VM are not Emulators, they are as their name indicate "virtual machines", the big advantage of VM is that they can be nearly or even as fast than the "real thing", but they have a deadly weakness when compared to emulators : They can only "virtualize" the hardware architecture on which they run. (with some little exceptions)

As in they won't work on a different architecture : you cannot virtualize an x86 on an ARM CPU like you cannot virtualize a PowerPC on an x86.

VM will only work if in 10-20 years we are still using the same x86/x64 architecture than we use today (or at least something compatible). If we are not using x86/x64 compatible CPUs or if there was too many changes (e.g. older outdated instruction sets dropped like it already happened before) compared to what we have today, then VM will be totally useless. In that case the only hope would be emulation... but for that you will need to pay copyrights to Intel and AMDs.

I recently had to move and in the process found my old computer in the basement where I had plunked it years ago and promptly forgot about it.. On a whim I attached a monitor and what not and turned it on and lo and behold not only did it run fine and dandy but in the cd slot was a disk from myst, which actually played with no problems. hallelujah. I had to do some 'splaining' to my hubby why I insisted on dragging that old monster with us in the move but I am glad I did. I only wish I had held onto my other pcs but seriously, who has the room ?

we are currently facing the distinct possibility of losing decades of not only games but publishing, art, records, personal history and data due to obsolescence of hardware. how many of us have pictures on disks/tapes no longer readable or correspondence disappeared into old locked accounts. Used to be when someone passed on the pictures and stories remained to be sorted out by heirs, now it just disappears into the cloud

at least we are aware of the problem today and I believe that individual human creativity will keep alive old tech when the corporations have long since abandoned it.

Hyperbole much? Shamus Young may be old and pessimistic, but the article (that you haven't even read by your own admission) is far from a histrionic fluff piece. He raises some good points there. The point isn't that nobody will be able to run old games period or any kind of doomsday scenario like that. What he basically says is that due to legal issues, DRM, hardware and software limitations, etc, it will be much harder to get the the games of today running in the future. And he isn't saying that as a journalist. He is saying that as an engineer.

Actually, a company called Eltechs developed a customized WINE VM called Exagear that converts x86 instructions into ARM instructions in runtime. I can run Sid Meier's Alpha Centauri/HoMM3/Stronghold Crusader/Caesar 3 on my android phone! Those are just the ones I've tested personally so far. It's actually a very sophisticated piece of software.

What a depressing article!

I actually found that article pretty good. Maybe I don't agree with all of it 100%, but many things seemed spot on.

I was especially delighted to see him mention also DRM as one of the obstacles for keeping old games alive. I was expecting him to dismiss it by merely "Well, DRM doesn't matter because as we all know, Steam will be among us forever, and even if it disappeared, Gabe has promised to unlock all Steam games at that point.". :)

Related to that, he could have mentioned about the tendency to make forced online features on single-player games. Diablo 3 being a prime example. Those can't be solved with mere cracks, like more classical authentication DRM can.

Also good to see he mentioned consoles are even worse off in this respect.

So all in all, keeping old games alive will become harder and harder. Many current games (especially on consoles) will disappear from the history at some point, something that rarely happens to e.g. books or movies anymore. Things like WINE will salvage maybe some, and the biggest enthusiasts (like me) will try to run old Windows systems on virtual environments, just in order to run some old games. Most probably having to use cracked Windows versions, that don't need online authentication (as the MS servers for those old Windows versions are most probably offline anyway).

GOG versions of modern games will probably be the easiest to run in the future. :)

EDIT: The ending summed it up pretty good: I don't feel quite as powerless as the writer though, as if "nothing can be done". WINE is one example of something being done. The project to create a virtual Direct3D graphics card (which could have several different feature sets) would be another. And as a customer, I can favor e.g. DRM-free games that are more future-proof.

Sure, we can't necessarily save them all, or even most (and naturally everyone has their own priorities for games worth saving; e.g. I couldn't care less if I can't play Candy Crush Saga in 2050). But we can try to affect the future.

One thing I didn't quite understand in the article: To me, the article makes it sound much worse than what the reality is. As if I would usually have to re-install DirectX/Direct3D with the exact version that the game shipped with. So even if I have Direct 9.0c already installed, I might still have to have the game specific version of DirectX 9 installed, again?

That is not my experience at all. For instance on my retro-PC which is running both Windows 98SE and XP, I've installed DirectX 9.0c only once, with the installer I got from MS homepages. That DirectX version has so far worked fine for all DirectX 6-9 (maybe even some DirectX 5, not sure) games I've tried to run on it.

As for graphics drivers, it is true that in some driver versions certain games are optimized better, while backwards compatibility for some older games might become broken. But still it isn't like you need to constantly switch between different graphics driver versions either; usually you find some sweet spot that seems to work. So what if it doesn't yield the extra 5 fps on certain game due to lacking optimizations, or certain shadow effect is rendered a bit oddly on another game? It still works "well enough".

I generally respect the technical knowledge of Shamus Young (especially since I'm not really tech-savvy myself) but I don't think he's right in this case.

To stream games for GOG on their twitch channel I have to get the games ready to stream that can be difficult. Just recently I ran into a lot of problems trying to get the GOG version of Populous: The Beginning "streamable". While the game ran perfectly for me when trying to play it, streaming it (and I was trying for hardware mode and high resolution) wasn't easy.

Not being tech-savvy I tried to find out what was causing the problems by taking my own old boxed version of Populous: The Beginning and trying to get it to run on my system. Slowly and step by step I could start to guess what GOG had been doing to the game to get it to work on modern systems, and I could then fire up the GOG version again and use dxwnd to be able stream it (not in the way I had originally planned, but, hey, I could stream).
Like I said, I'm not tech savvy, but as I see it, GOG already has the technology that Shamus Young fears in his article cannot be created for being too complex.
Also games he mentions like "Jade Empire" and "Thief: Deadly Shadows" are already available on GOG.
Maybe I'm in denial but I believe that Shamus Young sees a problem which is already being solved.

I guess my point is that this isn't news, and thus is a flimsy excuse to raise the same hue-and-cry over well-trodden ground. I think even non-technical folk are aware that as technology changes previous iterations fall out of compatibility with modern systems, and it takes more effort as more time passes to make something run with the same ease. This is simple reality, and doesn't need to be the pretext for heralding doom once again. Especially not when there are people demonstrating daily that these very difficulties can be, and are being, overcome. GOG itself stands testament to that.

As a side note, I do apologise to folks I have offended by my rather strident tone earlier in my posts here. I should certainly read an article if I am going to comment on it, rather than engaging in metacommentary on other people's comments. That's neither fair to the author nor those discussing it. So I will recuse myself from the rest of this, and merely offer my comments in the context of being on the technical issues as I apprehend them.

Shamus is more of a pundit anyway, so this type of speculation or awareness rising isn't unusual for his articles. I think it's the headline that sets up a certain doomsayer expectation, but the article itself isn't actually all that alarmist.

That is true, but he isn't quite talking about that. He's talking about running those games on the systems coming out 10 years from now.

That makes sense in context, and again, shame on me for not doing the required reading before sticking my oar in.

Though I have to admit that the thread title made me think at first that old game developers were being killed off by Javier Bardem with a compressed air canister. ;)

Same here, but for me it was also the Yeats poem. Curse of being an English major.

A good association to make. Bit of a Browning man myself, and of course Coleridge.

hmm, looking back at the history of our current PC architecture, it has a rather incredible track record of keeping backward compatibility. The x86 processor today is still almost fully compatible with the first incarnation of the architecture in the early 80s(!). And there are no signs or indications (as far as I know), that we are about to switch to a totally different, incompatible platform.
The suggestion by the author that in 10 years our current platform will have died and we are sitting on a totally different computers seems like alot of handwaving and groundless speculation to me.

I would be a lot more worried about all the consoles, portable gaming devices, arcade systems, etc. with their unique, custom (and mostly undocumented) hardware. Those face the danger of vanishing really quickly.

Among all those, I feel that PC games have the best chance to survive.

Yes, and see how much effort it takes to get the old games running again even with that incredible backwards compatibility. Often, the only solution is to emulate the old system on the new one.

10 years ago I was running Max Payne on 32 bit Windows XP without any great difficulty. Today, I had quite a lot of trouble trying to get Max Payne running on 64 bit Windows 7, a platform that is very similar to and in theory totally compatible with 32 bit Windows XP.

Where does he suggest that?