Slightly OT: Interesting MAME statistics.

Main > Everything Else

<< < (5/15) > >>

PL1:

--- Quote from: Howard_Casto on October 29, 2015, 03:54:59 pm ---Instead of a slash with a more specific category. (Fighter / Vs) I'd rather see multiple boxes that can be checked.

--- End quote ---
That's what I figured.

Each game/console/computer/bios described by listing any and all relevant categories.

--- Quote from: Howard_Casto on October 29, 2015, 03:54:59 pm ---The way ini files work in windows is you can default to 0 if not found, so all the categories a game isn't don't need to be added, just the values that are true.

--- End quote ---
Agreed, but I'd still include all possible MAME files in the "list" section of the .ini since it's easier to change a number than add a file name.

--- Quote from: Howard_Casto on October 29, 2015, 03:54:59 pm ---Of course all possible flags have to be listed somewhere. That is another minor flaw in the catver system.

--- End quote ---
Maybe make a complete list of them at the beginning of the .dat file or create a companion file that contains the list. :dunno

The game list editor could then use that list to generate a scrollable list of checkboxes on the left side of a GUI with a tab-selectable view on the right side of "Filtered List", "Complete List", or "Exceptions Only".

Selected games/systems shown with one background color.

Non-selected games/systems shown with another.

Exceptions color-coded to indicate that were manually added/removed.

Scott

dkersten:
I think I even asked here about a program that will allow me to filter my rom list down to what I want and then copy those roms to another directory. I ended up doing a lot of it manually since there are no easy solutions. It would be nice though.. have one master rom list of whatever flavor, plug in the folder and check off the criteria you are after (ie arcade cabinet games, best working versions, type of game, etc.) and select the name of the folder you want it in and push the button. THEN add support for front ends to build your lists so you can just import them. Boom, would have saved me about 100 hours.

Everyone wants their software to do everything, and the ones who made specific programs made it for what they wanted, which didn't ever fit what I was looking for. So no solution for this, and as fast as things change with MAME, anyone who tries is going to have a lifelong project on their hands. Every few months when a new version comes out, everything has to be gone over with a fine toothed comb to update it, not too big of a deal once it is done, as long as you keep it up.. Much harder to deal with halfway through the project. As much as I would like this kind of program (and data file) to be available, I'm not sure anyone is interested in tackling it. I don't blame them.

Ironically this is not unlike what I work on for my job these days. We are a wholesaler and carry around 19,500 parts that change by around 10% each year. Some parts are similar, but thousands of them are odds and ends that don't share any criteria with the others. My job is to come up with a catalog that any customer can easily use to find what they want, sorted into just a few main categories that you can break down from there. The inventory is so dynamic that no static solution works, so you need to come up with a means of categorizing stuff that will last, then build a front end that is flexible enough to withstand the changes. That's the easy part though, actually sitting down to categorize nearly 20k part numbers when you don't even know what 90% of it is for is the hard part, particularly when it is in constant flux. After a couple years we finally got a handle on it, but the categorization is (so far) only two tiers deep. Some products should be categorized up to 7 tiers deep, with 3-4 tiers being independent. Yet the front end needs to know whether to categorize it with the first two tiers or to start sorting by more. To add another level of challenge, about 4,000 products could be categorized under completely different means, but the database we have to work with is limited to two main tiers that are dependent on each other, and then several independent categorizer fields. I have taught myself HTML5, CSS3, PHP5, and MySQL just so I could do this, and while it is certainly something I enjoy digging into (And I am getting pretty good at considering my only formal training was on Basic programming in the late 80's), I wouldn't do it if I didn't get paid really well. I certainly can't see taking on a similar project as a hobby. Maybe building the front end (and that is a BIG maybe), but not categorizing the lines of data... No way.

After going through it once, I know what it takes. In the beginning several people are offering to help, but part way through (especially when things change and you have to shift gears) people drop off and soon you are left alone with an overwhelming amount of work to do before you can even start the fun stuff you set out to do in the beginning. It will turn out that the stuff others were helping with is such an uncoordinated mess that it takes longer to sort it out than to just start from scratch. Three quarters the way through you will figure out that you should have taken a completely different approach, but it is too late and that just adds straw to the camel's back. Then finally you will get it done and start on the part you really like, only to find that you no longer really care about it enough to put in the effort, so you just half-ass it enough to get it working. Two years later someone discovers it and loves it but starts asking you for the features you really wanted to add in the beginning but lost sight of when you got stuck with the boring stuff, and now you don't have time because you are stuck in another similar project that is going about the same as that one did.

I sincerely hope you pursue this, and if so, good luck!!

Howard_Casto:
Yeah that's exactly what happened to controls.dat. I got unwilling and sirpoogna got unable due to work. Then we both got unwilling when mame radically changed it's control reports.

Mame's naming system is more of a hassle than anyone can imagine unless they've dealt with it before. Even when the list is "stable" there might be a parent rom named "rom" and it's revision 3 of the arcade game. Well if revision 4 is found then rev 4 is now named "rom" and the one that used to be named rom is now called "romv3" In theory this is a good thing as the latest version of the game always becomes the parent, but when doing something like a rom manager or a controls database it's a nightmare. Often different revisions of a game have different control schemes like tmnt vs tmnt2p and even if they don't you need to check and make sure and there's tons of renaming. Then every now and again it's decided that all the roms for a driver need to be completely renamed...because.... reasons? So you have 200+ games that are now renamed and you need to know the original name as well as the new one or else throw out the data and start over.

Howard_Casto:
Well it looks like AntoPISA (progretto snaps) has taken up the cause. The last time I checked he was just as disgusted as I was and wasn't sure if he'd continue making lists. It looks like he finally bit the bullet and categorized everything in mame, even the "misc" junk. His full version is only on 166 though.

I don't like the idea of depending on one guy to get all the data.

Howard_Casto:
Just for those interested......

I sorted my games out. If we are talking just parent games with all the other stuff I mentioned filtered off, the roms total a mere 11.1 gigs. So yeah that's roughly 25% of the total collection size and worth filtering off. My filters seem to need work because the bios didn't get copied over, but once I get everything ship-shape I'll release something.

Navigation

 Message Index

Go to full version