[topicmapmail] subject proxies and subject maps

[Lars Marius Garshol]

* Johannes Schmidt
>
> while thinking about how (and where) to implement a simple  
> dereferencing
> mechanism for proxies (label -> proxy) I came across the question if
> subject proxies can exist independently from any subject map (both
> according to TMRM), e.g. in an application (context).

You have to remember here that the TMRM is maths, and that a proxy is  
just a set that conforms to some simple constraints. So your question  
is effectively whether the set

   { <{}, {}> }

can exist independently of any subject map. Looked at in this way I  
think the answer is inevitable. Of course it can.

* Lars Heuer
>
> But you cannot "travel" from a proxy to the containing subject map.
> Nobody forbids to implement such a feature, but the TMRM does not
> provide such a feature.

What Lars says here is of course true, and defining this feature is  
trivial. Let's say the function

   map(p)

returns the set of all maps which contain the proxy p. Well:

   map(p) = {m \in whateversetofallmapswasagain | p \in m }

That's the maths. How to do this in an API is another thing  
completely. Maths and APIs are quite different, and so an API can  
behave rather differently from the math.

* Johannes Schmidt
>
> (Apart from any technology/model/paradigm subjects "exist"  
> independently
> from any clustering strategy, of course.)

Yes. Subjects and their proxies are different things, however.

--Lars M.