Re: TMQL - Proposal: Removal of reifier axis

Hi Patrick,

Delayed response due to 3-day weekend here and uncharacteristically stepping away for the computer for most of it! :)

On 1 May 2011, at 12:20 PM, Patrick Durusau wrote:

> Andrew,
> 
> On 5/1/2011 6:42 AM, Andrew S. Townley wrote:
>> Hi Lars, Patrick,
>> 
>> On 1 May 2011, at 09:39 a.m., Lars Heuer<[email protected]>  wrote:
>> 
>>> Hi Patrick,
>>> 
>>> [...]
>>>>> If a topic reifies association A and A has an item identifier B and
>>>>> the topic has an item identifier C. What's the result for the "item
>>>>> identifier" axis? B and C? Or is it B in the context of the
>>>>> association and C in the context of the topic?
>>>>> 
>>>> Good question. Don't know.
>>> [...]
>>> 
>>> I think the answer should be [B, C] in any case since the assoc and
>>> the topic represent the same thing.
>> I'd agree with this too--especially given the way the implicit merging stuff works in TMDM. Subject indicators and subject locators are specific reifiers with given semantics. Both trigger merging in the effective map based on the spec, so you could reasonably expect similar things like your proposal to take place given the logical merge implied by reification of an association.
>> 
>> The main problem I see is that your proposal makes subject proxies in the general sense first-class citizens in the TMDM. Once this happens, there's no longer a need for the distinction of constructs that can only have terminal property values (TMDM topics) and constructs that can only have non-terminal property values - what I call links (TMDM associations).  This would therefore invalidate a number of fundamental assumptions baked into TMDM systems, and I think that makes it a non-runner (in addition to rather unpopular). ;)
>> 
> How does that "invalidate" the fundamental assumptions baked into TMDM systems?
> 
> They operate just as they did before but with an operator that returns more complete information (the identifier for both the assoc and its reifying topic).
> 
> So, before we declare something "unpopular" let's at least describe it accurately.

A fair point.  Reading this again, I didn't qualify my assumptions statement well enough.  Sorry about that.

If you leave the resolution of the type of the association completely within TMQL and not extend it elsewhere, you're right:  you don't really have to change anything.  I guess I was thinking more about the overall implications of this on the *implementation assumptions* of TMDM-based engines if they were asked to do this sort of thing on a regular basis.  I've never poked around inside one, so I don't know, but I'd think that cases requiring direct access to reifier properties of the reifying topic based on the identifier of the association might not be well catered for.

Wouldn't be the first time I've been wrong, and I wasn't intending to ruffle feathers. :)

>>> The more I think about the removal of an explicit reifier axis the
>>> more I love this proposal. ;) I think it brings us a step towards your
>>> TMDM-NG proposal [1], at least on the syntax/mental level. You can ask
>>> *any* Topic Maps construct about names, occurrences etc. Lovely. :)
>>> 
>>> [1]<http://tm.durusau.net/?p=2893>
>> Thanks for the pointer. I hadn't seen this. Very interesting. In many respects, this is part of what I'm trying to do with Pragmantix, but probably keeping fewer TMDM assumptions.
>> 
> Be mindful that *blind interchange* in based on standardized assumptions. I suspect the fewer of assumptions you have, the higher the cost of interchange.

No argument here.  I believe there are ways to do both, but it does require more explicitly articulating what those standardized assumptions are and how they impact serialization of any information to which those assumptions are applied.

> I think topic maps need to support domain-based, assumption laden models as well as more flexible and open-ended approaches, as well as steps in between.

No argument here either.

> I would like to think that if my blog has illustrated anything, it is that the world is a very big and complicated place.

We do need a reminder of this from time to time, indeed. :)

> Hope you are having a great weekend!

Likewise.

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