[sc34wg3] Analysis of TMRM Use Cases

[Dmitry]

On Apr 7, 2004, at 9:32 AM, Steve Pepper wrote:

> Thanks to Patrick Durusau and Steve Newcomb for submitting
> N490 (Topic Maps -- Reference Model Use Cases).
>
> The purpose of that document (which WG3 requested at its May
> 2003 meeting) was to help us arrive at a common understanding
> of what the *purpose* of the Reference Model is, because without
> such an understanding it is impossible to evaluate the need
> for a reference model or the suitability of the current draft.
>
> Having studied N490 it seems to me that the purpose of the RM,
> as the editors envision it, can be summed up as follows:
>
>    To enable the documentation of the specific merging
>    rules that are required to be applied by a particular
>    Topic Maps application [1]
>
> I believe that this objective can be satisfied by the TMDM
> in combination with TMQL (or by the TMDM in combination with
> TMCL, which in turn uses TMQL). If that is the case, then I do
> not believe that N490 provides sufficient justification for
> developing and standardizing a completely separate reference
> model.
>

Steve,

I agree with the opinion that TMQL can be used to express what have to 
be merged.

But I see in some cases a need to describe what merging means for some 
constructs which are not standard for
TMDM. For example, let's say that we would like to define ordered sets 
in Topic Map.  I can introduce PSI for ordered  set but I also need to 
define how merging works for ordered set. Merging depends on a way of 
representing ordered sets and may require custom merging definition.

Dmitry