Wow... that took a long time to read indeed... 3 hours in an internet cafe, and I'm feeling catatonic lol. However, think I ought to add my 2c simply coz I want to.
Personally, I think the device looks extremely slick indeed. What's more, it looks to overcome a few of the weaknesses of our old and extremely faithful friend, the GP32. So that's good as far as it goes.
It doesn't, unfortunately, have Bluetooth, wifi, or anything similar, which is a bit of a dissappointment, and yes, as others have said, I'm a little underwhelmed by the 200MHz cpu. However, I must say, I find a 166MHz GP32 quite amazing still, and the difference when moving up to an insane 200MHz OC is very much noticeable, so its still a plus.
With GP32 compatibility, this would, in my view, be a certain buy... as it is, I'll probably play the waiting game for a month or two to see what happens, possibly (you never know) in time to commit myself for X-mas
.
However, there's a few things I feel I do need to point out.
The first is that, a very very quick google for ARM920T yields the TumTum, a GPS device that, in some of its models, runs with that chip set at 400MHz - which is clearly a great deal more than 200MHz. Obviously, however, we don't know yet whether the chip will even overclock - or underclock, for that matter - a bit (possibly one of the things that limits backwards compatibility - remember a lot of GP32 proggies change the CPU speed every few minutes) let alone to 400MHz, but it is well worth pointing out that another device exists for sale at the present time with the same CPU core set at a good bit above the 250MHz mentioned on ARM's site or the 200MHz we've now been promised.
The second is that just as the GBA's ARM7 was MMU-able onto the GP32 in the form of GP-Advance, I don't see any reason why, with an array built to open and read components of an .SMC file sitting on an SD card, the same couldn't be done with the GPX2 to give us backwards compatibility. I imagine, of course, that again the CPU speed would be messed to hell and back, since even at an ideal time, I can't believe we'd get more than a 1:2 instruction ratio (as in for every 1 instuction in GP32 code, 2 CPU cycles would be required), but for old stuff or games, that would help a great deal.
This is an especially plausible idea considering I'm fairly certain the GP32 CPU is a different ARM9 chip, if not actually the same one, so the code ought yo be even easier for it to deal with than that from a chip a couple of generations earlier.
Furthermore, I remember Rattboi telling me in IRC once, during the time when GPAdvance was going through an update a day that he "knew it ould be done, but [he] didn't think it would be done that easily", which implies that, with a few references to our one decently speeded GBA emu one could probably re-write it to give a GP32 output instead. Especially since the reason it is slow is not because of bios wossnames, but rather because on top of the ARM7, there's the GFX chip as well.
The final point I'd like to make is to consider the reasoning behind not making it GP32 compatible. I would suggest that there are several possible reasons for this. The first I've allready touched on - the GP32 plays havoc with its own clockspeed, and we cannot be sure that the central chip we've got in the box now is capable of handling that havoc, even if the core itself is. There are two other obvioous hardware-based reasons. Namely, the screen is different, which could conceivably lead to BLU+ type problems all over again, except possibly worse (who knows?), and the old GP32 bios and SDK were both designed with SMC cards in mind, which, as many here know, are radically different in terms of how they work to an SD card.
So why didn't GamePark Holdings simply take their bios source and write a transparent layer convert SMC instructions to SD ones, possibly at the expense of a few cycles per second whenever reading or writing is taking place? In my view, that's a crucial question. The answer could be that other hardware problems ere also present, so they didn't think it' be worth the bother - the screen could be a LOT worse for compatibility with GP32 screen output code than the BLU+ was for all we know, or the CPU clockspeed changing issues may have caused too many problems with the 2D accellerator. Or there's another reason.
Ladies and gentlemen (mostly gentlemen, but it sounds better with both), I refer you to debates that have taken place in the last few weeks revolving about the presence, now, of not one but TWO companies - Gamepark, and the more homebrew-friendly Gamepark Holdings. I would suggest, perhaps, that the reason no layers or similar have been written is simply because they don't actually own the code the bios for the orginal GP32 is based on. If they did write GP32 compatibility into their new console using the method of shoving a modified old bios in, then conceivably Gamepark could take offence and sue - and you never know what might happen, considering that there's clearly been a quite substantial difference in opinions about which direction the company should head anyway, seeing as there are now two of them.
So what's the oucome of all this? How does it change the situation? Well, to be honest, I don't know. But I would think that it implies there is more to the "no backward compatibility" than merely them not wanting to try. It could me a bit like Gamepark's original stance on emulators - tolerance, but not support. The upshot, though, is, that *possibly* (by no means certainly), backwards compatibility IS a possibility, but NOT one that will ever arrive officially - its one some bloody skilled coder is going to have to pull all the stops out to get to.
Thankfully, we have no shortage of bloody skilled coders in this scene.
Anyone have any comments about the theory, please do say, because it'll be a break from the "GPx2 is crap - no it isn't" arguments that take up probably about 20 pages of this thread lol
P.S.: For those of you who've joined in the last 6 months or so and therefore haven't a clue who I am and actually think I know what I'm talking about as much as some of the people who REALLY know what they're talking about (Squidge, skeezix, many others), I'd better point out that I don't - I'm just not bad at thinking about solutions to situations; everything I'm saying is conjecture and nothing more - unless some people who do know what they're talking about say otherwise.