[sc34wg3] Almost arbitrary markup in resourceData

Kal Ahmed sc34wg3@isotopicmaps.org
11 Nov 2003 22:25:54 +0000


On Tue, 2003-11-11 at 20:01, Robert Barta wrote:
> On Mon, Nov 10, 2003 at 08:07:55PM +0100, Lars Marius Garshol wrote:
> > OK? If so, what does it mean? Is it OK for <scope> to appear here? For
> > <topicMap/>? For <resourceData>?
> 
> My first reaction to this would be to use the super-annihilator on
> this person, deport his family to Alpha-centaurii and erase every
> record of their existence. Everything additional could be regarded as
> extreme.
> 
> :-)
> 
> > This is clearly OK, however:
> > 
> >   <resourceData>XTM is <xhtml:em>really</xhtml:em>
> >   cool.</resourceData>
> 
> This would mean that _every_ resource data is now XML and not any more
> an opaque sequence of characters? Even images or LaTeX data? Can I
> flag that what is NOT xmlish?
> 

Actually your resourceData in XTM 1.0 is XML. It just happens not to
have any start or end tags in it. You could never have had tag-ish
content in resourceData (your examples below would not pass as
well-formed XML and so cannot be valid XTM 1.0, let alone valid XTM
1.??).

> And obviously must it be well-formed then to make the whole instance
> wf-ed.  But what if I want
> 
>     <occurrence>
>        <scope topicRef="#theBegin">
>        <resourceData>This is my <x:opening> tag for may fancy application</resourceData>
>     </occurrence>
> 
>     <occurrence>
>        <scope topicRef="#theEnd">
>        <resourceData>This is my </x:opening> tag for may fancy application</resourceData>
>     </occurrence>
> 
> And a TM processor would store it as o DOM o XML fragment o String?
> 

A TM processor would be required to give you back the content of the
resourceData element as a String, modulo XML whitespace processing.

> And a TMCL would be able to touch/constrain data in it? What about
> TMQL, does it offer now access to this? If not, why not, when that
> would be 'standard'?
> 

I think that those are issues for the TMCL/TMQL folks to decide. As it
stands, the TMDM does not provide access to an XML Infoset, so perhaps
the first question should be what does the TMDM reflect....my feeling is
that it should be a sequence of characters only.

Anyway, in what way would allowing TMCL to constrain (or TMQL to query)
the XML Infoset of the resourceData element be "standard" ? Surely what
is in the model is what is defined by TMDM and TMCL/TMQL are designed to
constrain/query *the model* not a particular interchange syntax for the
model.
 
> I find <resourceData/> myself rather inconvenient (fortunately, AsTMa
> shields me mostly from the 'niceties' of XML), but this does not look
> as if it would get better.
> 
> Between the 0% ( = 'dont embed data at all') and 100% (= 'fully embed
> data'), <resourceData/> is somewhere around the 1%. This move would
> now make it 10%. I would rather prefer a more non-adhoc-ish solution
> to be somewhere around 90%, while being clueless how this can/should
> be achieved.
> 

I think its a little more than 10% - for the uses I can forsee for it it
is anyway. But even if it is 10%, surely thats 1000% increase in
usefulness ;-)

Cheers,

Kal
-- 
Kal Ahmed, Techquila
Standards-based Information Management
e: kal@techquila.com
w: www.techquila.com
p: +44 7968 529531