[sc34wg3] fundamental types

Lars Marius Garshol sc34wg3@isotopicmaps.org
20 Nov 2003 11:39:48 +0100

* Patrick Durusau
| (This language also appears at "4.2 The fundamental types" so if
| changed here needs to be changed there as well.)

I know. The terms and definitions section is actually completely
generated by the XSLT stylesheet. This way we conform to ISO style
without sacrificing too much readability and without risking
inconsistencies. :-)
| What caught my eye initially was the phrase: "independent of the
| context they appear in" which struck me as odd, considering that the
| fundamental types are string, set and null. (see 4.2 The fundamental
| types)
| Then I thought, oh, it means that these value types never change, that
| is a string is always a string, wherever it appears, etc.

For strings and null it does, but not for sets, unfortunately. 

Other distinguishing features are:

  types that are not information item types

  types for which equality means: no differences between the values at
  all (also applies to locators, unfortunately)
| Well, then why do I need the explanation that it has "no topic
| map-defined semantics"?

Originally we had it as "no semantics independent of the context ...",
but we realized that they did, but that they didn't have any topic map
meaning outside of context. (Unfortunately, that also applies to
| For that matter, what is gained by gathering these three value types
| up into a class of some sort?

When I first read this question I thought "Obviously because we use
the term all over the place!" but searching the document shows that
that's not the case, so that leaves me wondering the same thing. :-)

Graham and I have discussed this and now think the term and the
definition of it should go, but that the section should keep its
heading and content otherwise.
| Suggestion:
| Move the terms string, set and null out of 4.2 (where they are mixed
| with equality tests) into the glossary.

We could do that, but it would hurt readability. Now these terms are
introduced in the logical reading order for the document. If we put
them in the glossary and don't have them anywhere else they get buried
with all the other equally important stuff in the glossary and no
longer have any form of thematic unity, so readability goes down.
| I think these are common enough to not need any defense as
| "fundamental" value types and simply defining them without further
| explanation is quite sufficient.

We agree.
| First, Unicode 4.0 is out so perhaps we need to change that in "2
| Normative References" to: "The Unicode Consortium. The Unicode
| Standard, Version 4.0.0, defined by: The Unicode Standard, Version
| 4.0 (Boston, MA, Addison-Wesley, 2003. ISBN 0-321-18578-1" (odd I
| know but that is the citation form given at:
| http://www.unicode.org/versions/Unicode4.0.0/)

Good point. Will do.
| Second question: The technical report for normalization (#15) is
| cited in Unicode 4.0.0 but does not actually appear in the volume.
| Should we make some more specific reference to the report itself?

You are right. We should. I wasn't aware of this, but checking now I
see that you are right.
| Note that the Unicode Standard will continue to evolve so would it be
| better to simply point to the Unicode website?

Yeah, that's actually an issue. Should we use a dated reference, or an
undated one? Do we want to require people to support Unicode 4.0, or
do we want them to support "the latest version of Unicode"?

Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
GSM: +47 98 21 55 50                  <URL: http://www.garshol.priv.no >