Hacker News Reader

Comment by ryandrake | Hacker News Reader

It really is a shame that "failed" projects tend to just get tossed away rather than open sourced. The excuse is usually vague licensing reasons but, sadly, I bet the real mentality is "if I/we can't profit from my/our hard work, then nobody should enjoy it."

I would love to adopt a few failed or semi-failed games from back in the 90s and resuscitate them to run on modern PCs. But the source code is likely lost.

View on HN

Licensing is a huge issue. If the game is on consoles, that's a whole big chunk of code that has to be removed from the open source version. Console SDKs are usually under very heavy NDAs -- even indirect references to how the SDK works might get you into legal hot waters. Many middlewares, even for stuff like compression, are also usually under pretty punitive contracts. Older games had fewer of those problems, but practically speaking most well known releases in the last 25 years would have to have someone go through and sanitize the codebase, and that can be extremely time consuming. And depending on the specific game, stripping out all that stuff might leave behind just a useless skeleton of code because so many of the basic components that make up the game engine are third-party. Imagine getting access to a game's source code where the entire animation, loading, physics, rendering and many other small systems had been ripped out. You're usually not left with a whole lot of other stuff in those situations.

Not just for code, either. It's common for games to use licensed graphic/sound assets, either wholesale (e.g. a finished character model) or as components (e.g. a texture or sound). Licensing terms for these assets are generally incompatible with an open-source release.

Yeah I didn't even want to wander into the minefield that assets might be. Code alone is usually enough to make open sourcing infeasible, assets are even more complicated because there's way more unique assets than 'code components' in most games. Tens if not hundreds of thousands of individual assets that need to somehow be curated for release.

I don't think the licensing is really the excuse. It is more of the fact that studios have put tons and tons of effort towards those assets including their creativity towards that partly finished end-goal. They most likely want to keep it in their possession if some point tides turn and they are able to finish the game or even small part of it.

And I totally understand that. After sacrificing years and years + thousands/tens of thousands euros towards project, you really don't want to see it taken by other people and putting their own creative ideas on top of your work(that can be potentially used later).

For the assets yes, especially. It's hard enough explaining to the Big Boss why the game your division worked on for X years is not going to release, but imagine having to explain the next year why some asset from it is all over YouTube.

> The excuse is usually vague licensing reasons but, sadly, I bet the real mentality is "if I/we can't profit from my/our hard work, then nobody should enjoy it."

I'm 100% certain that the primary reason boils down to "we don't have the rights to all of the source code, and we don't want to deal with any of the options for moving forward."

As sibling comment points out, a lot of console SDKs are pretty heavy in the "this code must never be open-sourced" space (hell, even talking about what's in them can run afoul of NDAs). But it goes beyond that. Building assets might require a suite of proprietary tools and custom extensions for those tools, where your ability to release those extensions may be legally unsound. Many games will have several libraries, not all of them open source (and some of them probably under pretty strict don't-even-think-about-asking-us rules--can't imagine many anticheat or DRM packages are going to say yes to allowing source to be released).

If the company doesn't have the rights to large chunks of the source code, their options boil down to:

* Release just the bits they have the rights to, which isn't going to be compilable. And hope that nobody complains about residual references to their code in the released stuff (litigation risk!).

* Spend time going through the code to sanitize references to unreleasable code, but still just release the not-working code drop.

* Do the above, but also replace the unreleasable bits with working releasable bits.

* Do nothing.

One option involves no work and no risk. One option involves minimal work but incorporates small-to-medium risk, but also sacrifices much of the benefits of open-sourcing (there is likely to be community blowback if the product is unusable!). Two options involve a lot more work to mitigate the risk.

It's going to be rare to find the corporation that's going to go for the no work/no risk route. If you want to change that, the best way is to create legal mechanisms to make the no work option much higher risk than the other options.

I think the "no work" aspect gets underplayed in a lot of these preservation discussions. At the point you've got a failed product you're either scrambling to start or do something new to pay the bills or winding up the company. There seems to be this idea that everyone involved who could work on the old product has a lot of free time and won't see a significant return for it.

It's similar to the rumor that if something apocalyptic happened to Valve they'd keep steam open and let you download all your stuff, if that scenario happened they'd be least able to be able to do that for a long list of reasons unless they had instantly been bought.

I think you are underselling licensing issues, especially of third party libraries I can imagine for things like ui, music, physics, etc.

Honestly assets can be extracted from a lot of older games, the issue is usually licensing.

Nightdive has a good writeup somewhere of the hoops they have gone through.

They say it usually boils down to:

"Mr Rights holder, can you license this property to us"

"No I dont know if we own that property, we would need to check paper records and that costs money"

"Ok, but then you wouldnt go out of your way to sue us if we did resurrect this old game"

"No if we saw you making money from it we would definitely spend the money to figure out if we had a viable lawsuit"

The source code is likely not even terribly useful, for a game that was never released. What this is talking about are the assets (sounds, 3D models, textures, etc) which could still be quite useful.

And that's why they don't just give them away (unless it's a very small studio/individual effort). Many "how it was made" histories of games describe how they used assets developed for a previous game or otherwise not used.

Could also be that open-sourcing something is probably lots of work.

I've seen internal projects that were open-sourced and there was a ton of time and effort involved to get across the finish line.

Probably anything involving more than one person make exponentially more work, let alone licensing/etc

Reply on news.ycombinator.com