Richard Stallman on the Painful Birth of GNU
blogs.computerworlduk.com
blogs.computerworlduk.com
To me, this is really amusing to read on hacker news, given the rosy lenses we put on when it comes to getting a startup running from venture capital. I would go as far to say that most of us are building the business and enabling this kind of management/carrot-holding types to continue their behaviors.
So basically, there are both people who enjoy this kind of thought and people who vehemently oppose it on HN. We can see it easily in HN's comments.
That's fair. I have this romantic view that there's a brand of ethical/hacker startup, but that's, of course, without data to back up that not everything goes to shit when you have demands for meteoric growth.
Your computer is the factory. The binaries running on your computer are the means of production. You use your factory to do office work, for example writing documents, whatever. Copyleft is all about protecting the user that is running the binaries.
Now, if you control the source code to your computer, you control the binaries. This means that you control (own) the means of production, as far as a binary can be owned.
However, if you don't have the source code, I can sell you binaries but you have no mechanism to control your own means of production. Really, you haven't bought the software from me, you have bought a license to use it. It's as if you work in a factory, and you have to pay someone else for the privilege to use the machines in it.
Copyleft makes source code public among its recipients, meaning that control / ownership is shared, and you can still do what you want on your own machine.
Copyleft depends on copyright law to ensure public ownership; if all RMS wanted was a lack of ownership, as opposed to public ownership, he would promote public domain and BSD code.
Note that at no point here have I said that public ownership is bad - there are cases where it makes a lot of sense - but I do believe its not helpful to deny that the model is fundamentally socialist, as opposed to the EULA model which is fundamentally capitalist.
I still wouldn't call your conceptualization of Copyleft "socialism," because that (at least currently) has a strong connotation of government-involvement nowadays. I'd probably call what you're describing communism.
Anyway, thanks for listening. You could also argue that the source code was the means of production (of binaries, and then of work), which might be more accurate. I do not know much about Marx or the history of communism, but the focus on social concerns in the GPL is similar.
I get the analogy but I don't think it draws a reasonable comparison because it confuses the software with the license. The confusion is understandable because proprietary companies deliberately spread it. They want to blur the line between software and license so people have a harder time understanding their politics. But like you said, the software is not the product, the license is the product. Proprietary licenses aren't about restricting someone's ability to sell licenses, it's about keeping you from using their sales pitch. Copyleft is about making sure that no one can stifle anyone else's ability to use a certain sales pitch.
I'm not assuming I don't control my computer (it's mine to burn), I'm assuming I don't control the specific binaries on my computer unless I have the source code to them.
But, the amount of capital required to produce a binary of OS X or Windows is beyond anything you or I or the vast majority of users will obtain in the forseeable future, compared to a Debian binary which we could do today.
It's true that some people have definitely tried to write programs in their own languages with their own secret compilers just to make things even more difficult for potential copyright infringers, but thankfully that practice is not common because it has way more drawbacks than it does advantages.
Ummm, no, not even close. Speaking as a LMI employee during this critical period who then became for a few months one of RMS's roommates after leaving the company, right when he formally started the GNU project:
Yeah, LMI was with some angel investment able to build a run of CADR Lisp Machines just like the AI Lab did (25 vs. 75 or so that Symbolics did of its repackaged LM-2 version), but the company came to a dead end long before its next generation machine was ready. And went to Texas Instruments for money (TI even ended up buying Western Digital's engineering workstation unit, which LMI's processor was designed to work in).
For a couple of months the management said "We've cut your paychecks and put them in the safe, working is of course optional", we all continued to work and TI then made their investment (a loan followed by buying a 25% stake exactly 30 years and 1 month ago). But even then the company would have failed, for the hardware talent consisted of Greenblatt and a physics student who'd learned the basics of digital design on the job and they couldn't get the wirewrapped prototype(s) to work.
I even considered doing that (I was also a scientist type, could program and do system administration, was hired as the only person in the community with a foot in both Lisp Machine and UNIX(TM)), but I managed to recruit a graduating classmate who'd in his digital design class made a baby Lisp Machine (and I then left the company in frustration after about 9 months of never being in a position to do what I'd be hired to do and ran out of "make work" like sending a software update to our existing clients).
He then put in many "100 hours weeks" and with Greenblatt and the physics student (and perhaps some help from a couple of other friends I'd recruited) got the LMI-LAMBA processor working.
See http://web.archive.org/web/20110719154038/http://danweinreb.... as fsck--off linked to for the viewpoint of a Symbolics founder, although not from the AI-Lab (he took a detour to a LLL Navy supercomputer project), a very honest guy who's name was also on the Lisp Machine manual like RMS's.
I would just like to point out that comparably unethical behavior is, unfortunately, still commonplace at some schools (and again, not limited to computer science).
"By then, Symbolics had given everyone at the AI Lab an ultimatum. They said: We demand that you choose a side. I had been neutral, I had been more favourable towards Greenblatt's company, but not particularly involved with it in any way. But at that point they demanded that I, like everyone else in the Lab, choose a side. And I said, in that case, my choice is obvious, I'm against you. "
Wow! It's interesting that arrogant actions by various corporations are still pushing people towards OSS and away from closed source.
http://web.archive.org/web/20110719154038/http://danweinreb....
Edit (a partial summary of the above link but not a direct quote):
It is really hard to bootstrap a top of the line computer company, especially when you need capital to pay for your own fab. Greenblatt insisted on bootstrapping and being CEO too, and many hackers at the Lab did not think he would be a good CEO. It had nothing to do with "backstabbing".
Symbolics hired AI Lab hackers full time because other AI spinoff companies were inappropriately using AI Lab resources, so Symbolics wanted to make sure this didn't happen. Hiring workers full time wasn't an "evil plan" to destroy the AI Lab, it was actually a well thought out ethical course of action.
Since MIT owned the source code to Lisp Machines (read: it was copyrighted and not free in any way), they licensed the software to several companies under the conditions that all proprietary changes could not be distributed to anyone else except MIT. AI Lab hackers got all the changes that came from a private company, so they didn't need to leave the lab to work on Lisp Machines. The AI Lab hackers had free will, too. No one forced them to leave, and Stallman's suggestion otherwise is actually really offensive. In fact, if LMI had been the only company, the same people would have left the lab.
Stallman thought that Symbolics's compliance with MIT's licensing agreement was actually an evil plot. His essentially wanted Symbolics to support LMI by giving them proprietary changes. If Symbolics had actually done this, they would have been shut down by MIT for breaking their licensing agreement. The same was true for LMI, which also had to make any changes proprietary. Stallman should have been angry at MIT, not Symbolics.
Stallman completely misunderstood this and his offensive (and quite possibly libelous) slander against honest and ethical Symbolics employees is a very off putting.
Symbolics was a huge company and developed an entire OS, architecture, and many programs for their platform that Stallman did not and could not copy. Stallman mainly copied some features from Zmacs (the text editor). Symbolics made high end machines that sold for tens of thousands of dollars. Nobody would have ditched an expensive Symbolics machine, one of the leading CG platforms of the 80s, because some other platform also an editor that sort of looked like Zmacs. Stallman had pretty much no impact on Symbolics's revenue.
Back in those days the serious CPUs were made from discrete logic, so both companies outsourced the construction of boards and I assume assembled them from there. ASICs came later (while I was at LMI TI consulted us about building a system on one or more of their gate arrays).
RMS did a bit more than copy new Zmacs features; Dan acknowledges "What he copied were incremental improvements: a new editor command here, a new Lisp utility there." and I e.g. remember something about a new condition system or the like.
A constant theme with RMS is his complete misunderstanding of basic business and economic principles.
I can't tell if Poe's Law is at work here. For all of RMS's 'misunderstanding' he's been so profoundly right that practically every single piece of hardware on the internet is a product to one degree or another of his understanding of economics and business in the internet age.