Patrick Dowler wrote:
>
>
> The problem with such an approach is that the dictionary (say the list of
> known instruments) changes all the time, so we would also need a process
> by which new tags are added to the XSD, we'd have many XSD versions, we'd have
> docs requiring various versions of the XSD to be valid. if you though shared
> lib versioning was nasty....
I thought about that (even though not in terms of XSD, I don't like to mix requirements with possible implementations) but I don't think that we will have so many new instruments every year ...
Nevertheless I indicated a way out by saying that "One could imagine the existance of specialised sets of context-dependent UCDs managed by each individual project"
In that way the HST part of the dictionary would be managed by the HST project (presumably STScI).
It could be a distributed dictionary, where only a certain number of well defined centres will be able to add to it (a parallelism with GLU here).
Alberto Received on 2003-08-13Z18:33:23