The simplest solution would be a straightforward law mandating that end users be given access to the Source Code of any software they run on their machines and to which end, nothing in the Source Code is secret from the end user.

How many of the end users can read & understand source?

Those who can't won't and very likely won't need to, those who can will.

How many can accurately interpret a legal document? Yet most of us are shown one we must agree to before we can install a piece of software.

If they weren't such private information vampires, this could be a good job for LLMs.

I remember someone advertised a free service which renders legal junk and shows whatever it means. It was either a service or software.

I take no position on the OP's suggestion, but I doubt whether that matters. If the code has to be made available, someone will be able to read it, and that would certainly "focus the minds" of the original authors.

Maybe they can't themselves, but at least they will be able to pay someone else to do so on their behalf. Which will create jobs.

Cloud cuckoo land calling,

"How many of the end users can read & understand source?"

More than zero, which is enough.

True, but the number of people who think they can read code and are able to misunderstand it is probably a far larger number.

40-something former users of Winamp? Probably most of us.

Cloud-hosted software as a service has entered the room.

Cloud-hosted software as a service has taken over the room.

Maybe everything you ever run is open source. Not everything I run is. Sometimes, the way to get people to write the code you want to run is to pay them for it, and I don't have the funds to finance the development process myself, but I can easily afford to buy licenses for stuff I want. I prefer to run that stuff on my own computer, but if they're required to hand me the source code in that case, they'll restrict it to running on their servers or whatever environment it takes to prevent me from accessing that code.

Proprietary software is not automatically evil.

[runs off to github to search for winamp]

Response: Whoa there! You have exceeded a secondary rate limit.

ROFL!

It seems to have been like this for a while. [1]discussion at GitHub Community

The github-actions bot has responded.

[1] https://github.com/orgs/community/discussions/141073

If so much 3rd party code was in the repository did they even have full rights to distribute the binary?

Probably not. But as long as they weren't making the Source Code available, no-one was any the wiser.

[Author here]

> If so much 3rd party code was in the repository

Hang on. I think you're making a big assumption here which for me ruins the joke. (It was a joke, right?)

"In the repository" -- *what* repository? What makes you think that they had a repository? Source repos are an element of a version control system. What makes you think they had one of those?

I think they probably had a big ol' Zip file, or a backup disk, or tape, or something like that, and some rookie staff member who doesn't know how Git works was told "put the source code on Github" and they tried.

Look, Microsoft *owns* GitHub, and when some staffer uploaded the rediscovered MS-DOS 4 source code:

https://www.theregister.com/2024/04/26/ms_dos_4_open_source/

... they screwed it up and did rookie mistakes like mixing up CR/LF versus LF line endings.

Git is horribly complicated, more than it needs to be for 99% of its users. Github adds an extra layer. It doesn't just need training: it needs deep knowledge and understanding. Most people using it do not have that.

Me included, and I did daily for over 4 years.

I strongly suspect a 20th century freeware proprietary Windows app had nothing like that. Maybe a shared network drive on an in-house server.

As a mercurial user, git's faffing around with the line endings of "text" files (and the need to declare things as binary) really pisses me off, and corrupts data that is used for test cases. It's a return to the dark ages. Every modern text editor can (and does) automatically adapt.

And the server was called 'server' and browsed to via it's netBIOS name.

And the shared drive was called 'shared' and it's permissions where set to 'everyone'.

Regardless of whether it was a repo or not, they were apparently distributing GPL licensed code as a proprietary binary. Otherwise why would the code have been in the zip file or whatever?

Agreed about the complexity of Git and Github. For personal projects I sometimes use Fossil, but Git and Github are what everyone knows. I mostly use local tools as far as possible even when using Github so keeping it (or at least trying to keep it) to being just a place to host a repo.

> Otherwise why would the code have been in the zip file or whatever?

Um. Look, I try to avoid saying "have you read the article?" -- but: have you?

Do you think every copy of WinAmp included the Shoutcast _server_ in it? Hint: no, it didn't.

The whole point of the article is that what they shared _was not just the source of WinAmp_. That is what the article is about: that they mistakenly shared lots of other stuff that wasn't in fact part of the source code they were in fact _trying_ to share.

Taking a company's source code and preparing it for release to the wild is not a simple process. I've been there, done that for [1]OpenSTA

back in 2001.

IIRC, it involved checking the copyright of everything, removing author's names (some no longer worked for Cyrano, we didn't want people being bothered by support requests) and not least, removing any jokes or bad language from the comments.

[1] https://opensta.sourceforge.io/

CVS has been available since the '80s, and is the most approachable of any such tool.

If Winamp was going to use anything, they would have tried CVS first.

> CVS?

I doubt it.

Look, TBH, if they were that clueful in the first place, then this wouldn't have happened.

In the end, I wish this had gone better. The no touch license made anything out of the gesture a non starter for pretty much anyone who could do anything with the project.

The third party source references made the personality even more murky. In the end, there's not much that can be done.

I wish there was a rough modern equivalent player that didn't suck. I didn't know that there is any such beast. It least not with anything that resemble the community and ubiquity of winamp at its peak.

For that matter, I'd like a good Linux version.

I loved Winamp in the day, but my music library got far too big for it.

I'm using Gmusicbrowser on my Linux system now and it does everything I need it for, and doesn't freak out when presented with 1/4 million tracks, unlike many others I have tried along the way.

Have you tried Wacup? https://getwacup.com/

> I wish there was a rough modern equivalent player that didn't suck.

Foobar2000 is my audio player of choice on Windows, macOS and Android these days.

https://www.foobar2000.org/

By one of the co-authors of WinAmp, I believe.

Sadly, there is no Linux version and it's not FOSS.

So, then, what's the point?

Are they afraid somebody might take their baby and make a better version than they ever did?

Developers are still being forked. But then, Llamas are known for spitting in your face.

Which is what this amounts to: being expected to do the work but then hand it over without being allowed to distribute their own work was never going to get a lot of devs on board.

> So, then, what's the point?

The point is you don't get taken down or banned from GitHub.

A bit academic in hindsight, but there you go.

When I wrote this, the repo and the code was still there, making it easy to find the (literally) thousands of forks.

> Are they afraid somebody might take their baby and make a better version than they ever did?

Yes, I think exactly that.

Now, of course, legions of copies of the code are out there and perhaps new versions and ports will appear after all. With the names and serial numbers filed off.

I was trying to figure out what the difference is between a fork and a "modified version" and all I can think of is that they have to remove every reference to the name Winamp and then it becomes a fork. But I'm not a programmer of FOSS expert. It does seem like it doesn't matter one whit what their WCL said. They showed that they included GPL2 code so now anybody can do whatever they want with it according to the GPL2 terms (though the other code that Llama legitimately had no right to distribute is probably still off-limits).

If this did get forked/converted to FOSS, it would just end up like most other software of that nature: used by a few technologically-literate people, made ugly and hard to use in the name of giving users complete control and customizability so it can never get anywhere in mainstream use.

If that GPL2 code was linked, yes, maybe. If it was something else, that does not automatically apply GPL2 to the rest of it. It just indicates that they can't apply their proprietary license to those bits. For example, if that was a separate binary which they called to decode something, quite common in media players, their code is not covered by the GPL, explicitly or implicitly. If that code was included in the repo by mistake and isn't required to use it, not covered. Only if they took a GPLed library and included it directly in the application is the code of that application covered as well. Given that code for unrelated projects was coming through, the presence of that code in the repository is not enough to conclude that it covers the whole. You have to read the code to determine what part if any it touches.

I'd suggest that if the compiled binary for the application (ignoring other projects and material included in the repo that doesn't compile into it) contains anything that is compiled from GPL's source code that is included in the repo, then that GPL license applies to the whole. IANAL though, and I only get one lifetime to play with, so I'm not going to spend it reading and analysing the GPL license terms for fun.

In any git-based source control system (such as GitHub), and indeed pretty much any other source control system (how is Subversion doing these days? Anybody?), the word "fork" has a very specific and well-defined meaning. That will be almost certainly be the one they mean.

> They showed that they included GPL2 code

Er, no.

They included GPL2 code in the stuff they uploaded; that does NOT prove that WinAmp contained GPL2 code.

They accidentally included the source code of their proprietary audio streaming server as well. It is I think reasonably safe to conclude that did not form part of WinAmp.

> Any mass-market proprietary software project this old is likely to contain lots of parts from other companies, if only so that it could interoperate.

Llama Group SA is valued (today) at €15M. There must still be companies out there that have some of their code inside WinAmp. With documented source available as evidence, law suites should be a doddle for carving up the 15 million. Maybe even have a bash at Maxximum S.A. (> €1B) which is the majority holder of Llama Group .

Webamp is winamp for web browser

Has anyone ever considered why one would wish to whip a llamas ass?

https://www.youtube.com/watch?v=HaF-nRS_CWM

For those who don't understand that reference. Absolute gold. Take me back!

I've always assumed that was a Jeff Minter reference, given that he has a thing for camelids.

Maybe ask Jeff?

[1]https://www.polygon.com/23613576/jeff-minter-profile-akka-arrh-atari-llamasoft-arcade

[1] https://www.polygon.com/23613576/jeff-minter-profile-akka-arrh-atari-llamasoft-arcade

Aside from the hilarious debacle they've landed themselves in, this is a great future example of why "if they won't make a Linux port, just open source it and someone will." Very few companies own and have distribution rights to 100% of their sources.

Apparently one Winamp is worth one-tenth of one US penny. I'm not even sure if there is such a thing as a Winamp (only this site lists it) but they apparently were trying to do NFTs at one point. Seems like an appropriate price though for one copy of the software.

https://www.coingecko.com/en/coins/winamp/usd

This is the exact reason why IBM can't open the OS/2 code. I am not sure if they are unhappy about it.

"Never mind the llama, the cat is no longer in the remote geographical vicinity of the bag."

> This is what we come here for

:-) Thank you very much. Now grinning widely here.

I can foresee absolutely no problem with feeding someone else's proprietary and copyrighted code into an "AI" LLM, trusting that it will remove anything problematic, and not use any of the copyrighted material for its own training model. None whatsoever.

> I can foresee absolutely no problem with feeding someone else's proprietary and copyrighted code into an "AI" LLM, trusting that it will remove anything problematic, and not use any of the copyrighted material for its own training model.

Excellent plan. I endorse this.

Let's do it with Windows 95 and see what comes out the end.

Unless of course the answer to that is "Windows ME".

I think it is a very bad idea to give the Winamp folk a hard time for releasing their source-code in any kind of context. This bad rap will likely scare off future companies from doing the same. If people don't like something, perhaps just file a polite bug request in future. Don't turn this political. Don't turn this into a witch hunt.

The fact they had some GPL 2.0 code in there and witheld the source code is naughty, yes but lets be honest, I imagine 25%+ of proprietary code does similar and from much larger money-grabbing companies.

> I think it is a very bad idea to give the Winamp folk a hard time for releasing their source-code in any kind of context.

A fair point.

I did not mean to give them a hard time as such, merely to point out that they fairly heroically fluffed up the job.

But you are not wrong.