The Prague meeting
WG3 in session
The ISO SC34 meeting in Prague was a big affair with five different working groups and many attendees. Working group 3 had a lower attendance than usual (for a number of reasons), and perhaps for that very reason had a highly productive three days focusing on TMCL. The status before the meeting was that we have a quite loose draft that shows in rough outline the intended functionality of the language and gives a good indication of the way it's intended to be specified. The task of the meeting was to process this to the point where the editors could write something quite close to the final specification. I'm happy to say I think that's what we did.
We settled a number of minor issues, like how to specify unlimited cardinalities in CTM, naming convention for PSIs, validation of the TMCL schemas themselves, how to define loose versus strict validation, how to handle documentation of TMCL schemas, which regular expression syntax to use, and so on. No less importantly, a number of new features were considered for inclusion and dropped.
Some minor changes were made, like collapsing the three types of scope constraints into one, adding support for constraining reification, moving regular expressions on names/occurrences to the individual type, changing the meaning of identifier URI constraints, etc.
The topic representing the TMCL schema itself will be added back, because this provides several different kinds of useful functionality, such as adding metadata to schemas and tracking which constraints came from what schema (after merging of schemas). However, using the associations connecting individual constraint topics with the schema topic will not be required. This makes it easier to write schemas in CTM, and simple tools can add the connecting associations if needed.
How constraints combine formally was decided a bit more clearly, and one big change was made: constraints are no longer defined using TMQL, since we want to publish TMCL before TMQL is finished. This means the editors have challenge before them in working out how to define the semantics of the constraints.
Two minor changes were also made in CTM. One is the use of the "*" literal for unlimited cardinality in TMCL, and the other is template overloading, which clearly is so generally useful that it just needs to be supported.
Some detail can be found in the slides from the meeting. The next step for the editors is to write the draft as soon as at all possible. The intention is that this draft should be ready for implementors and users. The further course of events will be determined by the feedback that is received on this draft.
Since the meeting was in Prague I took one day off before the meeting and one after in order to have time for sightseeing and beer tasting. That's a subject for a different posting, however, which may or may not get written.
The TMCL standard now seems more or less stable, and so now it is finally possible to explain to outsiders what the language looks like and how it works
Read | 2008-10-03 17:33
ISO has been working on creating a graphical notation for Topic Maps for some time now, and now the first proposal for level 1 has been published
Read | 2007-07-05 10:37