On Thu, 08 Mar 2001 17:57:14 -0700, Paul & Eleanor wrote:
>> > role=bookie, role=sketcher and role=book all look fine for many humans,
>> I gather that you are asking that the field be a enumeration type
>> instead of a string?
>John wondered if I was thinking enumeration. Yeah, something
>like that, but looking around at
>http://www.w3.org/TR/1998/REC-xml-19980210.html#dt-enumerated
>I can't see where in an enumeration you could specify a fixed set
>of values but also allow anything else, so this type of cross link
>either needs to get into IDREF, XPOINTER, or outside the XML
>DTD specify that certain values would mean special things.
But if you do that, you're basically right back where you started: unable to query the dataset for information. If
the types were more controlled (this could be done in an internationalization section for enums), you could
ask the dataset how many times person X was on a survey trip or the number of times X person did various
roles. This is the kind of information cave managers (ie owners like NPS, NFS, BLM, etc) like to know.
Some of the internalization could be done like a resource file. I think that would work through XLink since it
wouldn't fit into a DTD and you don't want a copy of it in each CaveXML file. It might look something like this:
<roles>
<tape>
<english string="tape"/>
<french string=""/>
<spanish string=""/>
</tape>
<fore_sight>
<english string="fore_sight"/>
</fore_sight>
</roles>
or some such format. The point being, enumerations can be handled in over languages but it will take some
extra work.
Allowing users to just type in PCDATA is, IMHO, allowing too much freedom, especially in certain sections. I
think the roles, data items, and data units are some of the sections that need more control.
What is the rule for data validation in programs? Something like 50% of the code (and probably 75% of the
effort) is dedicated to error checking and shielding the data from the user.
- Richard Knapp
_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com
This archive was generated by hypermail 2b30 : Mon Apr 02 2001 - 18:00:01 CEST