TOSEC?


skeezix

Internal Development
Joined
Mar 11, 2003
Messages
8,063
Website
www.codejedi.com
I've been looing at the TOSEC stuff, trying to make some sense of it, from the outside. (ie: Not knowing much about TOSEC, trying to peer in. The site helps a little, but not a lot)

How do you get a dump of the database as a textfile, mapping content to CRC32?

Theres TUGID which seesm to be to build the database and do renaming and such (and why isn't it open source? Why is there "attitude" with that site?)

I do see some ST references in one of the file listings, but its changelogs.. ADDED this or that to this or that. No complete listings.

You can do a search to try and get listings out.. but thats not useful.

I must be missing something :)

Seems like the source for all tools should be well published, and the database shoudl be readily available in numerous formats.. TXT, CSV, etc. If its in a tgd file or whatever, there should be a tool to spit it out in various formats. Thats how it gets useful :)

Am I being an idiot? There must be somethign there :)

jeff
 
Ahh, got it. Need to find the .cm files, which have .dat files in them, with raw usable text. I can parse it with a perl script to spit out something I can use for other purposes.

Google was able to scan the tosec site to find the "all releases" option :)

jeff
 
I dont claim to be an expert (far from it :) ) but you could try THIS LINK

Its a ClrMamePro datafile i think, but it contains ascii text in a logical format.. including name/size/checksums etc.. should be easy enough to decipher.

Just noticed the COMPILATIONS
are in a seperate file ..

Thanks for everything :)
 
Doh!

Doesn't appear to be useful... example:

game (
name "Automation Menu Disk 031 (19xx)(LSD - Was -Not Was-)"
description "Automation Menu Disk 031 (19xx)(LSD - Was -Not Was-)"
rom ( name "Automation Menu Disk 031 (19xx)(LSD - Was -Not Was-).st" size 737280 crc e8a4d7c5 )
)

OKay, so I can use TOSEC to nail down for sure the disk is Automation 31, but we still have no idea what games are hosted therein. Darn :/

game (
name "Pompey Pirates Menu Disk 003 (19xx)(Pompey Pirates)"
description "Pompey Pirates Menu Disk 003 (19xx)(Pompey Pirates)"
rom ( name "Pompey Pirates Menu Disk 003 (19xx)(Pompey Pirates).st" size 737280 crc c19a641e )
)

Theres another one for example.

Durn, so its a good idea, but not useful for our purposes of mapping disk-image to game listing :/

It could be used to map the file to the TOSEC, and then we could make a TOSEC to game list file. It woudl have been better if the TOSEC datafiles included some unique-id to identify the disk.. ie: IF you've got 3 automation-37 disks or whatever, then to refer to a TOSEC automation 37, you have to go by crc, which may not be unique in the TOSEC database. But h'okay fine :)

So I'm not sure TOSEC is useful to us at all.

jeff
 
I quite liked the idea of a user-editable filename (or CRC) to name mapping.. then a merge of everyones files. Although its not a huge concern to me atm, since compatability is higher on the non-compilation disks anyway, so I only use one or two :) The feature im really hanging out for is macros/definable keys and a gamma correct.. Already, caSTaway is almost perfect :)
 
Belgarath is the name of the guy you're looking for
 
yeh i'm the guy :)

people complain about the length of the tosec filenames enough as it is so putting the names of the games on the disks into the filenames wasn't really an option. a small number of them though have been listed with what they contain, on the page with the databases, simply download the idx file that appears next to the st compilations and st games databases

oh and there are no dupe crc's across the atari st databases
 
The database shoudl include the contents of the compilation, to be useful. I wonder how open they are to suggesiton, to extending the format? A TOSEC-id woudl also be useful, since CRC's arent' that reliable. A true unique ID would be very useful.

But okay, I'll probably make a CRC32 to compilation-contents database. Then later someone could merge it with TOSEC if they wanted to :) I could use TOSEC as an intermediate step.. something like this..

(the ClearMame format or whatever, thats text and easily parsed -- why did they use a binary format for the distribution? Silly :) At least XML ;)

Get a file, calculate CRC32 on it, look it up in TOSEC to get the long filename ("automation 031"). Then pick up the import [parts and toss out the kruft.. "automation" and "31". Then look this up in a master "automation" list for instance, disk 31, to get the contents. Making this database will be a bit fo tedium, but nothign a unix scripter like me can't handle.

So file -> CRC32 -> TOSEC long name -> break it up -> match to known database of compilatoin cotnents -> summary.dat filr gp32 :)

whew :)

jeff
 
OKay, one must also say that when he used a broken CRC32 alg instead of the standasrd ones floating around, he needs beetings :/

I tried 3 CRC32 alg's that all agreed on a value, then dug through their code to find out when they say CRC32, they dont' mean it. Its their own slightyly broken CRC32.

Garg.

But okay, 3/4 of the way there, just a bad taste :)

jeff
 
Back
Top