Paul & Eleanor wrote:
> My thoughts on allowing names or Ids was still along the lines of
> playing with the idea of allowing some users to generate something
> very simple maybe using a generic XML editor so that possibly
> someone with a laptop could type in yesterdays survey while in
> the field and send it in for processing (or add to the system be
> importing into some package).
I have trouble seeing how a "generic XML editor" could even be close, in
ease of use, to a rather simple "CaveXML editor". Such a program could
be tailored to the format used in YOUR notes, and could me a much
smaller, simpler program than an XML editor. The only part that involves
real work is reading and writing the XML, and there are libraries for
that. I would expect that entering data tag by tag with an XML editor
would be "clunky" to say the least, regardless of how much we simplified
the format.
> The problem there is that a certain
> flexibility to allow nearly raw caveXML probably is a challenge
> to not provide 'too much' flexiblity so that a more standardized
> form doesn't get broken.
That's the problem all right!
>
>> How do those editors handle attributes of type IDREF? If there is no
>> matching ID, the file is not valid XML. Do "XML editors" always produce
>> valid XML?
>
> I'd be interested in how much work these editors put into understanding
> IDREFs also. I'd be surprised if they do anything at all,
That would be my guess too.
> thus the need
> to
> allow just shots with names, to be later converted into some kind
> of canonical form. I'm all for some canonical form, but maybe some
> flexibility to allow a simplier "pre-canonical" form.
The risk, of course, is that some day someone is SURE to send you just
the "pre-canonical" file, without the context information needed to do
the conversion. Actually I sort of assume that (at least until CaveXML
tools become universally available) most people will enter data using
whatever tools they use now. In a sense, that makes the tool specific
format a "pre-canonical" form. Converting halfway from there doesn't
make sense to me.
Ralph Hartley
This archive was generated by hypermail 2b30 : Mon Apr 02 2001 - 18:00:01 CEST