Re: [topicmaps-comment] TMs & XTM [Was: skills to create topic maps]

["Thomas B. Passin" <tpassin@home.com>]

[Lars Marius Garshol]

>
> * Tony Coates
> |
> | If you find it easier editing textual formats for topic maps, then
> | that is what you get for using Emacs/vi/Notepad in the first place.
> | My own experiences, XTM and otherwise, is that using a real XML
> | editor for XML is certainly no *less* productive than using a text
> | editor for plain text.
>
> I think this depends on the user. Developers and other people who do
> editing of text in some formal language (XML, Java, Python, whatever)
> every day will most likely be much faster with a textual format than
> with a graphical XML editor, simply because using a GUI is much slower
> for them. This has been my experience, at least.
>
>...

> It would be good if this were to happen, but personally I am a bit
> concerned that XTM may be too hard to implement for this to actually
> happen. We may want to develop a special no-redundancies
> easy-to-process XML syntax for topic maps in order to better support
> this use case. (This would also be much easier to process with XSLT,
> incidentally.)
>

I think this is a likely thing.  I've done it myself - I sometimes use an
xml format for topic maps that separates out the topic definitions with
their names form occurrences and associations.  That way I can be sure that
all topic ids have been created before they are used in building the other
structures.  That has some advantages for parts of my code.

I do think of XTM as being hard implement for, mainly because so many things
have to be checked or retrieved to do anything.  For example, you can't
easily look for a topic with a given name; instead, you have to check in a
list of names (and possibly variants as well), and if you are applying
scopes you have to check against a list of scopes a well.  That's much
harder than running a simle query.

There is a positive side to this, though.  Compared to something more
general, like RDF for example, the strong structure of topic maps allow a
great deal of predictability in the processing.  It's like jazz - a jazz
tune is always based on earlier songs or musical structures even though
there is improvisation and creativity.  If it weren't, the performers
wouldn't have any idea what to do - there are too many possibilities - and
there would be just chaotic noise instead of organized music.

Topic maps offer analogous advantages, and it would be wise to make the most
of this fact.

Cheers,

Tom P