Historical Unix, Open Source Legal Battles, and John Lions

Invicta{HOG} was the ffirst to write us about today's new Salon piece. It covers the first legal battle open source faced, quite some time ago, John Lions and a look into the history of Unix. It's a pleasant read.

Re:Open Source is too young for this

[Gurlia]

IMHO this is why FSF needs to start coming out of their proverbial well and start a legal defense fund for Free Software, in particular, the GPL. So far it seems to me (correct me if I'm wrong) that FSF only cares about defending the GPL when the software involved is owned by the FSF, like GCC, Emacs, etc.. We need a legal defense fund for free software in general, not just for FSF products! Otherwise all it takes is one piece of obscure GPL'd software (that FSF doesn't care about) being abused by some company/person, and the court overturning the GPL -- and everything will collapse.

Disclaimer: IANAL.

Re:Open Source is too young for this

[Demona]

Could you go into more depth on how "the current widespread abuse of the GPL and Open Source is already part of Common Law"? (Based on my observations, common law has long since given way to equity, admiralty, bureaucracy and just plain stupid tyranny, at least in most parts of America.)

It sounds like your dislike for the term "open source" might be coloring your views, but I'd like to understand your position better. Please expound!

PS: A friend of mine noted:

Reading the paraphrase of the title, 'Commentary and Source,' and reading the description of it as _samizdat_, and the importance of possessing a cherished, water-damaged copy, one really gets a definite feeling of Unix hacking as secular yet religious vocation.

Re:Open Source is too young for this

[Rupert]

Yes and no. You make a vaild point:

Neither the GPL or Open Source have been proven in court - ANYWHERE. The current widespread abuse of the GPL and Open Source is already part of Common Law which (as we all know) is hard to overturn in most courts worldwide.

but you diminish it with the rest of your post.

Firstly, if, as you say,

Open Source is a marketing term for Free Software

then Open Source is not young at all.

Secondly, it may appear in the title of the article as "Open Source", but in the text it is "open source", and without the capitals this is what this story is about - the openness of source code, not the OSS movement.

The GPL will be fine.

[Paul Crowley]

Remember that if the courts decide that the GPL is so much toilet paper, the consequence is that distributors have no right to distribute code at all, since it's only the GPL would grant them that right. It's pretty much a fail-safe license: there's simply no legal route by which someone could distribute GPL'd code as if it placed no obligations on them unless the court returns a seriously perverse verdict.

You're mistaken about Common Law. It's possible you're thinking of trademark law; this isn't trademark law.
--

Re:Open Source is too young for this

[jflynn]

I must disagree that "open source" is nothing more than a marketing term for "free software." It may be that to you, but it's more than that to many people.

Neither open source nor free software have much to do with either money or law, IMO. Free software maintains it is a moral imperative to open sources because information shouldn't be property, while open source just says software works better and develops faster if you do. You can honestly believe in the efficacy of sharing sources without believing it has anything to do with morality. To provide an umbrella that includes such people is more than mere marketing.

Yes, the GPL has never been tested in court. So what? That argues against there being much reason to write cute articles to prepare legal positions that probably will never be necessary.

You really can't throw out things like "widespread abuse" and call for the overturn of the GPL and open source before it's too late without backing up your argument. I am not aware of any abuses.

If you don't like the GPL and Open Source, why do you think you have a right to the labors of those who do, and wish to voluntarily pool their own efforts together? Do you believe in nationalization of all intellectual property or just that of groups you disagree with? In actual fact, if the GPL fails, it should fail safe, that is, all GPL software will become undistributable without permission from all copyright holders. Is that what you want?

It was an interesting historical article that barely mentioned either free software or open source.

Available on amazon.com

[neophase]

The 1996 edition is on amazon.com if you search under Author: John Lions.

ISBN: 1573980137
Price: $29.95
Ships: usually 2-3 days
ISBN: 1573980137

I think I'll make a suggestion to the Christmas elf :-)

==================================
neophase

Re:The GPL will be fine.

[Chalst]

It does seem conceivable to me that the courts could rule that, while
the right to make software public domain is OK, the extra requirement
that further modifications have to be made available amounts to theft.
Not likely, but conceivable.

So I think there is some ineterst in seeing how the courts react.
I am conscious that I am talking about `the courts' irrespective of
national distinctions in IP law, but it seems a field in which the US
is most likely to see a first challenge and other OECD countries are
likely to follow suit.

Ran into someone interesting while buying my copy

[jms]

I attended a Usenix LISA conference about a year ago. The Lions book had just been released, and there was a table in the vendor area with copies for sale. I picked up a copy and started looking through it ... trying to decide whether to buy it then or wait until later, when I happened to notice that the person standing next to me was having a pretty interesting conversation with the booksellers. Checked his badge ... Dennis Richie! I'd no idea beforehand that he was on the program. Of course, I couldn't think of a thing to say. I settled on asking him nicely to autograph my copy (he did, nicely.)

Feel free to moderate this tripe down :-)

John Lions, remember?

[GnuGrendel]

I'm not sure why the reponses to this article have devolved into a flamewar over the GPL, since this article was about John Lions, not the GPL.

I had never heard of John Lions before this article, but he sounds like he was an intelligent and wise man, and I wish I had had the chance to have him as a teacher. It's good to learn the history of the movement to open the source to the people, and fascinating to hear source code discussed and dissected like literature. These are the types of intelligent conversation that I wish I saw more of here, instead of petty bickering.

Thanks, John, and I hope to read your book soon.

What I note is more interesting

[Anonymous Coward]

is the total LACK of comments about this article. Its like the Linux camp could give a damn about Unix history.

*walks away stroking greying beard, mumbling about kids, and how people who don't understand Unix are doomed to repeat it*

SCO will sell Unix release 6

[Anonymous Coward]

SCO will sell "personal use" licenses of the Unix release 6 code if you actually want to run the same code that the esteemed Mr. Lions talks about.

You then either have to buy a PDP-11 or find an emulator.....

Western Electric?

[theonetruekeebler]

There was a reference to Western Electric trying to supress the Lions Book. I never heard of them having anything to do with the OS. What was their stake in Version 6 or 7?

--

Running Version 7

[theonetruekeebler]

Is there any way of setting up an environment to run Version 7, possibly inside a VM of some sort under Linux?

I think that one of the most significant things about Lions's work is that it's a commentary on a complete kernel. Works like Bach's seminal Design of the Unix Operating System notwithstanding, there are no follow-the-code examples of a real OS out there today; modern Unices like Linux and BSD are far too complex to just sit down with and start understanding the deep mysteries of OS construction. Of those two, I'd prefer BSD for its more cohesive design, but even then, you're looking at twenty or a hundred times the code of the original Version [67] kernels. A 1990s Unix contains deep kernel hacks that make sense only in implementing advanced networking, scheduling, and virtual memory contexts; the study of these should be postponed until after the fundamentals are mastered. In contrast, there are wristwatches that could run V6; on a Palm Pilot or an embedded 386SX/20 system, V6 would scream compared to the now incredibly bloated 2.2.x Linuces.

I think the Lions book and a running system on top of it would be a tremendously Good Thing for a burgeoning Linus Torvalds or Alan Cox.

--

Interesting the low amounts of comments

[Col_Panic]

I find it interesting that one of the most interesting article that has appeared on /. in a long time has generated almost no replies.
I think this is interesting. The article was written in such a way that it offended no one (not that SOME people didn't find a reason to find something offensive anyway). It seems that slashdotters in general aren't interesting in entering into dialogue that doesn't involve confrontation.
Well, to prove my point, I shall point out how confrontational my post is!
The main thing I want to say, however, is good job on the article, I truly enjoyed it.

mark
(Was kestrel, but damned if I can get my password)

Great piece

[EvilIdler]

Here's me doing my Keeanu impression: Whoah.


It seems like I'll have to read Salon on a
regular basis now; good writing and interesting
material. And it's about something we all should
care about. Aaahh...

(And guys 'n gals - stop flaming eachother every
time there is mention of the three-letter acronym
described as open sores!)

Western Electric...

[Anonymous Coward]

Lessee, yes, I still do have my nth-generation photocopy. The front page of the Commentary (and the source listing) says,

"The UNIX Software System was written by K. Thompson and D. Ritchie of Bell Telephone Laboratories, Murray Hill, NJ. It has been made available to the University of New South Wales under a licence from the Western Electric Company.

This document may contain information covered by one or more licences, copyrights and non-disclosure agreements. Circulation of this document is restricted to holders of a licence for the UNIX Software System from Western Electric. All other circulation or reproduction is prohibited.

J. Lions
Department of Computer Science
The University of New South Wales
C Copyright 1977 J. Lions"

Postscript version of the book...

[epaulson]

I have a postscript version of the book, which I
think was made from the original nroff file. Is it legal for me to redistribute this?

-Erik

Re:Western Electric?

[edhall]

At the time, Western Electric was the manufacturing arm of the Bell System (i.e. AT&T). So when Bell Labs wanted to distribute Unix, they did it through Western Electric. That's just how things worked in those pre-divestiture days.

-Ed

Re:What I note is more interesting

[Gurlia]

I'm not speaking for other slashdotters, but for myself, I find this article "refreshing", in a sense, to know that people have fought for Open Source before. But I'm just not sure what to comment, other than, "Wow, so it's only today that we see the results of those few people's fight to open up source code to the public."

Re:Running Version 7

[IntlHarvester]

Take a look at the recent slashdot thread about a Java PDP emulator running Spacewar [slashdot.org]

Apparently the emulator can also boot V7 UNIX. Also see comment #73 for a link to DEC's non-java emulator.
--

Re:Open Source is too young for this

[vectro]

Actually, I believe the FSF does have a legal defense fund and is looking for somewhere to test the GPL in court, but the opportunity hasn't presented itself.

The problem is that the FSF can't legally sue unless they own the product they are sueing over. But my understanding is that if there were a willing author out there with a meritous case, then they would be willing to underwrite the lawsuit.

Re:The GPL will be fine.

[Chalst]

What the courts could regard as theft is that the GPL compromises ones
right to the product of one's own labour, ie. changes that one has
made to the source.

Yes it is absurd, for the reasons you argue, but while that is
clear to people who are familiar with the ideals of the FSF, it may
not be clear to someone whose legal education is grounded in
principles with very little similarity to Richard Stallman's beliefs
of how IP should work.

Sensationalism

[Denor]

I think this is quite interesting, too. Some statistics:

• Historical Unix, Open Source Legal Battles, and John Lions: 28 comments
• Interview with The Mind Behind Aibo: 50 comments
• Napster Attacks Open Source Clone: 112 comments

So we find that a story that was rushed and (repeadedly) flamed for being incorrect ends up having more commentary than two stories which seem fairly interesting, but non-controversial. I really would hope that Slashdot isn't starting to pander to the lowest common denominator like television does. I try to have more faith in the people here.

Re:Open Source is too young for this

[spacey]

Ever wonder why there's an Objective-C compiler as part of gcc? It's because RMS was willing to spank Steve Jobs and the next guys for not releasing the code for it. Next wasn't willing to try and defend themselves in court, and so in the end gcc-objc became a standard part of gcc.

The FSF does chase down suspected non-compliance with the terms of the GPL, and is willing to take it to court. It's just that most companies aren't willing to try to defend a violation of the GPL in court, because it's not defensible.

-Peter

Lion's Commentary vs. /.

[swordgeek]

Now that was a neat article. Poignant, even.

While I've got a copy of Lion's commentary on order, I still wish I'd been able to get a copy from the 'bootleg' days. Getting one of his original two-volume sets would be incredibly cool, but maybe a bit less personal since I didn't know who Lion was until after Unix 7 came out.

How many people here truly appreciate how important this work is, I wonder. I mean, an ENTIRE OS to read through and discover, and with commentary to boot. It's pretty cool, folks.

Of course, this seems to go right over the heads of way too many /. readers. Almost no posts on the subject, and half of them are stupid drivel about "turning teenage girls to stone." Has the entire user base been overrun with twelve year olds fantasising about older girls? Pathetic.

Just rambling folks. Nothing of importance to read here.

Re:Running Version 7

[theonetruekeebler]

Let me get this straight: I'm going to run the Linux I bought to replace Windows that I bought to replace DOS that I bought to replace my Atari 400 so I can run a Java RTE so I can run a PDP emulator so I can run Unix V7 so I can rebuild V6 from the Lions-commented source in a book legally reprinted from an illegally photocopied Australian book based on source code from that came from New Jersey by way of Wales. To quote Calvin and Hobbes, "The theological implications are staggering."

Should be fun.

--

John Lions

[Maclir]

I was fortunate - no, privileged - to have been taught by John Lions, and to have learnt about operating system theory using these notes in 1977. He probably has as much to do with the success of Unix, and establishing the intellectual freedom that inspired those who followed (Torvalds, Raymond, Stallman) as the other Unix pioneers. We owe him a great deal.

My only regret - I sold my copies of his notes in 1978.

Re:Postscript version of the book...

[epaulson]

Everything that I've gotten seems to say there's no problem, so here it is:
http://www.upl.cs.wisc.edu/~epaulson/lionc.ps

Re:Available on amazon.com

[Atomizer]

Thanks much! I dove into the comments just to look for a reference to the book. The only books I typically see for OS design that have source try to do some crappy DOS compatible thing. I would like to understand the Linux (or BSD) source, but ddin't really know where to start.

Ask Slashdot about Unix as Literature in the past?

[farrellj]

Good code is as important to keep as good literature. Sure, it looses some of it context, and can be somewhat obscure to read, but with the right commentary, it is priceless.

ttyl
Farrell

Re:Sensationalism

[sspiff]

I'd like to have more faith in people as well, but let's face it, controversy drives hits and that, unfortunately, is what it is all about today. Why else would Jon Katz write for Slashdot, for example? He's good at polarizing people and whipping them into a foaming frenzy. That equals more hits, pageviews, etc.

The only reason I read Slashdot is for the discussion on the less controversial subjects. Linux vs. Windows, PC vs. Apple, etc. will inevitably turn into a debate over political ideology, so I just ignore those articles. A lot of the sci/tech weblogs seem to do a better job on news than Slashdot these days, but then again, there is usually no forum for discussion.

Is it just me or do Slashdot stories seem to be getting more tabloidish? Too bad such junk sells!

I'ts nice to see us repeating history

[davecb]

... and improving on it.

Reprinting Unix source and providing a commentary on it was very much in the spirit of the times, and reflected to the credit of Dr. Lions.

It's also interesting that that was the public discussion of what was agreed to be Bell's intellectual property: v7 contained several improvements that were publically suggested by the Lions book. A win for both parties.

Finally, it's an existance proof that one need not take up the religious position that one's source code must be kept secret. In more modern times, it's an existance proof that it not be free in the Gnu sense to be worth publishing.

--dave (an author of a "free source" book) c-b

Re:John Lions, remember?

[Anonymous Coward]

I actually took a course on the UNIX kernel based on Lions' books during my senior year at Cornell (1977-78), apparently before the Bell lawyers clamped down on this stuff. It was my first introduction to UNIX, and I was immediately hooked.

Five years later, when I was working at Bell Labs at Murray Hill, I had the honor of working with John Lions, who had joined my department for his sabbatical. He was probably the first Australian I got to know personally, and I took a liking to his witty Aussie irreverence even if it cut a little sharply at times.

I had heard a while ago that John was sick, but I didn't know until I read the Salon article that he had died a year ago. How sad. He really did contribute so much to the open, academic study and development of real, live computer operating systems. Yes, I do think he deserves much of the credit for inspiring those who would later go on and start the Open Source movement.

And yes, I still have my copies of Lion's original books...guess they're collectors' items now!

Phil Karn
[e-mail]

John Lions' assesment style.

[Anonymous Coward]

When I attended the University of New South Wales I had John Lions for a couple of subjects including that where we learn't from the UNIX source code using the version 6 books talked about. Of the subjects I took, that subject was by far probably the most interesting. I am not sure if I still have the books and will be most disgusted with myself now if I did indeed throw them out in my last big cleanup earlier this year. One of the aspects of John Lions class which I found to be most unique was his manner of doing assesments. We were encouraged to write our own notes against the source code and the commentary and then we had a number of tests during the session where we could take the books including our extra notes we had scribbled onto the books into the exams. At the end of the test you could actually choose not to submit your answers if you felt you hadn't done well. Instead you could choose to sit a second different test the following week covering roughly the same material. You could also sit the second test even if you submitted the first in which case the results from both tests were averaged. This way if you didn't end up going well in the first you could sit the second and thus had a chance to improve your mark. For myself I was always happy with my first try and didn't sit the second test, but the whole concept I feel did encourage people to study the source code even harder when they realised they hadn't gone well in the first test. I was dismayed a year or so later when I heard that he had abandonded this assesment style and instead relied mostly on end of session exams and programming assignments. Anyway, when I get home now I will have search out my copies of the source code and the commentary. I dearly hope I still have them as I didn't realise what had gone on in regards to them. Two volumes that will be treasured if I do but still have them.

Literary criticism and programming?

[apsmith]

What I found most interesting about the article was the author's notes about how what Lions did resembles literary criticism. Of course Perl has long been advocating "literate programming" and Larry Wall has a rather interesting piece on Perl as the first postmodern computer language [wall.org]. But it really all started with C and UNIX.

I think we're reaching an interesting point here. Language of course was invented to communicate, and computer languages are called languages for a good reason - they are how we communicate in a deep way both with our computers and with other programmers who help maintain and develop our code. Before C (which came along with UNIX and made things like Lions' book possible) there were machine or assembly languages, which were too close to the machine to be very useable by humans. Or there were abominations like Fortran and Cobol, which generally insulted the intelligence of both the machines and the humans. C and later derivatives like Perl and Java somehow elegantly capture the essence of both machine and human ways of "thinking", and allow deep communication of meaning in relatively concise fashion. Just like a real language.

And this goes to the crux of the definition of open source itself. Binary executables are pure machine language, essentially unusable by humans, but since they contain the full "content" of a program (at least for a particular piece of hardware/configuration etc) why can't we just write good decompilers to convert machine code to source code? Maybe if our artificial intelligence efforts succeed eventually that will be possible, but until then the results of such machine translations are many times worse than the snarls babelfish and its ilk get into translating human languages... Things like variable names, the choice of loop or switch constructs, object-oriented constructs, even regular expression syntax are generally carefully chosen by the programmer for human readability and verifiability of the correctness of the instructions that the machine will carry out. What we're doing is really a new, and very interesting, form of literature... food for thought I hope!

Re:The GPL will be fine.

[J. J. Ramsey]

"What the courts could regard as theft is that the GPL compromises ones right to the product of one's own labour, ie. changes that one has made to the source." It isn't just the labor of the one who changed the source that is at issue, since one also has to consider the labor of the one who made the source available in the first place. The court could very well take the stance that the one who writes the original code is the one who has the right to give permission in how it is used (assuming that the rights didn't already belong to the original author's employer, or something like that.)

Re:The GPL will be fine.

[Chalst]

I would say anyone with the remotest sympathy would agree that that is how the courts *should* view the matter. The only point I am trying to make is that they *needn't*. These are novel ideas about property that are being talked about here and it is *conceivable*, not likely, that a court might think we are better off without them.

Re:John Lions, remember?

[craw]

I agree with you. It would appear that many ppl that peruse /. don't know how to intelligently comment on an excellent piece of journalism. The article was extremely bittersweet to read as I was unaware that Lions had passed away. However, the last section about how Lions was so pleased by the publishing of the book was very nice to see. That in itself made the article worthwhile.

As for bickering, how can one attack a story about a great OS, a true educator, and a significant series of events, all told superbly. Furthermore, the author actually married a Linux kernel hacker!

BTW, the book is amazing. How many other times do you get to check out code written by the gods of UNIX?

Re:Literary criticism and programming?

[FigWig]

I think that when computers are at the point to do `literate decompiling' they will be at the point that we will no longer need to write software ourselves - we will just tell the computers to do it.

I remember him as well

[Mr. Protocol]

I got copies of the Lions books when I was working at The Rand Corporation, the first commercial licensee of UNIX. We did a lot of userland stuff, some of which is still around (like MH). However, we were equally active in kernel-whacking, and invented things equivalent to named pipes, extremely buffed-up disk drivers, and a truckload of networking code.

The Lions books were a godsend. I'd learned UNIX by reading the source to Version 5 back at the U. of Illinois at Chicago Circle. Version 5 had no comments, except for one or two in the assembly-language assist, locore.s. Version 6 had comments! Wonderful! But the Lions books were a huge help in keeping us on track, and filled in an awful lot of background.

My originals remain at Rand. There's always the samizdat copies, but it's nice that they're in print again.

I never got to meet John Lions, but he was a huge help, at least to our corner of the early UNIX world.

Mike O'Brien

Re:Available on amazon.com

[judd]

Do.

I am very proud of my copy, and it makes great reading in small lumps.

The code is really quite beautiful, and Lions' commentary complements it well. he invites you to consider alternative ways of writing a particular function, and mostly, the original IS best.

The best thing is: it explains the famous comment "You are not expected to understand this".

Re:John Lions

[Harry Houdini]

I studied Elec. Eng. at UNSW starting
in 1978. When I graduated I worked there
as a Professional Officer for a couple of years.

I was a student in John's class - probably about
1979 or so - and I remember full-well most of us
being thrown into the deep-end with the UNIX
source code book and commentary. Few of us students
knew C - we were taught in Pascal - and John said
"Don't worry about that" and dived right into the
UNIX kernel. We learnt fast.

It certainly was the biggest program any of us
had ever read. I remember a colleague
after several weeks of reading through the source
code in lectures and tutorials turning to me and
saying "Where does it (UNIX) ever get time to do
anything !?" It took a lot of intense study before you
got to the first context switch.

The books were certainly in hot demand on the black
market and I think John was up on all the tricks.
A colleague tried to get a copy for a friend
who was not doing the course. John said, "You mean
to say that they broke into your car and all they stole
was your UNIX source code book ? Sorry, you will
have to share with someone else this session".

I remember the following year lending my books to
a colleague doing the course and regretting not
getting them back. Eventually I got a photocopied
version as a replacement.

I remember once staying up all night doing a tutorial.
Next day, I go into John's lecture and he says, "If you
haven't handed in your assignment it is too late now".
After all that work I was pretty pissed off so I decided
straight after the lecture I would go up to his office
and slip it under his door. When I went up, there were
already a whole bunch of other people's tutorials
already marked sitting in an old line printer paper
box. I was about to slip mine under the door when I
decided I better knock just in case someone was there.
I knocked and to my surprise John was in there
(how did he beat me up the stairs out of the lecture ?)
and he called "Come in". I was dumbfounded and decided
to lie. "Ehhh ... I just got my tutorial out of
the box out front and it hadn't been marked ...".
John was most apologetic "Oh, I am sorry, I must have
missed it. I'll do it right now. I am so sorry".
He marked it on the spot and I got a good grade.
For a young punk student versus the establishment
my little lie seemed like a real coup and brought
about much laughter in the student common room
when I told them.

His course and those run by Graham Hellestrand also
at UNSW were probably the most loved/hated depending on your
perspective. The courses those guys ran were very
practical and I really thrived on and appreciated them.

I enjoyed Rachel's story and it had an interesting
perspective.

Re:Literary criticism and programming?

[Anonymous Coward]

I thought literate programming was something that Donald Knuth came up with, and exemplified in WEB.

De-compilers: not likely

[UnknownSoldier]

> but since they contain the full "content" of a program (at least for a particular piece of hardware/configuration etc) why can't we just write good decompilers to convert machine code to source code?

Because the binary program IS NOT the full content.

Descriptive variable names just become numbers. Arrays become pointer walks. Not to mention code optimization.

THAT is the problem. The compiler has striped away the usefull [symbolic] information.

How is the de-compiler going to distinguishing between a pointer to a pointer and an array of poitners? etc.

IAAP (I am a programmer)

Cheers

Re:What I note is more interesting

[pb]

What do you find surprising? I found this story on slashdot, followed some links posted in the comments, and I'm reading it (the book, not the article) now, after everybody's commented on it.

That, long with some of the original papers on Unix, that is... Some of these examples are hilarious, since DOS and Windows both do things the wrong way, when compared to the Unix way.

(instead of real pipes, DOS uses temporary files, while Windows builds useless functionality into every program...)

At least some people took this message to heart, if not Bill G...

"But on the whole you will find that the authors of UNIX, Ken Thompson
and Dennis Ritchie, have created a program of great strength,
integrity and effectiveness, which you should admire and seek to
emulate."
---
pb Reply or e-mail rather than vaguely moderate [152.7.41.11].

Re: Sensationalism

[mdvkng]

> I really would hope that Slashdot isn't starting to pander to the lowest common denominator like
> television does. I try to have more faith in the people here.

Same here, so here goes my attempt at a possible explanation.

Speaking for myself, deeply technical articles cause me to read attentively instead of faster skimming. In addition, I like to look at more of the comments before I say anything, again because of interest but also to make sure my statements are not completely d'uh.

In this particular article, I read very carefully because, while I am no kernel hacker, as a technologist I do find the technical details of kernel hacking extremely interesting. Maybe someday I will do more than just look, but for now it's enough for me.

This makes any posting I may make more cautious in nature and I may not bother at all, choosing to leave the discussion level to those more qualified than I, and thus maintaining the intelligence of the overall discussion.

Perhaps I am not the only one to do this?

Yes, perhaps.

-M

You're not the only one.

[Tony-A]

After carefully reading the article (usually I read only the comments), all I can say is wow, well done. This post is to add a small voice that the small number of comments is not an indicator of lack of interest. Any of the inside history of how Un*x has managed to survive and even thrive is, I think, of great interest to us all.

Re:Literary criticism and programming?

[apsmith]

Oops - you're right. Knuth has been at this a lot longer than Larry Wall... Though, has anything useful other than TeX ever been written in Web?

Re:De-compilers: not likely

[apsmith]

> THAT is the problem. The compiler has striped away the usefull [symbolic] information.
> How is the de-compiler going to distinguishing between a pointer to a pointer and an array of poitners? etc.

But if the machine code is the same, then the original source code COULD have been written either way; it would produce the same correct result. Does it really matter if it's an array or a pointer? If there is bounds-checking going on then that's there in the machine code one way or another. The actual "content" of a program can legitimately be viewed as just the instructions the computer has to execute: the binary machine language version. As in perl, there's always "more than one way to do it" in the source code to get the same result (or essentially the same) in machine code. Anyway, that's why I suggested this would require real artificial intelligence: it's not an easy problem. But logically it could be done. It really is close to some of the issues with natural language translation.

By the way, I've programmed in assembly language in the past, and looked at "decompiled" binary code (it's trivial to get from binary to assembly language) - it's usually pretty horrific, but sometimes quite instructive. Just about any program COULD be turned back into C, but it would probably all be one long function full of goto's, arbitrary and frequently re-used variable names. If that was released as the "open source" version would it fly?

Re:De-compilers: maybe

[UnknownSoldier]

I don't disagree that decompilation can't be done. Heck, us humans do it, so it is possible. ;-)

The questions is, can we get a machine to re-engineer a binary program and turn it back into something semi-readable, faster than it is to just have a person reverse engineer it.

Do-able? Technically.

Usable? Not really. Well not yet at least.


> I've looked at "decompiled" binary code - it's usually pretty horrific, but sometimes quite instructive.

Have you looked at compiled Delphi and Visual Basic code. Horrific to say the least.


Cheers