[sc34wg3] Essence of the TMRM
Thu, 26 Aug 2004 10:39:07 +1000
On Wed, Aug 25, 2004 at 11:08:45AM +0200, Lars Marius Garshol wrote:
> * Robert Barta
> | As I understand XTM is using consistently URIs (not URLs) for
> | addressing purposes. This means that no one can stop me to use
> | a URN as subject indicator or even identifier: [...]
> You are of course right; you can define your own pointer mechanism
> with URIs, just like you can with HyTime. I guess the only difference
> is that HyTime makes it easier to do this.
So, I cannot see any problem with XTM (or TMDM) as long as there is
the 'escape hatch' URN. Good.
The remark the HyTime might have more infrastructure doing things
reminds me of the OSI/Internet, Hyper-G/Web, SGML/XML
dichtonomies. Despite what people usually say, OSI was pretty
sophisticated and - for most of its parts - designed with MUCH
effort. Hyper-G (now called HyperWave) is a brilliant hypertext system
and far superior to web. I know too little about SGML first-hand, but
would assume a similar bias there.
So what? The lesson for me learned is that complexity above a certain
threshold will only create 'experts'. That is good to fence off the
'everyday user and engineer', but that is probably not what we want.