When Free Software isn't better | Hacker News Reader

When Free Software isn't better

mako.cc

61 pointsvu3rdd45 comments

The author spends most of the article harping on the fact that most open source projects don't have any collaborators. The vast majority of software, open source or not, is completely irrelevant. Any conclusion drawn by talking about irrelevant software is also irrelevant.

I don't know why you were voted down, because yours is the most relevant point about this article.

It doesn't matter that 90,95,99,whatever% of sourceforge projects are dead audio mixers that not even the original developers are using. Those projects can be ignored. They don't take up physical space, and we're not running out of internet. Just ignore them.

What matters is the quality of the actively developed free software. Any study on the quality of FOSS should start and end with actively used and developed software and ignore those that have been selected out of the software pool, because that is the process of FOSS development.

absolutely agree. It seems as if the author is saying all open source, versus the small set of proprietary software.

I know I have lots of 'open source' half finished projects, which I may or may not be using, and it doesn't really matter to me if someone else wants to take the code and continue it, or fork it or what have you.

I certainly prefer for my web server an open source stack. That's not to say there is anything wrong with proprietary alternatives, but I worked for a very large company that transitioned from proprietary software to open source, precisely for the 'open source' credo (it's stable, tested, more secure etc) as well as its gigantic cost savings annually.

I think proprietary software has a place, but I would like to see the author argue when and what those cases are instead of just talking shit on free software. Perhaps desktop applications for MacOSX and Windows is such a case...

But I don't think that is a correct reading of his argument. His argument is that each one of those dead, undeveloped projects adds value to the free software world, even though they may have no positive impact on the open source world.

It isn't irrelevant if you're a proponent of open source (but not necessarily free) software, and want to argue its main advantage is better software through collaboration. That argument is challenged by the large amount of "irrelevant" and inferior open source software, because it suggests that better quality software isn't always a consequence of going open source.

Free software advocates don't have to fight that particular battle, since software freedom for them is the end, and not a means.

That's a terrible strawman though. By being open source the quality could be improved by anyone at any time. That software could be proprietary and it would be equally poor and irrelevant.

It could be, but it often isn't. One could imagine someone arguing against open source software, saying there isn't enough interest or talent to maintain all of the open source projects in existence. I think the article we're commenting on was written in response to that argument.

I agree with you about proprietary software, but that model is rarely attacked in the way open source and free software is.

Well if someone came to me with such an argument I would say it's silly to judge open source software by the long tail of least-successful projects. I mean no one chooses anything by focusing on the worst representatives of a category. People aren't evaluating the least-maintained open-source software any more than they are evaluating all the dead and gone software to be found in bargain basement bins.

There's plenty of great free, open-source, and proprietary software out there, and many interesting debates to be had about the merits of competing packages. If some pointy-hair insisted on an obtuse argument about how all the crap projects on SourceForge are diluting the power of open source, I would get another job and let him replace me with someone with a higher tolerance for incompetence.

Sturgeon's Law: 90% of everything is crap.

The more important metric is to measure is how good the top 10% (of each) is.

I've met the author, and considering he's one of the founding members of Ubuntu I would read this article in a different light than most of the comments on here. I think it's more of a call to action to make us think not just what Open Source can do for us, but what we can do for open source.

BTW: http://en.wikipedia.org/wiki/Benjamin_Mako_Hill

Mako's also on the board of the FSF and the Wikimedia foundation.

Why can't they call it LIBERATING SOFTWARE?

Seriously, it's not the software that's free (as in libre), it's you, to do as you want as long as others who receive your derivative software get the same freedom.

Liberating -- to bring about liberty to those who do not have it

It makes much more sense than "free" software. Why not take the opportunity to have everyone exposed to the name understand exactly what it means and be more intrigued?

In the business world, saddling yourself with higher support costs and lost productivity from lower-quality software would hardly be considered liberating. The quality issue is still relevant.

Well yea, but I am assuming that free software advocates are ADVOCATING for free software -- so if they call it liberating software, their task will be much much easier.

IMHO , changing names will create confusion among naive users. Example openoffice-> Libere office. We have been using the term "Free software" for years.

Yes, and people still don't know what it stands for.

That should've been a wakeup call for free and open software advocates since at least 10 years ago..

EXACTLY!!!!!!!

Software quality is absolutely orthogonal with software freedom. Did I misss something or is your comment irrelevant?

Software quality, documentation, testing is tedious and non-sexy for the most part. Because contributors to free software projects don't have to behave like customer facing pros they, for the most part, don't and you end up with poor UIs, non existent documentation and inadequate testing. The few exceptions (like GCC) merely underline how shoddy and amateurish much of the rest is.

So in a definitional sense they may be orthogonal issues, in a practical sense they most certainly are not.

Some free software is part of a commercial product, and is usually very well documented and reliable. Think WebKit, Firefox, Rails, and Google's many open source contributions.

Yes, and that just reinforces what I said. Free software done by pros is, mostly, done to professional standards, that done by amateurs is mostly done to amateurish standards.

What defines a professional or amateur, though? If you define a professional as someone that contributes to software as part of their paid work (and an amateur as anyone else), then Linux and GCC were both begun by amateurs, and are now maintained partly by professionals. Likewise, many projects are begun by professionals and receive contributions from amateurs.

That depends on whether you consider freedom in itself to be a quality or not, and if you do, how much weight you give to the freedom quality.

Then, the question is whether a certain proprietary software as enough other qualities to make up for the missing quality of freedom.

Also, freedom is not a pure ethical quality. It means greater independence as well as freedom of action, which both can be very valuable (in dollars), for companies even more then for single users.

But "Free Software" doesn't liberate you as much as "Open Source Software" - with BSD licensed "open source" projects, I can literally do whatever I want with the code.

With GPL licensed "free software" projects, I am not allowed to distribute in certain ways, locked out of packaging with certain other licensed code in certain ways.

>> Seriously, it's not the software that's free (as in libre), it's you

Except that it's the exact opposite

> Except that it's the exact opposite

You're wrong. GPL makes you free like human rights do: You are free to do whatever you want as long as you don't step on other people freedom. With the GPL, you can do whatever you want with the software except prohibiting users to do whatever they want with the software, with BSD you can do whatever you want with the software including prohibiting users to do what they want.

GPL promotes freedom, BSD does not.

> With the GPL, you can do whatever you want with the software except prohibiting users to do whatever they want with the software

The GPL can prohibit me from distributing the program modified to depend on proprietary closed-source code, even if that's what I want to do with the software.

> with BSD you can do whatever you want with the software including prohibiting users to do what they want.

Exactly - the point is that BSD doesn't limit me at all, and the GPL limits me fairly severely. The GPL guarantees that the code under the GPL and all its derivatives will remain free forever - the code is free. The BSD license guarantees that I have the freedom to do whatever I want.

> Exactly - the point is that BSD doesn't limit me at all, and the GPL limits me fairly severely.

But the BSD allows you to limit what your users can do, the GPL does not. Again, this is how all our democracies and human rights work.

No, it is most certainly not. In most, if not all, democracies, you have the right to restrict the freedoms of others through contract. For instance, if you want to buy a piece of furniture from me, I can impose conditions on how you use it. This is not common practice, because it isn't feasible to enforce and most people selling furniture don't care. However, it wouldn't be a violation of your human rights. Democracies allow us to limit the freedom of others as a condition of them engaging in commerce with us. No human rights are violated.

I never said human rights are being violated. I'm saying that human rights work the same way by allowing you to do whatever you want as long as you don't step on others freedom.

All of the BSD licenses, and also the GPL, are free-software licenses, as well as open-source licenses. If you want to argue about which particular licenses are more or less free or open, you should do it in a separate thread, ideally on a separate site, instead of trying to derail this thread.

Actually, the way I think about free software is that the GPL gives the software unalienable rights. Free software cannot be enslaved or killed, and neither can its progreny. BSD licensed software, on the other hand, is free, but anybody who pays for the upbringing of its progeny can enslave or kill that progeny at will.

I don't think this view matches reality. For example, imagine that I fork and modify some GPL'ed software and only distribute it within my company (or just use it myself). I'm not required to distribute source to anyone other than the users of my modified program, so (to use your term) I have "enslaved" my version of the software. I think it's clear that "free software" refers to the freedom of the users of that software.

(disclaimer: I am not an advocate of the "free software" moniker/agenda although I happily use and contribute to open source software and don't object to judicious use of the GPL.)

You are right, but I still find this imperfect analogy the best way to 'explain' why the FSF would attach the adjective 'free' to 'software'.

I also wonder what the FSF's response would be if such "within one's organization" use ever would become the predominant way some commercial entity used some GPL software package. I think they would classify that as a hole that needs fixing.

Edit: of course, I may be overlooking that such cases already exist, or that the FSF discussed this (I did browse the GPL FAQ to check the latter, but did not spell it out)

As a side note -- can anyone here shed light as to what the GPL says with regard to developing code based on the GPL within your organization? I mean, I would want my employees to NOT necessarily go and publish intermediate stages to the world or share them with my competitors. But then again, I am restricting my employees' freedom to do so, through a non-compete contract or something else. Is this compatible with the GPL? I guess this hinges on whether the employee installing a copy of the code in their development environment counts as "distribution". Does it?

I blame Sourceforge rather than the Free Software community. Take a look at my 150+ projects on Github; 90% of them have at least one commit by someone other than me. Look at the ones that have more than 10 watchers, and they all have commits by several people.

This is because Github makes collaboration and attribution easy. Sourceforge? I didn't even know it existed past 1998.

Wait a few years, and thanks to Github you'll see what the open source community is really capable of.

I think evolution of open source software resembles evolution in biology. There are many dead or one developer projects, but like in biology the projects that are actively developed and have many contributors are usually the strongest ones. They survived the competition with hundreds of other projects before they became what they are now.

From the article:

Free software is sometimes low quality. It is sometimes unreliable. It is sometimes inflexible.

Swap "free" for "proprietary" in that quote, and the statement is no less true. (Voice of experience, here) Continuing:

If people take the arguments in favor of open source seriously, they must explain why open source has not lived up to its "promise" and conclude that proprietary tools would be a better choice.

No, they need not. If even a handful of free or open source software meets the standards of quality one sets for proprietary software -- and I would argue not only that such a handful exists, but the so-called "standards" must be rather low to pass some of the proprietary software I've dealt with -- then open source has indeed kept its promise. Nor does the existence of some "bad" open source software imply that proprietary is better. If that were the case, we must also accept the reverse, that the existence of some "bad" proprietary software implies open source is better. Examples exist in both categories, so clearly each is better than the other.

Here's the deal: An open source package, once "in the wild", can become very hard to kill if it gains any traction to speak of. As long as someone keeps the source available, anyone else can not only acquire and use it, but build on and improve it. Not so, proprietary. With the source kept to the strict confines of a few personal or company machines, its growth and development are limited to what its owners see fit to pursue or permit, and they can choke it off at any time. As a consequence, we tend to see more "bad" open software roaming the net as abandoned projects are left to anyone who cares to pick them up, while proprietary packages that either don't make the grade or lose their luster meet a quiet demise behind closed doors. That's what makes proprietary software look better: the "bad" stuff just gets killed off more effectively.

I agree with author's observation. My personal open-source projects are developed 95% by me, and sometimes overhead of collaboration (documentation, discussions) outweighed the contributions I've got (small patches that I had to refactor anyway).

However, this should not be interpreted as a reason to drop free software or that closed/for-profit software would fare any better. I don't think that the "long tail" of projects in sourceforge graveyard would attract enough commercial interest to fund several developers.

The surprising conclusion of this article is that free software advocates have lower expectations. It sounds like for them every project released under a free license is a success, and features like quality, reliability, and flexibility are desirable but secondary goals.

Argh, are these old farts still making grunting noises? How many important projects started in the last 5-7 years are GPL licensed vs Apache/BSD/MIT? The battle is over, and the copyleft zealots lost. Time to move on.

This is all such bullshit.

If these folks spent 1/10th of the time they've spent on ideological wars doing something more productive, FOSS would be a heck of a lot better, especially for end users.

> If these folks spent 1/10th of the time they've spent on ideological wars doing something more productive, FOSS would be a heck of a lot better, especially for end users.

People down-voted your comment but I upvoted it. I agree, to a certain degree. I don't find it useless/wrong to use software for political purposes: the author is obviously trying to make a statement. I find it fascinatingly bizarre that the emphasis is so strong on the political expression of free vs open, though, instead of the technical merits of quality.

Thank you :)

Reply on news.ycombinator.com