Fwd: Banging the identity drum again, this time REST-style

Damn reply vs. reply-all... You'd think I'd have email 101 figured out by now... or not, as the case may be...

Begin forwarded message:

> From: "Andrew S. Townley" <[email protected]>
> Date: 22 November 2010 10:41:39 PM GMT
> To: Lars Heuer <[email protected]>
> Subject: Re: [topicmapmail] Banging the identity drum again, this time REST-style
> 
> 
> On 22 Nov 2010, at 10:13 PM, Lars Heuer wrote:
> 
>> [...]
>>> I still see some value for RFC 5988 but not for Topic Maps.
>> 
>> ... you can always map something to Topic Maps, but Topic Maps is not
>> a first class citizen and it will never be. You can always map
>> something to Topic Maps (regardless of TMDM or TMRM) but Topic Maps
>> isn't a player and I don't think that this fact will change soon. So,
>> we're forced to map something to Topic Maps (TMDM/TMRM) and we
>> shouldn't expect that we'd we popular.
> 
> Damn.  You mean I'm back in High School???
> 
> On a more serious note, I think that's a pretty tragic statement.  If it's true, then nobody's better positioned to fix it than those of us with Kool-Aid IVs.
> 
>> Why are CTM/AsTMa= templates popular? Because you can think of
>> triples. You can think of subject has-some-relationship-with object.
>> Users want: subject has-relationship other-subject. That's RDF.
>> Simple, effective. Not necessarily the perfect solution, but a
>> *working* solution.
> 
> 
> I agree with you, but only to a point.  it works right up to the point where it doesn't.  The platform I've been building uses an authoring format originally based on AsTMA= v1.3 (it's backwards compatible, but the most common mode would cause a v1.3 or otherwise parser to choke).
> 
> I find that only the most uninteresting relationships actually lend themselves to using the templates.  Many times have I considered and rejected ideas for extending that metaphor to allow me to specify additional properties without having to write something totally different.  I haven't managed it yet, but that isn't the issue.
> 
> I have well over 16K lines of definitions in this format at the moment, and I can say that I have lots of n-ary relationship proxies - proxies that would "ordinarily" be binary relationships in other approaches - simply because that's the way I feel makes sense to model the relationships.  Binary relationships don't cut it for my problem domain.  I think they're a good place to start, don't get me wrong, but in practice I find that they really aren't all that interesting or useful when you start layering more and more information into the mix (whatever you want to call it; Topic Map, Subject Map, "George", codex, information space, etc.).
> 
> If I wasn't feeling very charitable, I'd say it was a "superficial", if simple solution, but perhaps that's being extreme. ;)
> 
> Again, this is just based on my own experience and view of the world from an admittedly biased TMRM perspective.  Others likely have different views.
> 
> Cheers,
> 
> ast
> --
> Andrew S. Townley <[email protected]>
> http://atownley.org
> 

--
Andrew S. Townley <[email protected]>
http://atownley.org