"Chris Lack" wrote:
> I've begun the work on adding metadata for games/roms.  Currently each
> type has a typeitem.py and typedb.py file.  The typedb.py is just a
> big dictionary keyed by CRC values.  This data has been pulled from
> the mess project via a small script I wrote that parses their .hsi
> files and spits out the resulting dictionary.
>
> The data I'm currently storing in the databases are name, developer,
> publisher, year, country, genre, and notes.  Only some of this data is
> available in the mess databases.  Undefined values are filled in with
> Unknown' currently, other than notes which can be None.  For greatest
> accurracy the dictionaries will have to be hand edited at some point,
> which is about as much fun as doing taxes for your entire extended
> family (I've already done a few of the smaller ones).

Just for me to understand: you don't get the metadata by parsing the
the rom itself, you get it by just building a checksum over the rom
and look it up in a dict? That doesn't look like something for
kaa.metadata to me, but maybe other disagree. Can you at least detect
_if_ it is a rom?

> This will result in some rather large additions.  The nesdb.py file
> for example is over a meg.  I expect close to that for snes,
> megadrive, and gamebody definitions as well.  Is this acceptable
> within the scope of kaa?

I'm not sure if this belongs to kaa or freevo. And I'm also not sure
such a large py file is a good idea. Maybe store the db in a different
form, not sure.

I need to fully understand where and how you get the metadata and if
it is possible to access the information from the web. It always is a
bad idea to have a static database. Look at Freevo 1.5. It is over a
year old, the database will be outdated in that time.


Dischi

-- 
Voodoo is a very interesting religion for the whole family, even those
members of it who are dead.
        -- (Terry Pratchett & Neil Gaiman, Good Omens)

Attachment: pgpFCv8gfc5JL.pgp
Description: PGP signature

Reply via email to