Obviously, I'm fond of making things explicit. The difference in length
between a schema explicitly stating its min/max occurs and one that
doesn't
is trivial, and frankly, humans don't normally look at a schema *unless*
they're looking to it as a source of documentation, in which case,
why not
be explicit about what the schema is requiring?
On Sep 9, 2005, at 3:02 PM, Rebecca S. Guenther wrote:
> We recently had some messages about the use of minOccurs=1 and
> maxOccurs=1, which was changed in version 1.1 from what was in 1.0.
> It was
> pointed out that semantically including minOccurs=1 and maxOccurs=1
> was
> the same as leaving them out since they are the defaults in XML
> schema. (The same goes for the data type "string"; right now
> xs:string is
> used in most cases and is explicitly in the PREMIS schemas, but
> since it's
> a default in XML schema, it could also be deleted. We just
> discovered one
> place where it was left out that it should have been there for
> consistency's sake.)
>
> This really goes to the question of style and there are various
> opinions
> about this. Is it better to have a compact schema that only
> includes what
> is needed and eliminates redundancies or to have a more verbose schema
> that spells out the defaults explicitly? Either way is correct.
>
> As we go to version 1.1, before we finalize the drafts that are out
> there,
> it would be good if we came to consensus on this. Any opinions?
>
> Rebecca
>
Jerome McDonough, Asst. Professor
Graduate School of Library & Information Science
University of Illinois, Urbana-Champaign
501 E. Daniel Street, Room 202
Champaign, IL 61820
(217) 244-5916
[log in to unmask]
|