Main > Main Forum
the state of mame
ark_ader:
.
Rick:
--- Quote from: ark_ader on January 17, 2011, 11:11:08 am ---The Mamedevs of today are just standing on the shoulders of giants.
--- End quote ---
...and it's comments like this, that set you aside from those who sincerely appreciate their efforts.
ark_ader:
--- Quote from: Rick on January 17, 2011, 03:19:06 pm ---
--- Quote from: ark_ader on January 17, 2011, 11:11:08 am ---The Mamedevs of today are just standing on the shoulders of giants.
--- End quote ---
...and it's comments like this, that set you aside from those who sincerely appreciate their efforts.
--- End quote ---
Read and learn. There are two sides to every coin.
kagaden:
I don't understand. Why trivialize any of the work the MAME team has done? Past or present? This is a labor of love by the community of developers who make it so.
The underlying argument in this whole discussion is simple:
From an end-user perspective - playable performance on old/cheap hardware. (Hey, pirates are always cheap right?)
From a developer perspective - maintainable, accurate, and progressive code.
I, for one, admire and respect the development team's direction. Looking ahead, it lays the foundation to achieve the goal of true 100% accurate emulation someday. Whereas the end-user's perspective has a workaround now (use an old version) and will ultimately work itself out over time through Moore's law.
I only read the last two pages of the thread but that was enough for me. Thanks for the hard work old and new on MAME :)
My two cents.
Haze:
--- Quote from: Rick on January 17, 2011, 03:19:06 pm ---
--- Quote from: ark_ader on January 17, 2011, 11:11:08 am ---The Mamedevs of today are just standing on the shoulders of giants.
--- End quote ---
...and it's comments like this, that set you aside from those who sincerely appreciate their efforts.
--- End quote ---
and continues to show his complete ignorance to what actually goes on.
as for editing my posts, as I said, they all get edited while they're being written, to clarify points, so that I can work on other things at the same time.
If Pacman was to be dumped for the first time today, and you threw it at the current dev team it would be emulated within about an hour because the MAME architecture has developed to the point over the last few years where that's possible. That's why MAME is a beautiful thing to work with.
All this seems to come back to is him being stuck up on the idea that only a handful of games matter, and they were 'emulated' ages ago, and that everything since then has somehow slowly been destroying the emulation of them because they no longer run well on his anicent box. Therefore in his world, only the devs who did the original emulation actually matter, and the concept that taking something from 95% correct to 100% correct is the bulk of work is some fallacy the current devs make up to sound good while continuing to push out what he sees as shoddy, slow, broken code.
What the original devs did was good for the time they did it, and hardware + resources that were available at the time. What the current devs are doing is good for the time, and the hardware + resources available now. Over that time MAME has come to be a collection of a huge amount of knowledge which is ever evolving and ever improving. New knowledge and findings are regularly applied back to older emulations, improving them. There are very few drivers that haven't changed from when they were written, and very little of the code left from the older versions.
This isn't insulting towards the older devs, because they see this, recognize this, and appreciate the new findings, and understanding of the hardware. It's the same for drivers I've written, there are things I've done which I've not been 100% convinced by, then somebody else has come along, figured out how it actually works, and implemented it fully. I love it when that happens, because it's always good to see what the correct solution was when you couldn't figure something out. I look forward to an accurate emulation of the shadowing hardware in Jaleco Megasystem32 for that very reason, it's got me completely stumped. Devs are NEVER happy about leaving hacks in their code, or incomplete emulation, at least not in a project like MAME.
It's easy to fire up a lot of the older emulations and see flaws that were never fixed. I randomly fired up Taito's 'Hit The Ice' the other day, only to see it leaving trails all over the screens after attract mode. It hasn't seen any attention in a good few years, maybe it's time that driver got some, maybe with some study of the hardware to figure the rest out. No doubt the flaw is still preying on somebody's mind.
Going back to page 1, my concern is there there is no 'next generation' of devs, that was my point. The imperfections we leave in the emulation, somebody else is going to have to figure out while the PCBs still work, not controls, which can be changed at any point, not speed, which hardly matters, but genuine imperfections in the emulation, things which aren't emulated properly and are causing glitches which do not occur on original hardware, or are leaving the games completely broken. We are currently cleaning up after the previous generation of devs, fixing things they left broken/hacked because they didn't understand them, the next generation of devs should be fixing things we've left broken/hacked, because we didn't understand them and so on and that's without mentioning the stuff that actually needs to be written from scratch because it wasn't available to work on at the time. I blame Sony, Sega, Nintendo, and, to a lesser degree, Microsoft + all the big name publishers for creating such a locked down industry.
It's a concern, all these people saying 'it's good enough', 'it just needs to be made faster', 'somebody needs to fork an old version and start again' etc. are detracting from the real issues, and making it sound like there really are no remaining interesting challenges. Hacking up and optimizing something somebody else has already done isn't going to help in the long run. Doing nothing, because you think everything that matters has already been done isn't either. Forking an old version will almost certainly just mean you waste your time getting to exactly where we are now because it's the inevitable path! At the end of the day nobody is going to care if MAME requires a P3, P4, or Core2 to run well when all those systems are sat in museums. They will care if there are still a bunch of broken emulations, or the emulation code no longer runs at all because it was too tied to legacy systems and hardware.
Derrek is right, there is no hope here. Somebody so far distanced from the truth is just a lost cause.