Andrew Lahde Timed the Market, but Missed the Mark on Linux 48 comments
-
Font Size:
-
Print
- TweetThis
Heavy thinkers in the blogosphere are both commending and attacking Andrew Lahde for timing the financial meltdown correctly and apparently making a bunch of rich guys richer (by getting them out of the market at just the right time around mid September 2008). My only beef with him is a suggestion in his now infamous farewell letter that says (according to Alphaville):
My suggestion is that (George Soros) start and sponsor a forum for great minds to come together to create a new system of government that truly represents the common man’s interest....This forum could be similar to the one used to create the operating system, Linux, which competes with Microsoft’s near monopoly.
He's got the genesis and status of Linux vis a vis Microsoft (MSFT) so wrong that I wonder how much of the other facts in his letter are accurate.
Linux was created by* IBM, HP (HPQ) and other former IT systems monopolists that realized that Microsoft was taking their systems monopoly away from them. IBM, HP, Digital Equipment (now part of HP), etc. had banded together for this purpose in the early 1980s while Linus Torvalds, the nominal creator of Linux and who now works for one of the groups IBM, HP, etc. put together for its trust-like purposes, was still in short pants. Ten years later, the consortium chose a small piece of software code, "forked" by Linus from some other code while he was in college, to complement the still ongoing technical development effort by IBM, HP, etc. to come up with "one Unix." What is today called Linux is the result of that one-Unix effort.
So far, from a marketing perspective, Linux has mostly replaced the old monopolists' own software (AIX, HP-UX, Digital UNIX, MVS, VMS, MPeIx, etc.) and has not slowed down the new monopolist Microsoft at all.
As for Lahde's suggestions relative to the use of marijuana for non-medicinal purposes, I humbly suggest that his letter was the result of such a session.
(*"Created by" is not an accurate way in technical terms to describe the development of any software for a variety of reasons, but I will leave it be for purposes of this blog.)
Related Articles
|
Find People on SA
-
The Seeking Alpha 100
See allThe top 100 stock
market authors
selected for publication
Seeking Alpha 100 » -
Fastest Climbers
Authors with the largest
increase in followers
in the last 7 days-
1
David Fry
-
2
Bill Gross
-
3
Philip Davis
-
4
Tyler Durden
-
5
TraderMark
Fastest Climbers » -
1
-
Top Commenters
Top commenters,See all
as ranked by their peers
Top Commenters » -
Top Instabloggers
The 100 most activeSee all
Instabloggers.
Top Instabloggers » -
Top StockTalkers
TopSee all
StockTalkers.
Top StockTalkers »
This article has 48 comments:
Or is this a humor piece, and I missed the joke?
As for other facts in Andrew Lahde's letter, surely those can be judged on their own merit?
Mr. Byron - have you ever heard about something called 'research'?
You should try it.
whereas Lahde *might* have partaken a bit of the healing herb, you are clearly on crack.
The stench of your bullshit riseth to the level of the SUV windows which are presumably inhabited by your intended audience.
Hopefully most of them are by now so broke that they can no longer afford to do any more of the jocaine you're pushing ...
The author of this article does not have a clue and definitely does not check his facts. I bet Linus Torvalds is having a good laugh over this. The fsf did not come along till Linux was already first developed by Linus. IBM, HP, and the others were the beneficiaries of Linux development. To my knowledge they had nothing to do with the original Linux development. Later IBM, HP, and the rest certainly took advantage of Linux to break the stranglehold MS once had. Open source rules!!!
You do have the excuse that most people think of that suite as "Linux" - wrongly; there were and are many projects using the same tools which look and work in a similar manner but use a different kernel (such as Apple's MkLinux, a precursor of OS/X, which actually used a Mach kernel). But up to that point you committed no sin other than ignorance.
To say that "What is today called Linux" was " "forked" by Linus
(Torvalds) from some other code while he was in college" is more than demonstrably wrong; it is defamatory, and if there is actual malice you will be lucky to get away with your shirt, or short pants.
This is a shame. There was a somewhat serious point you could have raised, which is that the great majority of commits to the kernel are now written by permanent staff of major corporations like those you mention.
That does not help your diatribe against Lahde, however, first because he got his fact right, and second because what ultimately counts is not who paid the salaries of the coders but rather, the licence the code is released under.
So the FSF preceded Linux, not the other way around. Had it been otherwise, the kernel would have had no GPL to be released under. On the other hand, the Posix tools were grafted on to the kernel, not the kernel on to the tools, so in a conceptual way you are right.
Mr Byron is probably thinking of the X/Open group when he speaks of the "early 1980s" consortium, but even that he got wrong because X/Open was originally a European idea and the US companies like IBM and HP only joined around the time the Linux kernel was released.
Next thing you know he'll be telling us that the Internet is a series of tubes...
The whole article is so silly, but at least partially entertaining. The above quotation I found hilarious, because that would suggest this article is a product of some serious deadly substances. Or maybe this is an distorted expression of frustration when confronted with the fact that the holy cow of a free market is proven to be flawed? Or maybe it's based on the view that nothing bright ingeniously can be created outside the walls of mighty corporations?
Mr. Byron, brains isn't the exclusive privilege of corporations and their staff; money can't by brains. Hence don't be surprised that your little history lesson doesn't give you good grades.
The GPL existed before Linus T. created the first version of his kernel. The movement of a GNU operating system was started 7 years before Linus T. started to build a kernel that could make use of free software. Linus T. made use of the GPL license which is proof enough of him not being backed by corporate interest. The whole point of GPL is to protect it from being corrupted by greedy corporations, but not hindering them from getting on board as long as they honour the license. Thus IBM or any other company doesn't magically become good, or get their greediness cured, but their interest and involvement is regulated by a smart license, and the process is kept transparent.
It's a pity we need to waste our time responding to articles of such low quality. Unfortunately it's necessary because to many read and uncritically believe anything they find on Internet.
I have only one question...since IBM, DEC and HP already had licenses to the UNIX source in 1990, why on earth would they need to rewrite it *again* from scratch? Each of these major players had already written one or more proprietary operating systems, yet they felt the need to borrow code from a Finnish college student, and rewrite an OS for which they already had licensed source code?
As Linus would say, "I want some of what they're smoking"
the mid-80's in an effort to speed it up and make it usable. The
speedup was 3X but that still wasn't enough to run a 9600 baud
modem. The message-passing architecture was fundamentally
too slow. Tannenbaum even advertised this in the form of one of
the chapter problems; the goal was modern computer science,
not speed.
Linux did not start with a message-passing architecture. Linus
went after speed and UN*X compatibility from day one. From scratch.
Though, if you read it carefully you will find that this loon is almost successfully skirting withinin the bounds of fact, almost.
Yes big companies tried in the past to have a compatible UNIX, he even missed out Microsoft who was nominally on the fringes of that one, but he says so in such a way as to conflate what failed effort with what happened next. This was not an anti-microsoft thing, Microsoft was not a player in those markets.
Linus Torvalds created Linux. It is not a fork of code from anything else. The lie that was out there before, it is just a version of Minix, has been refuted by the creator of Minix, Andrew Tannebaum, himself. Not only in that but he has been involved in a long term disagreement about the way to build kernels with Mr Torvalds. Their ways of achieving the result are diametrically opposed.
IBM amongst others have helped Linux develop, but only that. They put into Linux ideas they had developed for other Operating Systems, like OS/2, ideas original to themselves.
The parent article is a transparent attempt to make a conspiracy out of events where none exist. It is a fabvrication and, in effect, a fraud
So assuming he got paid, he should realize that "the big lie" works because it's plausible. Because it's not easily refuted. Such as damn near everything this author has stated in this article.
Just though of something - the only one that actually makes sense. But my god man, if you're going to post while stoned *and* drunk, at least have the decency to retract and/or correct the article when you wake up the next day.
www.groklaw.net/articl...,
for an article that makes clear the actual history of the relation of Linux to the large companies. Mr. Bryon has made some mistakes.
In other news, earth only 4000 years old, and tooth fairy real.
Some of what you say is somewhat true (e.g., some large Unix vendors banding together in the 80s trying to make their divergent Unix flavours more compatible), but that effort was not in any way related to the GNU/FSF effort to write a free-as-speech Unix equivalent from scratch or Linus' writing the Linux kernel (also from scratch).
Most traditional big Unix companies wouldn't touch GNU/Linux with a ten-foot pole until the late 90s. If you don't believe me, you can check for yourself - the advantage of open development is that the source code repositories (including who contributed what when) and the developer mailing lists are available for all to see.
What happened was that GNU/Linux grew as a free grass roots project, while at the same time the big *nix vendors' efforts at creating a more unified commercial *nix market failed. Eventually the big *nix vendors realized that putting their weight behind Linux was in their self interest.
Trying to cast the development of Linux as some "great gameplan" from HP/IBM/DEC in the 80s is completely contrary to what really happened.
Cheers!
Best regards.
Here is an exerpt from an article by Tufty on Groklaw.
Caldera used to make publicly available the documentation to OpenLinux, which was also released under the GPL by Caldera, and while they've removed it from the Internet, I quoted from it back in 2003 -- and yes, I think there is a connection to my article and its disappearance thereafter -- and here's part of the history section:
"Linux was started in the early 1990s as a small research project by a Finnish college student named Linus Torvalds. Soon after Linus started his project, hundreds of others began to participate in its development via the Internet. A cooperative venture grew in which thousands of people were working together to create a new operating system. The inclusion of the GNU utilities from the Free Software Foundation (see www.fsf.org) and the release of Linux under the Gnu General Public License (GPL) furthered the spread of this work. The GPL provides that the source code to the software is released with the product and that no one can restrict access to it. Software licensed under the GPL license is sometimes referred to as Open Source software. With this type of software, anyone can examine and extend the source code, but all such work must be released for public use. Other licenses provide for inclusion of source code with its associated software, but to date the GPL is the most common Open Source license."
Unbelievable to some. Enraging to others. But that's really how it happened.
So, yes, the "history of Linux" Byron offers is very silly. So many histories. So much defending of Microsoft's proprietary ways. But since Byron wrote what he wrote, and Google will collect it, I deliberately wrote this article so that when folks search for a history of Linux and come across his silly stuff, hopefully they'll find this accurate information too.
I have been using Linux since 1994 and I have never seen such an innacurate statment about Linux in print since SCO started litigation.
I sugest the next time you decide to print something so far from reality, you take a step back and go to the sources for where things come from. A quick email to Linus Torvalds or a quick session with a search engine would do wonders for your credibility.
However, I will make sure to make a copy of this to use for April Fools Day 2009!!
The GNU foundation and the FSF had nothing to do with OSF, and should not be confused with that effort, simply due to the similarity in initials. OSF wanted to recreate Unix so that they could compete with the creator of Unix. The FSF wanted to recreate Unix so that they would be free to modify it and give it away. Two totally different mindsets, two totally different goals.
Since when does publishing an outright lie constitute anything other than libel?
I'm sorry that the truth is hard to bear, Mr. Byron, but the person who missed the mark is you -- entirely.
Best of luck giving your degree back to K-Mart. I'm sure there's a blue-light special on new degrees.
Ehud
Talk about 'The pot calling the kettle black' MVS is a proprietary O/S from IBM. VMS is a proprietary O/S originally fro DEC/Digital and neither has anything to do with Unix's
Unix was created at Bell Labs but extensively added to by U.C. Berkeley using an open source license. AIX, HP-UX etal are licensed proprietary derivatives of Unix.
Minix was written from scratch by Andrew Tannernbaum and is not a derivative of Unix.
Linux was written by scratch by Linus and is not a fork of Minix or a derivative of Unix.
Given your track record for 'facts' in this article, nobody should ever trust your facts again.
And, good luck if the law suits start rolling :-).
/ikh
Consider the Linux operating system source code first:
Linux 1.2 [Weight: 2.64MB] Release date: 1995/03/08
Linux 2.0 [Weight: 6.55MB] Release date: 1996/06/09
Linux 2.2 [Weight: 14.8MB] Release date: 1999/01/21
Linux 2.4 [Weight: 27.7MB] Release date: 2001/01/04
Linux 2.6 [Weight: 49.4MB] Release date: 2003/12/17
See www.krsaborio.net/rese... for more details.
It's easy to figure out the lines of code contributed by IBM, HP and other companies since 1999 by looking at the weight and complexity in MBs of each Linux release.
Therefore, Dennis Byron asserts:
"Linux was created by* IBM, HP (HPQ) and other former IT systems monopolists that realized that Microsoft was taking their systems monopoly away from them."
Linux, in it's current form, was obtained thanks to the millions of dollars and resources provided by IBM, HP and other companies.
The first part of the following assertion clearly refers to the Open Software Foundation:
"IBM, HP, Digital Equipment (now part of HP), etc. had banded together for this purpose in the early 1980s..."
For information on the Open Software Foundation see www.krsaborio.net/rese... .
Dennis Byron then continues by getting Linus Torvalds involved in his assertion:
"... while Linus Torvalds, the nominal creator of Linux and who now works for one of the groups IBM, HP, etc. put together for its trust-like purposes, was still in short pants."
Indeed, Linux in 1991 was still in short pants. It might help to hear it from Alan Cox himself, www.intercom.co.cr/med...
The following sentence isn't entirely clear:
"Ten years later, the consortium chose a small piece of software code, 'forked' by Linus from some other code while he was in college, to complement the still ongoing technical development effort by IBM, HP, etc. to come up with 'one Unix.'"
One of the problems with the above sentence is "the consortium chose a small piece of software code".
www.krsaborio.net/rese... and www.krsaborio.net/rese... clearly demonstrate that IBM, HP and other companies contributed millions of dollars and resources on improving Linux 2.2 so it would perform properly within the enterprise.
In 1999, the weight of Linux 2.2 was only 14.8MB. The first robust version of Linux for the enterprise was version 2.6 and its weight was 49.4MB.
So how small was the small piece of software code chosen by IBM, HP and the other companies?
Dennis Byron concludes:
"What is today called Linux is the result of that one-Unix effort."
It's probably correct taking into account the millions of dollars and resources provided by IBM, HP and others.
Dennis Byron finishes his assertion with the following paragraph:
"So far, from a marketing perspective, Linux has mostly replaced the old monopolists' own software (AIX, HP-UX, Digital UNIX, MVS, VMS, MPeIx, etc.) and has not slowed down the new monopolist Microsoft at all."
Has Linux, as we know it today, slowed down Microsoft?
That talk by Alan Cox was given on April 20, 1999 in Reykjavik, Iceland. See www.krsaborio.net/rese...
To obtain a better perspective of GNUs participation, lets listen to Richard M Stallman and a talk given on September 27, 1997.
GNU, GPL issues: www.intercom.co.cr/med...
Linux, Hurd issues: www.intercom.co.cr/med...
Stallman's talk was given at the Individual Network e.V. Convention in Aachen, Germany.
Cheers!
Um, I'm in awe. I didn't think anyone could screw up this badly in an article, at least not since Ken Brown of ADTI.
If you were trying for page hits Dennis, you probably succeeded, since many like myself will have dropped in to see if you actually wrote what was reported. For accuracy I'll grade you with an F, and I suspect that your public school teachers would be aghast at how badly written this is.
"Dennis Byron assertion isn't entirely wrong..."
This is like saying that it is not entirely wrong to claim that the Moon is made out of green cheese.
First of all, it it quite a remarkable claim that the increase of kernel weight is evidence that "Linux, in it's [sic] current form, was obtained thanks to the millions of dollars and resources provided by IBM, HP and other companies."
By that argument, it could just as reasonably be said that Linux, in its current form, is obtained thanks to the millions of man-hours provided by the US Navy, European governments, and university students.
Or any other random group of kernel contributors.
As your numbers show, the kernel has grown steadily, roughly doubling or tripling in size every year or so. If anything, the growth rate has declined in the later years.
The growth is caused by many contributors, and IBM and HP are responsible for far from all of the growth. I suggest that your "study" should be expanded to examine _who_ contributed how much code, and in which part of the kernel. Naturally, IBM and HP has contributed a fair bit of driver code, so that the hardware produced by them can be used by Linux. But that is not the same as "Linux in its current form".
Secondly, just because Byrons statements might refer to OSF, it does not mean that they are not entirely wrong. Linux was neither developed nor planned by OSF.
If someone said that the iPhone was made by old, British musicians, they would be entirely wrong, even though The Beatles were from the UK and had a record label called Apple.
As for your idea of Linux in short pants, I must point out that Byron said that Linus (the person), not Linux (the kernel) was in short pants in the 1980s.
And, "the consortium" did not choose Linux. Early distributors such as Slackware, chose GNU software to go with their Linux distributions. And commercial distributors like Red Hat chose GNU and Linux - and many other software projects - as a basis for their business.
Only later on, in the late 1990s, IBM, HP, and others decided that it would be good for them if all their hardware ran with Linux. At that time, the choice had already been made by others.
You say that it is probably correct, when Byron says
"What is today called Linux is the result of that one-Unix effort."
But the original "one-Unix" effort was aimed at creating a commercial Unix. In the beginning, that was a direct competitor to Linux. So Linux as of today is here _in_spite_of_ that effort.
Granted, Linux would not have gotten quite as far now without contributions from (amongst others) IBM and HP, but it would still be here.
On the other hand, Linux would never have been here, if it was _created_ by IBM and HP.
In the end, you ask:
"Has Linux, as we know it today, slowed down Microsoft?"
It has never been the purpose of Linux to slow down Microsoft. But still, it has.
Had there not been real alternatives to Microsoft, it might not have been relevant for the EU Commision to examine and fine Microsoft.
If GNU/Linux did not run quite well on desktops, companies all over the world would not hesitate as much in switching to Vista. Simply because Microsoft could then just kill off XP without fearing losing customers.
And, the server marked would have been far more infected by Windows.
IBM released an interesting study in January 2003 titled "Improving Linux kernel performance and scalability".
See www.krsaborio.net/rese...
Figures 1 and 2 show the increase in performance. The first graph shows the performance of different versions of the Linux kernel from May 15 2001 thru June 26 2002.
www.krsaborio.net/rese... shows the most important contributions of IBM in the form of patches to the Linux kernel.
While the contributions of non-profit organizations and University students were important, the millions of dollars and resources provided by other companies including IBM and HP were key to obtain the performance of Linux as we know it today.
Non-profit organizations and University students didn't have access to multiprocessor machines to test and improve the Linux code. www.krsaborio.net/rese... shows some early messages that illustrate the lack of access to multiprocessor machines.
Therefore, IBM and HP made those machines available to non-profit organizations and University students to help develop a suitable solution to replace their aging proprietary Unix operating systems.
"Follow the money" is a good approach the understand the impact of IBM and HP in Linux as we know it today!
"The performance of Linux in the enterprise is a cool way to measure its growth since 1999."
It is? That's a non-sequitur. It does not neccesarily follow that growth (whether in source code size or in usage) correlates with performance improvements, so that is a flawed premise which you cannot therefore use to support your (flawed or at best biased) view of Linux's history.
Please listen to Alan Cox talk about the status quo of Linux in 1999. The date of the following talk is April 20, 1999:
www.intercom.co.cr/med...
Cheers!
Yes? Your point being? I'll note (as per Alan's speech), the 2.2 kernel was already out by the time he made his speech (It was probably tagged as 2.2 near the beginning of that year, and in any case had been in development for several years in the form of the 2.1 "development" kernel tree.) Furthermore the "production" 2.0 kernel series was as Alan put it, the one "being released by the vendors" [i.e. at that time] and so was also clearly ready for, and in mainstream use, used in business etc. at that time.
As he says (at 4:38) "Linux 2.0 is the actually the real Linux for business, the Linux for commercial use, and Linux 2.2 is the latest Linux, it's much more scalable, multi-processor machines."
The IBM report concluded:
"The Linux scheduler needs to be modified to more efficiently support large numbers of processes. This is a large system scalability feature that many people believe is required in order for Linux to become enterprise ready (see "Making Linux a World-Class Enterprise Server OS" on The Linux Scalability Project home page). This is not just a Java technology issue. Other applications also wish to make use of threads to improve performance."
"The Linux kernel needs to be able to support a many-to-many threading model. At the moment, the Linux kernel cannot support a many-to-many threading model (see, for example, a FAQ on kernel threads). A many-to-one threading model solves the scheduler bottleneck, but has other problems including lack of SMP scalability for the application, or insufficient thread state to support some application requirements..."
The above was reported in January 2000.
Throughout the years 2000, 2001, 2002, the Web was full of discussions about the instability of the Linux kernel mainly related to SMP, Virtual Memory management, and related issues. See page 27 of www.krsaborio.net/rese... for a list of these issues.
Some of those discussions were:
2000-01-27 Strange scheduling behaviour in SMP www.krsaborio.net/rese...
2000-05-08 Recent Virtual Memory fiasco www.krsaborio.net/rese...
2000-05-15 Virtual Memory stable again? www.krsaborio.net/rese...
2000-06-28 A joint letter on low latency and Linux www.krsaborio.net/rese...
001208 Linux Scalability Effort Proposed Roadmap www.krsaborio.net/rese...
2001-01-07 Low-latency scheduling patch for 2.4.0 www.krsaborio.net/rese...
2001-04-03 A quest for a better scheduler www.krsaborio.net/rese...
2001-09-15 Virtual Memory problems www.krsaborio.net/rese...
2001-09-16 Major Virtual Memory merge www.krsaborio.net/rese...
2001-10-05 Google's mm problem www.krsaborio.net/rese...
2001-11-30 Linux, what changes are needed? www.krsaborio.net/rese...
2001-12-10 2.4.16 & OOM killer problem www.krsaborio.net/rese...
2002-01-03 Ultra-scalable SMP and UP scheduler www.krsaborio.net/rese...
2002-02-14 Linux 2.4.17 VM problems www.krsaborio.net/rese...
2002-04-14 VM updates for Linux 2.5 www.krsaborio.net/rese...
2002-06-19 VM with reverse mappings for Linux for 2.5.23 www.krsaborio.net/rese...
020626 Buried alive in patches: Picking up the pieces of the Linux 2.5 kernel www.krsaborio.net/rese...
2002-12-12 VM with reverse mappings www.krsaborio.net/rese...
2003-01-xx IBM: Improving Linux kernel performance and scalability www.krsaborio.net/rese...
A full timeline of events is found at www.krsaborio.net/rese...
The above shows why the Linux kernel wasn't enterprise ready in the years 2000, 2001, and 2003. While the kernel was used in some enterprises in those years, maintenance costs were quite high due to the issues mentioned on page 27 of this IBM report www.krsaborio.net/rese...
In conclusion, Linux 2.6 released in December 2003 was the first Linux kernel ready for the enterprise with much less maintenance costs than its predecessors.
With all due respect, your view of history is far too black and white I think which is why you seem to end up with some odd conclusions. You seem to talk about "enterprise ready" as though it's some sort of black/white binary 1/0 type fact, either on or off, yes or not. The reality was rather more gradual/grey. The fact is the Linux kernel was already in use in businesses (which is after all "enterprises") since the mid 90's, as I've pointed out already, and as you actually admit above as well.
Sure there were issues and these were worked on as parties interested in those issues started helping out on the kernel. This is nothing new, and is how the kernel development has proceeded prior to Alan's speech in 1999 as well as after. There's wasn't suddenly a magical moment when someone flicked a switch (with the release of say 2.6 kernel or any other version for that matter) and *everyone* then started using the Linux kernel in earnest in "the enterprise" (whatever that means) where *no-one* used it before.
Rather there was a continuum of development, improvement and slow gradual increasing use, as people (and businesses/enterprises... became aware of Linux as an alternative, as it improved and as it became more capable in more niche areas they were interested in. Even so, as I've said, businesses (enterprises), were already using Linux since roughly the mid 90's which is an undisputable fact. So, It would be wrong to think that Linux was "nothing" before big companies became involved, which is what you seem to be syaing. That they've made sizeable contributions is obviously however true.
Tell me, what was the first version of Linux you used? For me, it was one of the pre 1.0 alpha versions (think it was 0.98), which I downloaded onto floppies, and installed onto my (at the time) 386sx machine. I continued using Linux, watching it grow, watching it starting to be used a few years later to run an entire computer science undergraduate department including FTP, mail, web, compilation on a multiprocessor Pentium Pro machine (at the time cutting edge) alongside "bigger iron" boxes already in the department, including Sun's, SGI's and PowerPC's amongs others, eventually watching it being ported to other architectures (initially DEC alpha and then later others IIRC) including Sparc, as the 90's progressed, watching a friend start up an ISP using only Linux in the mid 90's (for which it was even then already eminently suitable) which is what they use to this day, watching Mark Shuttleworth (now of Ubuntu fame) start up his first company in the same timeframe, again using only Linux, eventually to sell the company for profit in 1999 somewhere (as I recall) close to the dot com boom. And so the list goes on.
As I said in the beginning, this is not the black and white question you're seem to think it is.