Hello,
To sum up more roughly :
- We need Unique Ids, as otherwise we're not dealing with any efficient
database model known to date ;
- We need Flags to speed up searches and normalize institutional work ;
- We need Keywords because users like to personalize their stuff : cf.
Internet's tags frenzy.
I agree with this. Here are some comments :
1. Unique Ids are also essential for a reference system. With unique Ids,
we could link to another game in the database, maybe even in any database,
instead of trying to merge every games into every games.
2. We need to distinguish when a flag denotes a game as a whole, or a
position. For instance I don't know how to identify my repertoire by games
and not by positions. More generally, we would need something like a
semantics for tags. If the CentriScid comity is in charge of flagging
games, we need at least to describe what is the usage. I personally think
flags should be predicate for games. Lack of clear semantics is the main
reason why I don't use flags. What exactly means "User" flag, anyway ?
Miscellanea could be reserved to keywords.
3. I suggest we have a flag that either tells
Hey ! This game is checked with reliable sources, overchecked by
multiple persons and we don't need to bother about it's accuracy.
or
Hey ! We need to check the actual accuracy of this game.
Do we need to identify what we know or what we don't ? I need to think more
about that. My gut feeling is that the critical moment comes when a
database merges into another. Thence comes the errors, usually. Correcting
errors is frustrating alone : it's an impossible task when you can't even
tell what is done and what remains. At least, we should be able to tell who
did what.
4. I suggest also that we have a flag that tells if a game can be used as
an example or an exercise. Basic training dichotomy, IMHO. We could let
that flag to be checked by the trainer himself. We could also let him use
personal tags. So that suggestion is not essential.
5. I don't believe in terminological classifications. But if someone wants
to indulge into creating another ontology, the best one I saw for chess
comes from Berliner and Kopec. I can provide some references if needed.
6. A possible alternative to #5, i.e. structural classification, may be
done solely with search files. So there would be nothing in the code to
change. What we would need is to collect and provide patterns.
I'll return later this evening. That's all the tranquility of my kids can
allow me for now.
So bye for now,
B
-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Scid-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scid-users
