[dsdl-comment] Re: changes to DSRL ?

Dave Pawson wrote:
> 2008/9/23 Rick Jelliffe <rjelliffe@allette.com.au>:
>
>
>> I think this is a matter of expectations too.
>> The DTDs days of home-made enumerations for yes/no values are well and truly
>> done: dead and buried. I don't recall the last time I saw a schema (either
>> XSD or RELAX NG) that didn't automatically opt for using xs:boolean, with
>> the equivalence of 1 and true. I think David's surprise at 1 and 0 being
>> available *may* perhaps a sign that he hasn't used XSD datatypes that much.
>>
>
> So what does this say about your expectations of readers Rick?
> "you must be familiar with XSD prior to reading this spec"
>
Yes. That is the point of using other standards, you don't have to spell
things out. Or, at least, if the spec
says "we use XSD boolean datatype" then to find out what it is you look
at the XSD spec.

It is a myth that using other standards necessarily simplifies things.
> I'm with David, specs should attempt to be models of clarity with no
> preconceptions?
>
I think they should be clear, but not tutorial nor redundant. (There
was a rule of thumb that standards should be pitched
at a subject expert: in the case of the XML spec, for example, this
means a developer of parsers. In the OOXML
BRM there was discussion about whether to remove the schema fragments
from the text: I won proposals to
include the full schemas into the text, and to make the full schema the
normative one, but I lost my proposal to remove
the now-informative fragments, because reviewers found them useful: so
I certainly admit that extra material is not
always superfluous.

Patrick Durusau's plans for the ODF spec are instructive here: rather
than debating whether a standard should
be minimalist or tutorial, have both through multi-publishing!

We have a couple of strong model here for having minimalist specs with
annotations, from within the markup world;
Charles Goldfarb's SGML Handbook and Tim Bray's Annotated XML website.

The simplest thing would be to have a Wiki with each page having three
parts:
1) a section with a clause from the draft or standard, edit locked 2) a
section with interpretation/tutorial comments,
open to the public, and 3) a section with suggestions for improvements,
open to the public.

Actually, I am not sure that Wikis are the best kind of collaborative
mechanism: we use PageSeeder here for our
Lending Industry XML Initiative specs and it has better support for
multiple kinds of groups and views. I don't know what
is happening at the W3C on this at the moment, for their inhouse
development. I guess a Topic Maps-based
system would be best, on the "eat your own dogfood" principle.

By the way, IBM released a discussion paper and a policy that they would
withdraw from all standards bodies that were
not open according to their definition. Obviously they mean ISO, but
actually that involves ANSI and almost all national
bodies, so it will be interesting to see if they really give up on the
Linux ABI, SQL and so on: seems dubious to me.

Cheers
Rick

--
DSDL comments
To unsubscribe, please send a message with the
command  "unsubscribe" to dsdl-comment-request@dsdl.org
(mailto:dsdl-comment-request@dsdl.org?Subject=unsubscribe)