Le 24/02/2010 14:37, Alexander Wagner a écrit :
>
>> 2. Should we ask TWIC again?
>>      
> IMHO not needed at all. But I don't see why Mark would
> refuse.
>
>    

Some web sites are commercial, and I think it is fair to ask before 
using them in an automated manner (for example hooking 
http://www.shredderchess.com/online-chess/online-databases/endgame-database.html
 
was not allowed).

>> 3. If we had an ok, would it be tough to implement?
>>      
> It is actually trivial.
>
>    

There is already some scripts around for that. But integration in Scid 
is certainly better for the user. The question is : how long will the 
service last without teaking a bit the code and/or URLs ?
I never tried with chessDB (indeed, dead project now), but I was 
reported that this did not work and nobody was fixing it...
But if TWIC's setup seems stable, then maybe it is time to plug it into 
Scid ?

> BUT: there is currently no defined interface for these
> things. Therefore you'll have to do a specific script for
> each site you want to consider. And you'd have to either
> invent a pretty complex parser that allows the user to
> specify how to handle the site or you have to hardcode
> things in Scid. The latter is considered broken by design.
> Especially if you don't talk about persistent sites.
>
> Just to name a view points: TWIC used some internal path in
> the ZIP files for some time while it was skipped later on,
> reintroduced and seems to have gone again, now. All file
> names are given in sequence and called twic<number>g.zip
> containing the PGN. Trivial to do, but...
>
> The structure of ChessCafe is quite different. The file is
> always called Default.zip and covers some undefined time
> frame. The ZIP again contains a pgn file to unpack and load
> into Scid.
>
>   From ICCF Webchess on the other hand you get a single PGN
> for each game.
>
> Conclusion: there're far to many different ways to handle
> PGN donwload to have a general interface PLUS you have no
> way to tell if any of the things you do today will work
> tomorrow. So keeping things up to date is considered
> troublesome.
>
> IMHO it would even be better to set up a webservice
> somewhere that collects the games from the various sites and
> offers them for download via some webservice/soap-calls.
> Then we could add a simple inferface to Scid that allows to
> select certain collections that get fetched to the local
> database. Something similar to RSS for chess games.
>
>    

I agree. This would be really useful. But this is another project ...

> Another option would be to have such fetch scripts done in
> some sort of plugin architecture. But that approach didn't
> get to much positive response when I suggested it way back.
>
>    

Plugins add trouble. My reasoning is simple : if a plugin is really 
useful, why isn't in the core code ? What is the difference between a 
broken plugin and a broken feature in the core (Tcl plugins should be 
loaded the same way other code is) ? Howto handle i18n with plugins ? etc.

Pascal



------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Scid-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scid-users

Reply via email to