The Torvalds Transcript (InformationWeek)

Posted Mar 21, 2007 14:56 UTC (Wed) by zooko (guest, #2589)
Parent article: The Torvalds Transcript (InformationWeek)

There's a middle path here. You don't need to sign up on a proclamation that proprietary software is immoral, but you can still believe that, all other things being equal, Free Software is better for the society that depends upon it than proprietary software is. There's no need to go to an extreme stance in order to value Free Software as such -- for its freedom in addition to its usefulness.

Likewise, the other extreme is the belief that any constraint imposed by a software licence is immoral and that all software should be useful enough to survive under a permissive (BSD-style) licence. This is an unnecessarily extreme position -- imposing GPL-style constraints can help produce more and better software, and this is one of the reasons why Linux is so successful.

If you read the entire interview with Linus Torvalds it would seem that he too follows a middle path, but if you read just the out-of-context sentence quoted in this LWN article it would seem that he is being extreme.

Regards,

Zooko

to post comments

The Torvalds Transcript (InformationWeek)

Posted Mar 21, 2007 15:36 UTC (Wed) by TRauMa (guest, #16483) [Link] (40 responses)

It's interesting that that's the first time I read something pragmatic from Linus about the GPLv3. The points he has about license splintering are valid, IMHO, but then again what you get by combining multiple GPL offerings with different added restrictions is the GPLv3 with all restrictions applied. So when releasing something under the GPLv3 you should perhaps be comfortable with all the possible restrictions it contains, as code flowing back to you may have them. Seen this way there is no splintering at all, or at least not more than with GPLv2, where you can add any permission you like, too, but code coming back may not have those permissions anymore.

The other point he has about market forces sorting out the closed vs. free development models may be true, but I don't really believe him when he says he doesn't care about it politically. If tomorrow closed source would just crush free software and he would be forced to live with an IT infrastructure that wouldn't cater to his needs, but those of the vendor (or to live the digital equivalent of a hermits life), he'd mind it _very_ much. I see ignoring DRM issues as one way to get there.

The analogy between CS vs. FS and Alchemy vs. Chemistry is a nice one, it alone made it worth reading the article. Of course, with closed source, the alchemy actually works to some degree, and can be harmful to the user.

The Torvalds Transcript (InformationWeek)

Posted Mar 21, 2007 16:44 UTC (Wed) by anonymous1 (guest, #41963) [Link]

Closed Source in todays world is not just alchemy, it alchemy along with the coercion of copyright.
Public domain binaries without the corresponding source code is analogous to alchemy. But copyrighted binaries are not a good analogy to alchemy.

The Torvalds Transcript (InformationWeek)

Posted Mar 22, 2007 9:35 UTC (Thu) by malor (guest, #2973) [Link] (38 responses)

I remain convinced that the bad blood between Stallman and Linus over 'GNU/Linux' is at the core of Linus' objections. I don't think he would be attacking the GPLv3 so fiercely if he weren't personally invested somehow.

First, you see him talk about 'motives'; that's a code phrase for 'what Stallman is trying to do'. Most of what he's saying here isn't even arguing the technical merits of the license, he's saying, obliquely, that it's unwise to trust Richard Stallman. (and considering the GNU/Linux imbroglio, he's not entirely wrong to do so.) Most of his explicative text covers variations of this theme.

When examining his specific objections, it's obvious they come from the standpoint of the single most successful Open Source project; at this point, the kernel devs get more personal fame and fortune from broader Linux use, and to get that, they seem willing to trade away future freedoms. It's okay, they argue, for manufacturers to use their code to give us devices we can't change or use in ways of which the manufacturers don't approve.

It appears the devs don't want this restriction added because it will interfere with their own success; they are thinking of their own pocketbooks and fame. Closed-device manufacturers, after all, often hire people to work on the kernel code, increasing the 'jobs for Linux kernel hackers' market. The devs are not, unfortunately, holding end-user rights paramount, one of which is, and always has been, the right to use devices freely that are dependent on Free Software.

If you can't remove and replace the code on a device, but the manufacturer can, particularly without your permission, you don't truly own it. The code has been hijacked. You are enslaved by that code and dependent on it. (see: Microsoft Vista.) Stallman is trying to prevent this from ever happening with anything under the GPL, and he is right to do so.

Linus et al are clearly in the wrong on this issue. They appear willing to trade away our freedoms for their own success, and I'm saddened by this.

The Torvalds Transcript (InformationWeek)

Posted Mar 22, 2007 12:43 UTC (Thu) by jospoortvliet (guest, #33164) [Link] (4 responses)

I'm totally with you on this. I think the freedom to RUN the software is as important as the 'you have to share the source' part of it. The world could fully consist of Open Source Software (GPLv2) running on paladium/TCP hardware, so nobody could do with it what they want, except the big hardware and software vendors. I'm pretty sure Linus wouldn't like that, but apparently he's not worried, doesn't care, or doesn't think it's possible.

The Torvalds Transcript (InformationWeek)

Posted Mar 22, 2007 15:45 UTC (Thu) by zlynx (guest, #2285) [Link] (3 responses)

I don't think that people are really worried that *all* computing devices will be locked down with TCPA or whatever. It always sounds to me that they're worried cheap and easy computers will be locked down.

Does anyone really believe FPGAs, 8 bit micro-controllers, embedded ARM and MIPS chips will *all* be running trusted computing?

Heck, someone with time on their hands could breadboard a 6502 or 4004 chip. It'd be big, ugly, expensive and slow, but it's a computer.

It would never be "nobody could do with it what they want". It'd be "nobody could do with it what they want for less than $5,000".

I guess I'm in the "not worried" and "don't think its possible" camps on this one.

The Torvalds Transcript (InformationWeek)

Posted Mar 22, 2007 18:29 UTC (Thu) by mmarq (guest, #2332) [Link]

"" Does anyone really believe FPGAs, 8 bit micro-controllers, embedded ARM and MIPS chips will *all* be running trusted computing? ""

Absolutely. Its a good way to prevent tampering with some OEM product. And those microcontrollers will have the TCPA in the VLSI design, because at 32nm node they would be too small to be only 8 bit FPGAs.

TPA everywhere?

Posted Mar 25, 2007 2:49 UTC (Sun) by kevinbsmith (guest, #4778) [Link] (1 responses)

With the proper political climate, it could become illegal (in some/many countries) to sell a non-TPA computer. Or to build one. Or even to own one that was built pre-TPA. Hopefully that won't happen, but I can't see any reason right now that it absolutely positively could not happen in the next ten years.

If we have to live in that world, I would still want a thriving FLOSS ecosystem.

TPA everywhere?

Posted Mar 25, 2007 4:44 UTC (Sun) by dlang (guest, #313) [Link]

sure, and it could become illegal to own a car, illegal to drink, illegal to smoke, illegal to be unemployeed, illegal to have a job if you are a Jew, etc

stupid things happen. if they happen in one country, that country will suffer and fix itself or go down the drain.

and the probability of non TPA computers being illegal is infintesimal (just like everything else I listed above, there are people who want it)

if you think about the cost of implementing a full TPA environment, and consider adding that to everything that's run by computer (down to your toaster, coffie maker, vending machines, cars, etc) the cost of forcing all new products to implemtn TPA would scuttle it, and if you look at how it's basicly been impossible for the HD-TV to replace all the TV receivers around the world in anything resembling a reasonable timeframe, what makes you thing that there would be any more sucess in replacing all these other devices?

easily programmable computers are too much a part of the worlds economy for any country to sucessfully outlaw them, even if you ignore people's home PC's

The Torvalds Transcript (InformationWeek)

Posted Mar 22, 2007 13:31 UTC (Thu) by paulpach (guest, #20903) [Link] (17 responses)

> If you can't remove and replace the code on a device, but the
manufacturer can, particularly without your permission, you don't truly
own it.

What about when you dont own it?. A company I worked for LEASED appliances
that run linux and gave the source with it in full compliance with GPLv2.
To ensure the safety of our customers, we used DRM to
restrict what the appliance could run. This was a SELLING POINT because
the customers did not have to trust anyone else but us. We where clearly
distributing software and that would violate GPLv3 if it looks anything
like the drafts.

BTW, if you drink the kool-aid, GPLv4 should restrict some other uses that
restrict user freedom:
* Code shall never be used in ROM because you can never run your
modifications in it's intended device (looks a lot like DRM doesn't it?)
* Code shall never be used to run jails
* Code shall never be used in weapons ( if you are dead, arguably you are
no longer free to run the software)
* Code shall never be used in security system because people are no longer
free to enter the building
etc...

The Torvalds Transcript (InformationWeek)

Posted Mar 22, 2007 18:08 UTC (Thu) by malor (guest, #2973) [Link]

I'm not sure how it would apply to a lease, but with a normal sale, you'd be required to give them a master key; they don't have to exercise their right to modify the device running Free Software, but you have to make sure they can do so. You can't reserve rights to yourself that your downstream clients don't have.

No, you as a business can't do everything you want in that situation. That's your cost for using Free Software. If you want to give your customers devices where you have more rights than they do, find code that's under a more permissive license. It's really that simple.

As far GPLv4 not being usable for jails and wars and crap... that's obviously a strawman. This modification of the GPL is merely to make sure that downstream users get all the same rights as everyone else... that no matter where you are in the chain of code transfer, you can do everything that every other person in the chain can do.

That is the spirit of the GPLv2; version 3 simply extends it to provide additional freedom, so that everyone is empowered. You can't hold people hostage with hardware if it's running GPL3 code. Period. If you want to, use code that's licensed more liberally.

The Torvalds Transcript (InformationWeek)

Posted Mar 22, 2007 18:40 UTC (Thu) by mmarq (guest, #2332) [Link] (15 responses)

"" We where clearly distributing software and that would violate GPLv3 if it looks anything like the drafts. ""

Well... not necessarily

Private contracts are private, and if a costumer trust a vendor with a DRM implementation, that is ok... what the vendor shouldnt be able to do with GPL, is to push a DRM scheme down the trough of every possible user...

The Torvalds Transcript (InformationWeek)

Posted Mar 23, 2007 4:27 UTC (Fri) by drag (guest, #31333) [Link]

Well with this sort of DRM there is absolutely no need to give anybody the keys besides the people who own the machine.

That is why things like 'can't use GPLv3 with voting machines' is such a red herring.

(assuming the draft makes it through)
Yes you can use GPLv3 in voting machines and YES you can use DRM to protect those voting machines from tampering.

Same way you can sell machines to people and use DRM to protect those machines from being tampered with.

The deal is is if your customer wants the ability to change the software in there you have to give them, nobody else, a way to modify it. If they don't want to modify it and if they don't want to disable the DRM then they don't have to.

It's the end user that gets put into control of the hardware and the software not the manufacturer.

So with voting machines, since they are owned by the government, nobody but the government needs to be allowed access to it. And they can regulate it how they please since the machines belong to them. Which is exactly how it would be any other way GPLv3 or not.

So it's a pretty spacious arguement that your going to use DRM for end user's own good. If they want trusted computing or other such things to help make their systems more secure then there is absolutely nothing in the GPLv3 that prevents them from taking advantage of TPM or whatever.

Remember that the goal is here is to allow the end user control over the software.

Not the hardware.
Not non-GPL code.
Not DRM'd media.
Or anything like that.

All that is perfectly legal and if you invent a way for DRM to work while having end users being able to understand how the programs work and be able to modify the code, then it would be perfectly fine to use GPLv3 licensed code.

It's not designed to prevent that or limit that stuff in any way.

GPL and private contracts

Posted Mar 23, 2007 22:48 UTC (Fri) by giraffedata (guest, #1954) [Link] (13 responses)

Private contracts are private, and if a customer trusts a vendor with a DRM implementation, that is ok...

No, it's a violation of the license conditions under which the vendor acquired the code. The original author's copyright is violated.

If you have an enforceable promise from your licensee not to avail himself of the "license," then you haven't actually given a license. A condition of GPL is that you give an actual GPL license when you redistribute the code.

That's the weird and rather unique thing about GPL: it interferes with private transactions that don't even involve the copyright owner. It's meant to be a social engineering device, not a commerce tool.

GPL and private contracts

Posted Mar 23, 2007 22:52 UTC (Fri) by malor (guest, #2973) [Link] (12 responses)

... "it interferes with private transactions that don't even involve the copyright owner." ...

If you're using my code, it involves me. If you want to give my code to someone by loading into a device, that's fine, but the price for doing so is that they get the same rights that you have. You can't lock them out of their hardware, because the hardware is a necessary component to running the software that you licensed from me.

If you don't like that price, you can pay for other code with money.

GPL and private contracts

Posted Mar 23, 2007 23:49 UTC (Fri) by giraffedata (guest, #1954) [Link] (11 responses)

My statement is (for the sake of simplicity) over-broad. One could write pages on what it means for a transaction to "involve" someone else. John Donne said (The Tolling Bell) even the death of a stranger damages you, because you're both part of Mankind.

But the fact remains that the level of involvement you have in a transaction between two people using your code somewhere down the chain is fundamentally different from the kind of involvement most private legal acts (licenses, contracts, deeds, etc.) care about.

How many copyright licenses other than GPL seek to give value to some total stranger down the chain who will give the copyright owner nothing in return for it except the warm feeling of knowing Right has prevailed?

Missing this special quality of GPL is what leads people like the poster above to believe that GPL couldn't possibly stop a recipient from voluntarily giving up his rights (in return for something of value from his distributor).

GPL and private contracts

Posted Mar 24, 2007 1:12 UTC (Sat) by malor (guest, #2973) [Link] (10 responses)

The purpose of the GPL is to keep the code free and open for everyone. No matter where you are in the chain of transfer, you have the same rights and responsibilities that everyone else does. GPLv3 just shuts down some methods of removing rights from downstream recipients.

Anytime you're transferring a copyrighted work, the original creator is 'involved'. You have the right to transfer your only copy of something. You do NOT have the right to copy it and keep a copy for yourself.

With ordinary copyrighted code, you have to pay money for copies; Microsoft is just as 'involved' as a GPL author in that case. Or, you can opt to pay with freedom; you give up some of yours, and you increase the freedom of the transferee. But it's the same fundamental transaction; you are buying the right to use the code.

You do NOT have the right to give someone a copy of Windows without paying for it. By your assertion, Microsoft is 'involving itself' in a 'private transaction' between you and a customer. You can't legally steal Microsoft's code. Why should GPL code be any different?

GPL and private contracts

Posted Mar 24, 2007 2:51 UTC (Sat) by giraffedata (guest, #1954) [Link] (9 responses)

I wouldn't say Microsoft is involved in the transaction per se, but rather that it is involved in the copying that is part of it. And in contrast to the GPL publisher, Microsoft wouldn't care about the details of the transaction -- who paid whom for what. Unlike a sale of Windows from Microsoft to a Microsoft customer, where Microsoft is involved in a whole different way and is deeply interested in the terms of the deal.

I really don't know what battle you're fighting; it sounds like you're defending the use of GPL, but nobody has attacked it.

We agree the copyright owner is always involved. We agree the purpose of offering code under GPL isn't to selfishly protect the author's wealth, but rather to bring about a social goal. We agree that it's perfectly legitimate for a person to offer, and accept, code with GPL restrictions.

The only thing I'm saying, which I don't see how you can disagree with, is that GPL is special (I also used the term "fairly unique") in its effect on downstream copiers.

Maybe I could say, "unlike conventional copyright licenses, GPL interferes in transactions in ways in which the copyright owner doesn't have a material interest."

GPL and private contracts

Posted Mar 24, 2007 13:59 UTC (Sat) by vmole (guest, #111) [Link] (2 responses)

The GPL is special in its effect on downstream copiers because it's very rare for licenses to *allow* downstream copiers.

GPL and private contracts

Posted Mar 24, 2007 17:29 UTC (Sat) by giraffedata (guest, #1954) [Link] (1 responses)

The GPL is special in its effect on downstream copiers because it's very rare for licenses to *allow* downstream copiers.

The license doesn't allow downstream copiers, but the copyright owner does, by issuing additional licenses to them. Same for GPL and non-GPL.

But look at what the "stream" is: The stream is the modification of the software. A writes code; B adds stuff to it and passes it on to C, C does the same to D. With non-GPL, A insists on royalties for the C-D copy, but otherwise A usually doesn't care what the deal between C and D is.

GPL and private contracts

Posted Mar 24, 2007 18:09 UTC (Sat) by malor (guest, #2973) [Link]

> With non-GPL, A insists on royalties for the C-D copy, but otherwise A usually doesn't care what the deal between C and D is.

With commercial code, I've never seen that kind of chain relationship, where B had full rights to become a competitor to A. If it's a value-add chain, where each step adds more stuff, A's not giving up any rights. They insist on being paid at each step of the way, and remain involved and active throughout the chain. If I buy a custom database written on Oracle, Oracle gets paid (a lot!) for the database instance I'm running, and that's above and beyond whatever I pay for the custom database code.

With GPL, A insists on being paid by making sure that all downstream users have the same rights to all code that was added... including, likely, him or herself. The payment terms are "freedom and code" instead of money. B through D voluntarily give up rights in either GPL or non-GPL scenarios. With non-GPL code, they cannot freely make copies and must, generally, pay per running instance. With GPL code, they don't have to pay, but have to give away any improvements they make along with their improved binaries. In general, B through D give up fewer rights under the GPL than they do with regular copyright, but they always give some up.

Commercial A, as recompense, gets money. GPL A probably gets access to downstream code improvements. That's not a GPL requirement, but in practice, that's usually what happens.

Ultimately, there's little difference. Both As are still getting paid. As B through D, you're buying either codebase, you're just paying differently.

As GPL B through D, you're getting a great deal; you can set yourself up as a full competitor to anyone else in the chain. You're not hostage to anyone's code, and in exchange, you can't hold anyone else hostage. It's a pretty sweet deal, overall. It's why the GPL prospers.

TANSTAAFL. If you want good code, you usually have to pay for it. Fortunately, the GPL is not particularly onerous.... unless you intend to enslave your customers instead of serving them.

GPL and private contracts

Posted Mar 24, 2007 16:10 UTC (Sat) by malor (guest, #2973) [Link] (5 responses)

> "Maybe I could say, "unlike conventional copyright licenses, GPL interferes in transactions in ways in which the copyright owner doesn't have a material interest."

No, you can't say that either. As the other poster points out, the copyright owner most emphatically DOES have a material interest in all reproductions of the copyrighted work.

You get the unique ability to reproduce *and modify* the work freely, in exchange for ensuring that everyone downstream from you has the exact same rights that you do.

Again: yes, the GPL changes the nature of the relationship you can have with a customer. If you don't like that, you can pay for code with money, or you can use BSD licensed programs instead. You don't have to accept the GPL if the unique privileges and powerful utilities it offers aren't worth the price. v3 makes the price a little steeper, but I think only irrational people can argue that v2 wouldn't have said exactly the same thing if DRM had existed twenty years ago.

GPL and private contracts

Posted Mar 24, 2007 18:16 UTC (Sat) by giraffedata (guest, #1954) [Link] (4 responses)

Maybe I could say, "unlike conventional copyright licenses, GPL interferes in transactions in ways in which the copyright owner doesn't have a material interest.
No, you can't say that either. As the other poster points out, the copyright owner most emphatically DOES have a material interest in all reproductions of the copyrighted work.

Irrelevant; the quote above doesn't say anything about material interest in the reproduction. It talks about material interest in ways of interfering with a transaction.

You don't have a material interest in whether some stranger gets the source code for some code written by someone else. Through the magic of GPL, you may control it anyway.

And don't try to define "material interest" so that all interests are material. There's a separate term for a reason.

GPL and private contracts

Posted Mar 25, 2007 0:38 UTC (Sun) by malor (guest, #2973) [Link] (3 responses)

> You don't have a material interest in whether some stranger gets the source code for some code written by someone else. Through the magic of GPL, you may control it anyway.

If you are commercial, you have an interest in making sure you get paid for all copies of your code. If you are GPL, you have an interest in making sure anyone that adds to your program honors the license.

Both types of code have a price. One is money, but allows the freedom to keep your code closed. The other is usually low- or no-cost, but denies you freedom to hide your code.

Pay one or the other, your choice.

(or go BSD code, if you prefer.)

GPL and private contracts

Posted Mar 25, 2007 0:53 UTC (Sun) by vonbrand (subscriber, #4458) [Link] (2 responses)

Both types of code have a price. One is money, but allows the freedom to keep your code closed. The other is usually low- or no-cost, but denies you freedom to hide your code.

This is wrong. You can certainly keep your modifications to GPL code secret. If you distribute the modified code, you have to share the source to the changes.

Note that in the case of closed source you don't have the right to distribute anything, not the original version nor your modified one.

GPL and private contracts

Posted Mar 25, 2007 0:57 UTC (Sun) by malor (guest, #2973) [Link] (1 responses)

Yes, this is true, but this particular thread is talking about a chain of people, from A to B to C to D. giraffe seems to be asserting that the GPL is somehow unethical or something in that it 'interferes with' other business relationships, and I'm trying to point that this is the PRICE of using the code, and that it's entirely optional. If you don't like the price, buy commercial code instead.

You give up rights in either case, it's just a matter of which rights fit your particular situation the best.

GPL and private contracts

Posted Mar 25, 2007 1:15 UTC (Sun) by giraffedata (guest, #1954) [Link]

giraffe seems to be asserting that the GPL is somehow unethical or something in that it 'interferes with'

It's "or something," and in particular I'm asserting that GPL is unusual among copyright licensing schemes. Nothing more. I jumped into this thread when someone made an incorrect statement about what one can do with GPL code, based on reasoning that would be sound for any conventional licensing scheme. I went beyond correcting the particular misstatement to explain you have to use a whole different mindset when looking at what's allowed with GPL code.

The misstatement, IIRC, was that a downstream distributor could make a private deal with his distribuee in which the latter doesn't avail himself of GPL freedoms, and the copyright owner wouldn't have anything to say about it.

The Torvalds Transcript (InformationWeek)

Posted Mar 22, 2007 23:06 UTC (Thu) by barbara (guest, #3014) [Link] (14 responses)

>I remain convinced that the bad blood between Stallman and Linus
over 'GNU/Linux' is at the core of Linus' objections. I don't think he
would be attacking the GPLv3 so fiercely if he weren't personally invested
somehow.

You're probably right. I also think the reason that Linus can't stand
Richard Stallman is that he reminds Linus of his parents who were involved
in left-wing politics (his Dad was a communist and spent one year studying
in Moscow -- horrors!!). Politics was a big part of Linus' family, he's
rebelling against that in my view, and he now sees *any* kind of politics
in Linux as a Bad Thing (tm).

With all the threats to our freedom to use our computers as we wish, to
use an unfettered Internet, and to simply have the freedom Linux brings
us, it's important to keep our eye on and fight these threats (be it the
DMCA, the RIAA, the MPAA, and a whole host of other control freaks).
Politics is indeed important despite what Linus says and thinks.

The Torvalds Transcript (InformationWeek)

Posted Mar 22, 2007 23:48 UTC (Thu) by malor (guest, #2973) [Link] (7 responses)

Well, I can't speak about that directly, as I don't know him personally; I saw him do a Q&A session once, but that's about as close as I've gotten. :) I am quite sure that there's something more to his GPLv3 hatred than he claims, but what it is, exactly, I don't know.

Stallman hatred seems plausible to me, but your idea of it being rebellion against his home politics could also be an explanation. Regardless, his apparent disdain for supporting the idea of Free Software has even bitten him, very badly, in the BitKeeper fiasco. Apparently, he didn't learn that it's important never to make yourself hostage to code.

In my view, the GPLv3 is the best way to make sure that doesn't happen, and I find it curious and disturbing that he'd fight so hard for the ability to be chained up again.

GPL3 = BitKeeper2?

The Torvalds Transcript (InformationWeek)

Posted Mar 23, 2007 7:03 UTC (Fri) by dlang (guest, #313) [Link] (6 responses)

even stallman used closed source tools when there weren't opensource tools that would do the job.

at the time the bitkeeper use started there wasn't anything available that would do the job nearly as well (for that matter, when the bitkeeper use stopped there wasn't a good option available, that's why he started the git project)

stallman could have done all his programming on a system with no closed source tools. it would have inconvienced him, but it could have been done (not all computers HAVE a BIOS on them), he choose not to do this and instead use existing closed source components until he considered the opensource componenets 'good enough'

Linus does the same thing, it's just that what each of the two is willing to define as 'good enough' is different.

beyond that there is a fundamental difference in approach in winning users

Stallman wants you to switch to opensource (free) software becouse it's 'The Right Thing To Do'

Linus wants you to switch to opensource software becouse it's the best choice available to you.

Personally, I think that Linus' approach is going to be far more sucessful. you always do better when you involve people's self interest then when you tell people that they should suffer for the cause.

David Lang

The Torvalds Transcript (InformationWeek)

Posted Mar 23, 2007 12:12 UTC (Fri) by malor (guest, #2973) [Link] (5 responses)

He was widely and strongly criticized for choosing BitKeeper, and as it turns out, the critics were correct for doing so. Code that can hold you hostage is bad.

Closed source, in and of itself, isn't automatically bad: what's important is whether it's used to assert control over how you use it and what you do with it. BitKeeper did that, and it bit Linus very badly.

The anti-DRM provisions of the GPL3 are there to prevent people from being BitKeepered via hardware restriction. It amazes me that Linus, of all people, would come down against that.

The Torvalds Transcript (InformationWeek)

Posted Mar 25, 2007 0:17 UTC (Sun) by vonbrand (subscriber, #4458) [Link] (4 responses)

He was widely and strongly criticized for choosing BitKeeper, and as it turns out, the critics were correct for doing so. Code that can hold you hostage is bad.

Yes, Linus was criticized for using BitKeeper. No, he was absolutely right in doing so: The then imminent "Linus meltdown" was avoided, the kernel development picked up speed, and in the end all have got better SCM tools.

Was it painless? No. Was it worth is? Yes!

The Torvalds Transcript (InformationWeek)

Posted Mar 25, 2007 0:41 UTC (Sun) by malor (guest, #2973) [Link] (3 responses)

It was unnecessary pain. If he had simply refused to be chained in the first place, the whole fiasco wouldn't have happened. Kernel development essentially came to a dead stop for ages while they redid their tools; had they simply done the tools in the first place, they'd never have had the problem.

Code that can hold you hostage is bad. It's always bad. Anytime someone wants to tell you how or when you can use code, you're at a profound disadvantage.

I can't believe you're painting BitKeeper as a net positive. You sound like a Republican spinning the Plame outing.

The Torvalds Transcript (InformationWeek)

Posted Mar 25, 2007 2:34 UTC (Sun) by bronson (subscriber, #4806) [Link] (1 responses)

First, it really wasn't as painful as you describe. Kernel development definitely did not come to a standstill. While Linus ran off and wrote Git for a bit, everyone else continued using whatever tool they liked. No big deal (unless you happen to think that only Linus can develop the kernel...?)

Because existing distributed SCMs were crap (they were), Linus had to write his own tool no matter what. It isn't like one path involved great pain and the other path involved zero pain. What scenario are you picturing where pain is somehow avoidable and there's no need to write Git?

Now, would you rather:

- Linus write his tool cold, having very little first-hand experience of distributed development.

- Linus write his tool after thoroughly using the most advanced distributed development tool available at that time.

Seems to me like either way works pretty well.

For the record, I agree that the Republican spin on Plame (and now Gonzales) are just stupefying. But they have pretty much zero in common with Linus choosing an SCM.

BK and git: Lessons learned (or not)

Posted Mar 25, 2007 3:05 UTC (Sun) by kevinbsmith (guest, #4778) [Link]

Linus (and the world) got lucky with git, and many folks have learned the wrong lesson from the whole affair. I don't mean that Linus isn't a great programmer or that git isn't a great tool. What I mean is that the BK folks *chose* to give the kernel folks several weeks to stop using BK, rather than terminating all kernel dev licenses immediately (which they could have done). And git turned out to be relatively quick and painless for Linus to write.

If either of those (or other factors) had come out differently, the kernel could have been hurt very badly by the whole BK experience. Was BK net positive? With hindsight, I would say yes. Could it have been a disasterous net strong negative? Absolutely. The risk was high (and was known at the time BK was selected), the bet was made, and Linus got lucky and won.

Wrong lesson: Depending on closed source for a key tool is fine, because git easily replaced BK.

Right lesson: Depending on closed source for a key tool is very risky, and you really don't know how things might turn out.

The Torvalds Transcript (InformationWeek)

Posted Mar 25, 2007 4:33 UTC (Sun) by dlang (guest, #313) [Link]

if he had refused to use bitkeeper at all then kernel development would have ground to a halt as Linus overloaded, and people would not have had the experiance with distributed SCM tools to be able to create git or it's equivalent.

if the opensource SCM systems weren't up to the job of being able to handle the kernel after several years of bitkeeper being an example, what makes you think that any of them would have been suitable for it back then?

as far as I can tell, other then git, there is still no opensource SCM that can handle the scale and distribted workflow of the kernel development.

yes bitkeeper was a gamble, but the options at the time were gambles as well (what decision isn't a gamble for that matter). Yes, if Larry McVoy had been an evil person out to scramble the kernel development process things could hve turned out worse then they did

remember that even stallman says that you are justified in useing a closed source tool if there are no opensource options.

The Torvalds Transcript (InformationWeek)

Posted Mar 23, 2007 21:07 UTC (Fri) by h2 (guest, #27965) [Link] (1 responses)

barbara and malor, thanks. It's been, especially since the Groklaw rantings by Linus, very easy to see that the foundation of his opposition was clearly not rational, even though he's now of course trying to paint some rational cover for this. But what was a mystery to me was the actual source of that irrational opposition. Now it finally makes sense, even though I'm usually not a big fan of psychological explanations, this one just fits a bit too well. I think there are some other core reasons too, equally irrational, but there's not a lot of point getting into details.

What continues to amaze me is that Linus and the majority of the kernel team day in and day out enjoy the freedoms a license like the GPL explicitly envisioned when it was written, but they seem somehow unable to admit that their actual day to existence is the direct result of the freedom the GPL made a space for. That's the core freedoms, not the fiction of Open Source.

They live good, free, creative lives, almost dreamlike in fact. But they don't seem able to connect the dots. Odd. The fact that they have come to mistake some corporation for the actual real person end user, that's kind of sad, but predictable too, that's the world they live in.

The GPL always restricted the freedoms it imparts enough to make it able to sustain itself against a world that is generally quite hostile to that concept. And BSD type licenses have always been there for users who don't want that restriction. And the world in the past 15 years has grown more aggressive, you can without any problem see patent attacks on the horizon, so failing to take something like that into account would be a serious error on the part of the FSF. And it's not an error they have any plans on making. The entire Tivo thing collapses as soon as you remember that the GPL is all about the rights of the end users, not some corporations. People that is. That also has not changed, nor I hope will it ever.

The Torvalds Transcript (InformationWeek)

Posted Mar 28, 2007 19:10 UTC (Wed) by wilck (guest, #29844) [Link]

Linus understands the GPLv2 perfectly. His statements in the Newsweek article demonstrate that clearly. He acknowledges that the GPLv2 fits his needs _although_ the intentions of the FSF when they published the GPLv2 may be different from his own.

Whatever Linus' true motivations may be, it's pathetic to dismiss his statements because of psychological speculations. You can disagree with him (I do, too, at least partially), but please respect the man and his arguments.

The Torvalds Transcript (InformationWeek)

Posted Mar 25, 2007 23:55 UTC (Sun) by jschrod (subscriber, #1646) [Link]

> I also think the reason that Linus can't stand
Richard Stallman is that he reminds Linus of his parents who were involved
in left-wing politics (his Dad was a communist and spent one year studying
in Moscow -- horrors!!)

You are either a troll, or have quite obviously never been to Finland, or to Europe for that matter. Postings like yours and those of marlor who accuses Linus Torvalds and the kernel developers to "behave out of their pocketbooks" make me want to ask Jon for a KILL file feature in the LWN comment section.

Joachim

The Torvalds Transcript (InformationWeek)

Posted Mar 28, 2007 18:57 UTC (Wed) by wilck (guest, #29844) [Link] (2 responses)

I also think the reason that Linus can't stand Richard Stallman is that he reminds Linus of his parents

That's a dangerous line of arguing. Why not go one step further and ask what what Stallman is rebelling against (man, that guy's parents must be weird!)? And what are your motivations? Mine? Why the heck are we fighting against industrial giants, and trying to defend rights that 99.9% of the population don't care about? Dr. Freud, please explain!

Even if you were right, that would have nothing to say about the validity of Linus' arguments.

The Torvalds Transcript (InformationWeek)

Posted Mar 28, 2007 20:30 UTC (Wed) by malor (guest, #2973) [Link] (1 responses)

Well, I can't talk about barbara's idea, as I don't know anything about his past in that regard.

I CAN, however, see that the arguments he presents are largely based in 'don't trust Richard Stallman', and that's why I made the observations I did. I don't he likes Stallman, and I can't blame him: Stallman's GNU/Linux name grab made ME furious, and I'm not even one of the people in the project.

Regardless, I think it's a bad reason to try to sabotage GPLv3. I don't KNOW that that's the reason, but it sure looks likely to me, given their past and the way he's framing his arguments.

I mean, he couldn't use the GPLv3 even if he wanted to, so why try to mess it up for everyone else?

The Torvalds Transcript (InformationWeek)

Posted Mar 29, 2007 9:13 UTC (Thu) by wilck (guest, #29844) [Link]

I can't see that Linus is trying to "sabotage" GPLv3. He says that he doesn't like it, and that he won't adopt it if key parts of its current draft for remain unchanged. Why shouldn't he?