That sounds like a pretty good idea to me as long it is standardized like you said and not lots of variations on the same thing. It will save late comers some of the tedious work of trying to reverse engineer some game that isn't so popular, aswell as encouraging people to share what they've found.

By the way, what's going on with Data Crystal? It seems to be non-existent now.

Funny thing... I had the same idea this weekend, partly because I'm currently reorganizing my stuff. However, it would need to be viewable in a really good application, so you can have overlaps, mark rom, ram, etc as necessary. Also, for all the asm coders out there, it would probably make sense when it would include various registers and you could check how they are changed, or what they are initialized with. Like, a = Demon number, bc = xy offset etc. And after that maybe c set when a = 0 etc.

So afterall it would probably have to do a lot with coding and then nobody might use it...

cYa,

Tauwasser

I hope so.

Also
The first application might be best written in something like Python or Java so it's cross-platform. Then some faster native versions could be made. I noticed alot of people migrating to Linux recently, I'm thinking of doing the same.

I'm planning on uploading all my notes and such to my site sometime in the future, sort of like the bonus materials in games (or what some people would call 'omake'.) I especially want to upload the full scripts to the games I've translated. How's that sound?

While random offset lists would help, I also have lots of notes within snippings of ASM trace logs.
That could be a bit of a pain to sift through logs dozens of megabytes long.

Interesting idea, just a few comments:

Usual coding style for different consoles differs greatly, with developers of one platform coding in pure assembler while others utilizing compilers. For most SNES games, variables are essentially stored globally at predefined constant addresses, with catalogization being easy to achieve. Games for other (more modern) consoles might just have heap-like structure of dynamic objects with few constant global addresses. One then might want to catalogize the recursive structure of certain data objects, but how do you specify data structures in a formal and universal way? The list could go on, and there are many different types of things catalogizable for different games. How can you provide a universal format for doing all this? There is an unavoidable conflict between structuredness, completeness, and universality.

For each game there usually is a bunch of roms, consider for example all the official translations. Many people do their analyzing and hacking work for their localized version of choice. Often, especially if the original developers used linker tools or else, assembler code routines end up at different places, and sometimes even the addresses of global variables change. This means that address information and trace logs in general only hold valid for some specific version of a game. Doing the same analyzing work for each localized version of game is of course a waste of resources. Can a unified format account for differences between various versions of a game? Or is it better to just stick to the original (most of the time Japanese) release?

Because of the usual lazyness or lack of time, it would be a difficult time convincing sufficiently many people to adapt to this standardized format (even if some kind of portable magic graphical utility were avaliable). Would it not be easier to just constantly remind people to write clear, structured text files and upload them? Or to just upload them (better than nothing), and later have someone else messing with the game do the cleanup work? A well written plain text file can be the most informative thing on earth. There are numerous successful examples, just look at the documents archive on this site.