[topicmapmail] XTM Datatypes [Was: Adding weigths to associations]

[Murray Altheim]

Martin Bryan wrote:

> Murray
> 
> 
>>I don't know why you'd want to ruin an interchange format by
>>allowing all manner of markup, making the bar to entry and interchange
>>quite a bit higher than necessary, unless you're keen on keeping the
>>ability to use topic maps within a small cult of experts.
>>
> 
> I was reading the message in parallel with SAM as referenced in Lars
> accompanying message, which is not just about XTM: it is expected to replace
> 13250. However, XTM is being toted as the only interchange format therein.
> What I am keen to ensure is that those of us who need to be able to use
> topic maps within other applications can still do so. At present it is not
> clear that SAM will allow this type of application to survive. Hence my loud
> shout.
> 
>  > Look at RDF
> 
>>-- it's extremely difficult for a designer/programmer to implement a
>>system that can intelligently import "wild" RDF content. Or look at
>>HTML -- what a mess.
> 
> Yet HTML is the most successful document interchange format ever, and RDF,
> if we are to believe W3C, will invoke a Semantic Web! (The problem with RDF
> is RDFS and its extensions. RDF did too little, as did HTML.)


HTML is not successful as an interchange format. The Web is successful
in spite of HTML as the means of communication only because enough
people used the mainstream browser features to ensure the network effect
could take hold. And of course take hold it has. But for those that
don't use IE or have JavaScript turned on, for those whose native
language is not English, for those disabled, for those whose OS is not
Windows, their experience is less than satisfactory. Many web pages
don't display correctly or at all, many web sites come up completely
blank or menus aren't available, et cetera. I browse with Netscape on
Mac OS X, and with JavaScript turned off and cookies disabled I might
as well be blind. It's hardly a paradigm of communication to consider
as a model. We can do better.

If you wish topic maps to operate in the same way, where due to mixed
formats, proprietary formats, continually-changing formats (the formula
perfected by Microsoft), we can be guaranteed an interchange scenario
equal to the Web. Everyone's experience is different, some topics don't
show up because the feature-of-the-week isn't supported, et cetera. Yuck.

And you believe the marketing if you believe RDF will invoke the Semantic
Web. RDF is graphs-in-XML, which by itself will do nothing. If there are
problems with RDFS (as you say, which I agree), then a Semantic Web that
relies on RDF/RDFS is bound to fail. But obviously you and I are in the
minority on this issue, though being in the minority doesn't mean we're
wrong. Time will tell whether there's anything real to the claims, just
as AI promised us house robots and "intelligent" systems back in the 70's,
and space stations and rocket ships flying off to Saturn by 2001.


>>Then look at DocBook, or XTM. It's nice being
>>able to rely on *something* in this industry, and I think having a
>>fixed notation for interchange is a good thing.
>
> I hate having to drop those bits of applications that do not fit with fixed
> DTDs. Try writing an ISO standard with DocBook, or using XTM to identify
> whether a topic covers a particular period you are interested in.


If XTM can't identify whether a topic covers a specific period, it
sounds like a failure in the semantics or representation of the topic
map, not a failure of the format itself.

Try distributing an ISO standard using Word. Anyone without Word (or
even the same *version* of Word) couldn't be sure they were reading
the same standard, eg., how would it be if those using WordPerfect
couldn't read specific sections of a standard? This is a problem to
avoid with all effort.

I don't think there's a solution that can both allow reliable
interchange as well as everyone's favourite features. You have to
strike a balance, which I still believe we did quite well with XTM.
And yes, time will tell. I just hope it doesn't get polluted by
all manner of features and extensions like XML has. I consider the
ability to mix in any other namespaces as one of those death knells.

Murray

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

            If you're the first person in a new territory,
            you're likely to get shot at.
                                                     -- ma