Linus Defends Proprietary File Formats [Updated]

Simon (S2) writes "Torvalds launched a blast against OpenOffice.org, and defended Microsoft's right to keep its binary Office formats proprietary. 'I'm happy with somebody writing a free replacement for Microsoft Office. But I'm not fine with them writing a free replacement just by reverse engineering the proprietary formats,' said the Linux founder. 'Microsoft has its own reasons for keeping them proprietary, and I can't argue with that.' At the heart of Torvalds' decision to refrain from using Bitmover's BitKeeper source code management tool last week, a day after BitKeeper decided to drop its limited functionality free client, is a dispute between BitKeeper developer Larry McVoy and Samba developer Andrew 'Tridge' Tridgell. It has subsequently emerged that Tridgell was working on a clean room reverse engineered implementation of McVoy's proprietary software, and Torvalds has come down on the side of his friend McVoy." Update: 04/13 17:24 GMT by T : As reader Daniel Callahan points out, this is a goof. "The Register article made up the Torvalds quote. The article offers the quote and then continues: 'Actually he didn't - we just made that quote up. But what Torvalds really did say this weekend is only slightly less bizarre.'"

Comments

Dupe and a lie

[Sanity]

This is essentially a story about the last time Slashdot posted a story about this.

The only addition is the false quote from Linus, I think it is pretty unforgivable that CowboyNeal would put a deliberately false quote in the blurb of a story, but its not surprising given that slashdot editors really don't appear to give a flying fuck any more (even after I sent an email to the "on duty editor" after seeing this in the "mysterious future").

Re:Dupe and a lie

[maotx]

I am a subscriber. You have an option of the * showing up or not. Look by my name now as it should be there.

Re:Dupe and a lie

[Edgewize]

In this reframing: Linus has clearly come down against reverse-engineering. TFA is further correct in pointing out that this is inconsistent with what Linux, OpenOffice, gcc, and a bunch of other open source projects are all about.

No. This has nothing to do with what Open Source is all about. GCC wasn't created by examining the bytecode output of an Intel compiler. Most open-source "clone" projects are based on available documentation and end-user appearance, not examination of binary data.

Linus is in favor of implementing available standards in a free and open manner. He is not in favor of reverse-engineering someone elses implementation against their wishes.

There is no inconsistency to be found.

Re:Dupe and a lie

[1u3hr]

Journalists have a rule that anything between quotation marks has to be an exact quote. You're not even allowed to correct the grammar or make irrelevant changes to help it fit into your sentence better.

Yes you are. I was editing a book some years ago and the author was apparently taking delight at quoting grammatical mistakes his non-English speaking subjects made, which I thought a cheap shot. Looking up some reputable texts on journalism supported my view that minor errors can be silently corrected in quotes unless it's from a published text, and this is common practice. Actually listen to what someone says in an interview and compare with a written article -- you won't see the "ums" and false starts that almost everyone makes, unless they're trying to make the subject look like an idiot. Of course, trying to make any sense of what GWB says off the cuff may require more than that.

Both Slashdot and the poster also screwed up, but The Reg is the one who really blew it, IMHO.

I don't know if you're a regular reader of the Reg, but pisstakes are a feature of their writing. Their logo is a vulture; their slogan is "Biting the hand that feeds IT". They don't post lies but they sometimes do sex things up a bit. The poster is obviously a troll, he knew what he was doing. However, there is no excuse at all for Cowboy Neal. The "we just made that quote up" is prominently in the third paragraph. CN is just lazy and sloppy, like they all seem to be now. They collect a salary for editing this, they should be ashamed. But they're not -- I've sent several messages to him via the editor's address on similar issues, and they all bounce, he doesn't even want to know when he fucks up.

Re:Dupe and a lie

[Anonymous Coward]

He believes that you should get permission to mess with things that other people started. I don't se anything wrong with that position.

the problem is in the context and is called vendor lock-in. If the vendor refuses to provide you with an export tool (after all, why should he, as you're a leaving customer) you pretty much have no option but reverse engineering. Well, that or manually moving everything over, when applicable. Why should a vendor be able to effectively restrict what you can do with your data? He can restrict your use of his tool, but the data his tool stored is still yours.

On the other hand, whether reverse engineering will be used simply to extract the data or to make a compatible tool is a closer to the issue here.

Misleading headline... RTFA editors!

[garcia]

Torvalds launched a blast against OpenOffice.org, and defended Microsoft's right to keep its binary Office formats proprietary. "I'm happy with somebody writing a free replacement for Microsoft Office. But I'm not fine with them writing a free replacement just by reverse engineering the proprietary formats," said the Linux founder. "Microsoft has its own reasons for keeping them proprietary, and I can't argue with that."

Actually he didn't - we just made that quote up.

Well, thanks for another misleading headline Slashdot! While I applaud your recent efforts to fix crappy editorial comments and duplicate removal you still are showing that you refuse to even read the articles that users submit. Now on to the rest of the article...

You know Linux is a clone of Unix because Linus couldn't run Unix on his 386 machine. He wasn't pleased that he couldn't do something and he worked around it. Why can't someone be displeased with other proprietary systems and create workarounds for them?

I'm preaching to the choir here but reverse engineering is a Good Thing for all communities. There is absolutely no reason that we should not support working around what others have obfusticated to make money for themselves.

Linux wouldn't have nearly the same capacity in the Windows world we live in if it wasn't for Samba. Yeah, there is NFS for Windows and various other file sharing protocols that could have been used but Samba makes it easy for anyone to fit their Unix clone right into their pre-existing Windows network without much trouble.

The free client was costing Bitmover $500,000 a year, explains McVoy. "At that point we started looking at what it would be like to discontinue the free BK.

So? It's obvious that the pay-for client offered nothing worth what you were asking if the free client can do the job. Either price properly or make the pay-for product much better. I'm not talking about crippleware or nagware. I'm talking about creating a much more superior product that entices people to buy rather than hobble along with what the free version offers.

Plenty of companies out there have been doing it just fine by basing their business model on Linux. Why can't McVoy find the same happy existence?

"What Larry is not fine with, is somebody writing a free replacement by just reverse-engineering what he did. Larry has a very clear moral standpoint: 'You can compete with me, but you can't do so by riding on my coat-tails. Solve the problems on your own, and compete honestly. Don't compete by looking at my solution.'

They are competing honestly. They are doing it in a clean lab. They aren't trying to steal your code and use it themselves but they are trying to take a great idea and make it better. Welcome to the real world. Crying doesn't do anything but piss people off. Do something to your own software that will make it stay one+ steps ahead of the reverse engineered competition.

Erm

[pmc]

From the article:

Actually he didn't - we just made that quote up.

Sheesh.

Linus did NOT say that, RTFA!

[EnronHaliburton2004]

But I'm not fine with them writing a free replacement just by reverse engineering the proprietary formats

Linus never said that. From the fucking article:

"Actually he didn't - we just made that quote up."

Please don't put words in Linus' mouth. That's very sleazy, Mr. Andrew Orlowsk.

Also from the fucking article:

So is Linus going to come down hard on other efforts to create a free and open alternative to a proprietary product - say, for example, a UNIX(TM)-like operating system?

Does the author understand that this is a different situation? Linus did not reverse engineer Unix.

Re:Linus did NOT say that, RTFA!

[Anonymous Coward]

Really? I would have thought that that's exactly what Linux is.

Please explain how Linus writing Linux was not reverse-engineering of Unix.

Re:Linus did NOT say that, RTFA!

[man_of_mr_e]

Linus (and others) wrote Linux to conform to the POSIX specifications. They didn't reverse engineer any form of Unix, not even Minix (though Linus did start with it, he quickly threw it away).

Re:Linus did NOT say that, RTFA!

[EnronHaliburton2004]

Linus didn't have the POSIX specs when he started so he couldn't possibly have written to them.

But he did eventually get the POSIX specs, and made Linux conform with the specs.

As an interesting piece of history, here's the post which got the ball rolling.

Re:RTFA - "we just made that quote up"

[mcg1969]

Nor did CowboyNeal, apparently.

Before you get upset...

[ceswiedler]

Before anyone who didn't RTFA gets up in arms: No, he didn't say that, and the article header really should explain. The Register is drawing a comparison with his attitude towards BitKeeper. s/BitKeeper/Microsoft and s/Tridge/OpenOffice.org.

Were the submitter and editor confused, or are one or both intentionally trying to provoke a reaction by providing an inaccurate summary? At least the Register article has a clear "No, he really didn't say that" line. The /. summary acts as if it's a real Linus quote.

Linux is about reverse engineering?

[LordNimon]

From the article:

If McVoy thinks that reverse-engineering is so 'dishonest', then why did he offer to give free tools to a worldwide project whose primary focus is to reverse-engineering an entire OS?

I'm assuming the "project" in question is the Linux kernel. Well, I'm sorry, but Linux isn't about reverse-engineering an entire OS. Which OS do they mean, anyway? Unix or Windows? In either case, they're wrong. The Linux kernel is not developed by reverse engineering some other operating system. With the exception of a couple device drivers that were designed by reverse engineering their Windows counterparts, it's completely original development. Sure, it has Unix-like behavior, but that isn't gleaned by reverse engineering.

Sounds to me like the article author has a overly broad definition of "reverse engineering".

Re:This points out Linus' inconsistency very well

[Undertaker43017]

Number 3 is reverse engineering, as long as some sort of code disassembler is not used to actually look at the code used to implement said functionality. Otherwise it is no longer "clean" and NOT reverse engineered, but copied.

Re:This points out Linus' inconsistency very well

[PhilHibbs]

I disagree. Reverse engineering can be done with the use of a disassembler or debugger, so long as the disassembled code is not just copied. Clean-room RE often makes use of one team that uses disassemblers and debuggers, they then write the functional specfications for the programmers who write the code without access to the disassembled original. They can run the original to see what it does on-screen and to create and test files. How blurred this line is depends on the ethical framework, during the war this line did not exist at all.

Clean room reverse engeneering

[maxwell demon]

Done correctly, you have a final product with the same external behavior as the target, but with no possibility of IP contamination.

With no possibility of copyright violation. However, it could still be patent-contaminated.

Re:This points out Linus' inconsistency very well

[Jeremy Allison - Sam]

Thanks Bruce. This is *exactly* what happened. I'm hoping that when tridge's code finally becomes available people will be able to see the truth of the matter.

Jeremy.

Nope

[JoeBuck]

OpenOffice cloned Microsoft formats by reverse engineering. While the original protocol Samba implements had documentation, Samba had to reverse engineer all Microsoft's extensions to that protocol. Novell/Ximian reverse engineered Microsoft's Exchange to allow Evolution to hook in. And many device drivers in Linux required reverse engineering to develop.

Proprietary software developers also engage in reverse engineering. It's completely legal if done in a way that complies with the license (Tridge, the guy behind Samba and the free BitKeeper data extracter, uses captured traces of network traffic as his preferred method). If you think it's unethical, then you are basically saying that you believe in monopolies, and you might as well just buy all your software from Microsoft.

Linus replies

[Anonymous Coward]

Sorry Dean...
http://www.realworldtech.com/forums/index.cfm?acti on=detail&PostNum=3322&Thread=2&entryID=49312&room ID=11

Bloody nonsense!

[Danuvius]

Ethics are hard to nail down, but Linus clearly believes that if someone gives you a gift, you don't bend them over and ram them up the rear for their generosity, even if it's perfectly legal for you to do so.

What the hell are you talking about??

Tridge was not given a gift.
Tridge tried to reverse-engineer the network protocols used by bitkeeper, without using a copy of bitkeeper.

Ethics are hard to nail down? In this case WTF??

Re:Bloody nonsense!

[Danuvius]

Think of it this way. Your employer or company you contract with has an agreement that their employees or contractors won't do X with their software. Contractor/employee knows this. Contractor does X. Is this ethical? Debate :)

This did not happen.

OSDL "agreed" to not use BitKeeper for reverse engineering. Tridge did not use BitKeeper **AT ALL**.

And the reverse-engineering that he did was not of the type (take apart the program to figure out the source code) that Larry McVoy keeps trying to misrepresent Tridge's efforts as.