Bill Gates calls Linux/Open Source/Free Software Communist! accurate description or not?

http://news.com.com/Gates+taking+a+seat+in+your+den/2008-1041_3-5514121.html

I think "communism" is a daft word to describe this, but I can see that it's a significant change to how things might work in business; likely to be in favour of larger, more established firms I guess. So for example, if I have a great idea about search technology, then google can just copy it without paying me anything.

And why do you think that is?

Or is it that obvious; perhaps I'm being stupid.

I think sometimes that Gates' phenomenal early successes may stem from exactly this kind of thing, the CES disaster the other night, and other things like this - he's got such an utterly nerdly demeanor and misspeaks so frequently people probably underestimated him completely. I mean would YOU think he had any business acumen?

For example, Microsoft held patents that, they claimed, covered the FAT filesystem used in DOS and Windows 9x, and used on just about every floppy disk, memory card, etc, in the world. They had started to claim royalties from memory card manufacturers, until said patent claims were overturned in court.

Also: Microsoft and their friends - SCO, for example - often try to portray Free Software as somehow "anti-copyright", when the Free Software licenses rely on copyright law just as much, in a way, as other software licences. Free licences effectively say: "you must accept these terms, because if you don't then under copyright law you have no rights to a copy of this software at all".

I do believe they do this. They also do something I can't remember the exact term for, which is encircling patents, i.e. someone patents an idea, and then someone else patents every other conceivable idea closely related to it in practical use terms, hence reducing the scope for the individual's idea to come to fruition.

It's clear there's a lot of design goes into a file system. Where would you draw the line? Or do you believe that everyone should be allowed to nick everyone else's stuff regardless of who invented it?

With respect to free software and copyright, I think the problem is that these terms (particularly "free", which seems to only mean free in the way I understand it in BSD licenses) simply don't give the whole picture; saying someone is "anti-copyright", means very little; I mean, you can see kind of where you're coming from, but as you rightly point out, all of them come with copyrights.

I guess this moves into the arguments around digital music/film etc!

I actually see GPLd software as less of a danger than other open source licenses to business, in that largely, businesses cannot use GPLd software unless they use them as is, like using Emacs for example. Libraries cannot be used under GPL unless you want to open source all the code that is using the library; something I suspect almost all companies would not want to do.

Open source software is clearly a terrific thing and puts the thumbscrews on to business to make them perform as well as providing free and (in some cases) highly reliable, well-written alternatives. It seems likely that IE7 will turn up soon enough as a result of the gains made by Firefox recently.

If they had patented FAT back in 1981, and had insisted on licensing it out to other companies from the start, I would probably think differently. The software market would have developed very differently, too.

(I'm not sure what exactly was patented in FAT - I don't agree with algorithm patenting, but I would consider that some aspects of a filesystem design could be patented)

Libraries cannot be used under GPL unless you want to open source all the code that is using the library

Where do you stand on the LGPL license?

LGPL is absolutely fine; that removes the aforementioned restriction.

I like GPl, but I understand the wariness of it. The BSD license is good because it is simple, unrestricted and doesn't have any ideaology behind it. (believe me it was much asier to persuade my company to accept PostgreSQL and it's BSD licence than MySQL and it's funny mix of GPl and commercial licences.

I can't really remember the order of things, it was 1987 or so the last time I used CP/M, but I suspect all sorts were copied between these operating systems.

Do IBM hold a patent for virtual memory I wonder? I mean that sounds like a reasonable thing to patent, and yet it did it first; now all OSs use virtual memory.

Bill Gates and "Modern-day sort of communists"

Completely agree. In fact (particularly with the forthcoming windows version) I suspect PostGreSQL will overtake MySQL as the number one open-source database in time, basically because of this.

(I assume it's what used to be Ingres... Never been able to confirm that anywhere)

I personally agree about LZW, which is way too obvious an approach to be patented. Lord knows how you legislate for that sort of thing though. Perhaps you're right, it may just be a case of shortening the length of time involved.

Cheers; that clears that up! What about hierarchical file systems, databases, VSAM, ISAM, time sharing etc. Were these all patented by IBM too? What's the score, did the patents run out?

This is inccorrect. So long as a business is not planning on distributing a modified version of GPL'd software, it can keep the modified source code to itself. Upshot--if run a business that wants modify emacs for *internal use only*, I am not required to release the changes under the GPL.

For more info, check out:
http://www.fsf.org/licenses/gpl-faq.html#GPLRequireSourcePostedPublic

Thanks for that; that's obviously been a long held misconception of mine. The definition of "distribution" still concerns me a little though and what the implications are for publicly available web sites.

Caitlin...

Looks as though Ingres and Postgres are related but not the same thing.

http://databases.about.com/od/postgresingres/

There's some company currently trying to do the same to JPEG as Unisys did with GIF. Not sure how well they're getting on with their lawsuits though.

Richard Stallman is apparently rather concerned with this 'loophole' in the GPL, so the upcoming GPLv3 will probably include language to specifically prevent GPL code from being used in web apps without the source being available. So BSD/LGPL is still probably the best way to go.

Hey, just fork the GPL and refuse to use the new version.

Anyway, RMS is not the FSF, and the FSF is not RMS, so I'm not sure that what RMS apparently thinks is a 'loophole' will really end up being 'closed' in the GPL revision. I personally don't see any reason why someone using modified database software and having the *results* of that use available to the public should be considered a 'distribution' of that software. How different is that from using internally modified software to create a physical item and then offering the physical item for distribution? RMS wouldn't see that as distribution of the software.

What do you mean by 'web apps'? Are we talking about server-side our client-side? I think there's an important distinction there, with major implications for how this should be analysed, and I would bet, what RMS is talking about.

more fun and a 4-part interview