It's far from negligible, especially if you're not using a platform which supports dynamic translation of x86 code. Even so, you can easily run into performance issues. There are many dosbox games that run like shit on my PC, which is only four years old.

> I'd also say having the source doesn't mean it's portable.

You're right, and I'm well aware of the fact. Which is why I said giving people the source is an important first step. If you read my other comments in the thread, you know I know that the game's developers can also affect the portability of their code, if they care.

> which would mean a direct port would be less practical than a reimplementation.

Sorry but this is very uncommon. Even with weird unportable dependencies, it is almost always easier to take existing code and port it than to completely rewrite it (reverse-engineering the original in the process). If you think otherwise, you're grossly underestimating the amount of work it takes to reverse-engineer and make a game.

> I don't think old games necessarily need to be open source.

Games do not need to exist in the first place. But people want them (this is why GOG exists). People also want them to run well on their computers (this is also why GOG exists). Why shouldn't people enjoy old games, just like they can enjoy old books and movies? This is what GOG attempts to let them do. Why shouldn't game developers agree? Why not make it easy for people to enjoy their work any time? Why should games bit rot to hell? I'm talking about all games, not just old games.

I brought up an old game as an example because it shows off what open sourcing the game can do. There is a vibrant, active Doom community to this day. New map packs are being made, new source ports are being made, new mods are being made. There's new stuff and there's improvements to the old stuff. And it really makes a difference. I don't think the current community could exist if we only had DOS Doom. I know for a fact I wouldn't be playing it anymore.

My PC can't even run it in dosbox well enough. Performance is an issue.

And yes there are other old games that are very difficult to run properly on anything modern. If there isn't enough community around it to hack around all these problems, these games will slowly be forgotten. It's sad.

> If developers / studios wish to release the code for research or academic curiosity, then kudos to them.

Do you really think research and academic curiosity is all the source is good for? Maybe you do not care about old games and all the new games that have spawned as a result of old engines being made available to us. That's fine. But for a lot of people, it's much much more than a code museum. People really do play source ports of old games, or new games that are built on the sources of old games.

EDIT: If someone knows of a port or reimplementation of the original System Shock, I'm interested. I'd love to play the game, but not in dosbox, and not without some UI tweaks.

Also, I have reverse engineered games before. Albeit on a very small scale. I've also ported existing code on a number of occasions. I'm aware of the positives and negatives of both methods. Sometimes though, it's just easier to reimplement something than port old incompatible code to a new platform because the old code might have the graphics calls embedded in with the core game engine, and what not. Thus meaning that huge chunks of the core game engine would need to be rewritten regardless. Or in instances where the core game logic is written in 16bit assembly, so couldn't be ported regardless.

I'm not saying having access to the source isn't an advantage, what I'm saying is that sometimes it's not always as much help to the port as you claim with your generalisations.

Edit, freeing the art and sfx resources would be a massive help in almost all instances though