Author Topic: Helping us help you  (Read 4539 times)

Offline TKaos

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 535
Re: Helping us help you
« Reply #15 on: August 11, 2010, 01:56:59 AM »
Yes there's nothing like a "spread" flag, the only flag I ever used if I didn't know if it was cracked because someone just wrote his name there was [h] and I think it's enough, no need to add more flags eventhought I have some ideas for new flags.

Offline PandMonium

  • Administrator
  • Hero Member
  • *****
  • Posts: 1323
Re: Helping us help you
« Reply #16 on: August 11, 2010, 05:06:31 PM »
I guess we should have a couple more documents and guides on our faqs to help new members to contribute, renaming, checking and improving existent dats.

Trying to give a brief introduction:
As you known, TOSEC is a project with 10 years now, with the purpose of creating databases to correctly identify software images, creating what you called a nice history of past systems. To achieve this, over the years dats were created using a naming convention (TNC) to store the information on the setnames.
Due to several reasons, the correctness of these dats was always far from perfect, greatly varying between different systems. Whether it was due to TNC changes that were not updated in various dats, renamers not following all rules, lack of ways to check if those rules were being followed, renamers that focused on adding new media and so on, the result is what we have now. It is a great project but still there is much space for improvement. TNC rules are now +- enforced, at least we can detect when something is broken and most of these cases are fixed. The remaining problems are all related with the content of the flags.

As said, long ago, many sets were added with very little info, making it possible to add tons of sets, using only 19xx and no publisher name, or leaving out all other flags even it there was enough information to fill them. As you already verified, with a few more checks it is possible to find the dates, publisher names or to add countries and other information.

Currently this is our main goal (or at least mine :P). To improve our dats increasing information for each set, some systems are already really complete but many smaller or not updated for long are not. To do this i try to provide info and tools, complementing the dats information with some extra info that can't be stored there. That is why i asked for information about groups, sceners, persons and publishers and so on. Knowing what is valid and invalid will help us fix it now and in the future but also provides valuable extra info.

Currently we focus more on scene groups and sceners, publisher lists are way bigger unfortunately.
If you like you can keep working on it while you're renaming and i will provide you extra information you might find useful. Currently, on Atari ST - Games - [ST] from the last release, 317 different scene groups were used. For all other Atari ST dats the total used is 382, of which 9 are invalid and need to be fixed, 12 are valid and i miss information for all others.

Continuing on Atari ST - Games - [ST], some infos:
Code: [Select]
Sets: 4948
ZZZ-UNK: 49 (1,0 %)
Unk Year: 237 (4,8 %)
File Types: 2 (maybe there are some sets using .ST and others .st, was expecting only 1 here)
Language flags: 749
Dif Languages: 7
Country flags: 0
Alternate: 1080 (22,0 %)
Alt Desc.: 3
Bad Dumps: 294 (6,0 %)
Bad Desc.: 1

It would be better if all ZZZ-UNK sets were properly renamed, all years were discovered (if available, with month and day). Maybe many of the language flags should be instead country flags. Also, as usual there is an high number of alternates and bad dumps, it would be good to have a descriptor if possible instead of just [a2] [a3] ... [a200] but i know this part is harder, unfortunately for some systems the alts were just changed by emulators that alter the rom content or by someone just to increase the number of 'garbage' floating around.

Anyway, feel free to help us and make part of the project too :)