Mime types on resourceData WAS:Re: [topicmapmail] Fragmented XTM for web...

[Murray Altheim]

Thomas B. Passin wrote:
> [Murray Altheim]
[...]
>>Except that there's absolutely no need to break compliance with XTM to
>>do this kind of thing. This approach is by its nature completely
>>proprietary and ignores the ready machinery of the XTM interchange
>>syntax.
> 
> Well, yes, that is why I would like to see something like this _made_
> standard.  I definitely do NOT want anything proprietary in the basic
> machiner.

But my point is (and has been for since the release of XTM) that most
everything people need can be done without resorting to breaking the
current interchange syntax, that while there is a string tendency in
some approaches to create new markup, I think for most things in XTM
we can solve the problem without mucking at the syntax level. And by
not doing so, we keep the language easy to understand, easy to process,
and easy to interpret. Once we open the door to the ugliness of XML
Namespaces (and yes, I think the whole thing is a bad idea to a great
degree), we end up with markup that is enormously more complex and
has much more ambiguous meaning, sort of like embedding three or four
natural languages together in the same sentence and expecting anyone
to understand it.

>>And looking at the example provided, it doesn't make any sense
>>to me semantically. It shouldn't be using a <resourceData> to reference
>>an external resource, since <resourceRef> is available,
>
> I slipped up here - I was thinking mostly of literal values, not URIs - I
> should have changed the OPs example.  I do want to be able to include
> literal values that are intended to be markup or have some other special
> format.  I discovered that when I found myself marking up the text of
> literals with HTML elements (mainly <b> and <br>, but <ul>,<li> would be
> nice too) so they would display with useful formatting in  a browser (rather
> than just apply a <pre> element when producing the html page).

But where do you draw the line? Do we raise the bar on XTM applications to
now require an ability to correctly process XHTML? MathML? DocBook? SMIL?
SOAP? SVG?  The list goes on, and I'm absolutely certain that even within
this small community there are those that would like to implement one or
more of these markup languages as content within XTM documents, and then
our ability to interchange them is completely shot to hell. Or we find, as
we do in the browser wars, that only a few players can afford to play, and
that we have also created a competition in who can support the newest, the
latest markup. Ugh.

> I found my app could not really know whether any given literal was supposed
> to include markup, or javascript code examples, etc., etc. (they have to be
> handled differently for display).   That is when I realized it would be
> useful to be able to indicate a datatype for a literal value.
> 
> Even the RDF people have now accepted that it should be possible to give a
> data type to a literal.  Why not Topic Maps.  I do not really see how PSIs
> give a good solution for this.  It is true that I could create an occurrence
> type for a note and use a PSI to say that it could contain html.  But then I
> would have to create a different occurrence type, with a different name,
> using some other PSI.  That does not sound too good.

What's a better solution? Simply embedding whatever manner of markup in XTM?
Your document might work then in one application, and not in another. The
Web should indicate the road we'd be then travelling -- and just try browsing
the Web with JavaScript and cookies turned off for a day or two.

> Or maybe it would be a good place for scopes. That would make more sense to
> me.  But I sense you have something else in mind altogether, and I would
> appreciate it if you would explain it more fully.

Well, to me occurrence scopes are not an expression of a constraint or
type on an occurrence, they are a way of saying the context in which the
resource is connected/relevant to the topic.

But today is Sunday, and after some lunch it's kite flying time!

Murray

...........................................................................
Murray Altheim                         http://kmi.open.ac.uk/people/murray/
Knowledge Media Institute
The Open University, Milton Keynes, Bucks, MK7 6AA, UK                    .

        "There's a lot of intelligence out there that you don't
         know if it's true or not."  -- Anonymous US official
         http://news.bbc.co.uk/1/hi/world/middle_east/3014850.stm