non-overrideable ones.  It's neither obstinate
>nor perverse for an author to expect a system that calls itself
>SGML-compatible to follow the rules of SGML.

Nor for an author providing tagged data over the WWW to expect that it will
be compatible with other tagged data formats over the web.

These two reasonable expectations conflict. We have resolved these
conflicts in SGMLs favor (DTD syntax) and in HTML's favor (entities) and in
new ways altogether (end tags). I agree about overriding, but the issue of
expectations is very complex.

>Besides that, considering the way the Major Players have treated
>the issue of putting mathematics on the Web (Netscape:
>"There does not appear to be a market for it, but if you
>really want it you can always use JavaScript"; Microsoft:
>"There does not appear to be a market for it, but if you
>really want it you can always write a VBX"; W3C: "We're
>working on the spec.  Trust us.  We're working on the spec")
>only the most optimistic Web users expect to see ω
>recognized by a non-fringe browser any time soon, and even
>they would be surprised to see thinks like ℑ, ∅,
>∏, ♣, ⊥, and ∇ suddenly becoming keywords.

Well, that overlooks the fact that the Adobe characters are _easy to
implement_ on Unix, Mac, and PC. They won't do it because it's a good idea,
or because people need it, but because the standard says to, and it's
almost no extra work to implement it. I won't hold my breath waiting for
integrals or summations though. The mathemeticians are unlikely to accept
any notation other than TeX anyway, and for reasons that are actually
pretty good.

>If the XML general entity namespace is to be at the mercy
>of what Netscape decides to implement six months from now,
>it would be safest for authors to avoid declaring their own
>entities altogether.

The suggestion was that the XML entity namespace agree with the W3C HTML
entity namespace when the standards officially issue (in some timeframe in
the future). That is why the proposal is in advance of current practice.

   -- David

RE delenda est.
I am not a number. I am an undefined character.
David Durand              dgd@cs.bu.edu  \  david@dynamicDiagrams.com
Boston University Computer Science        \  Sr. Analyst
http://www.cs.bu.edu/students/grads/dgd/   \  Dynamic Diagrams
--------------------------------------------\  http://dynamicDiagrams.com/
MAPA: mapping for the WWW                    \__________________________