letters to the editor

OSS: Responses to Matthew Craighead
From Tim O'Reilly <[email protected]>:

Matthew, I wanted to urge you not to give up on Open Source yet! While it's true that there are a lot of people on slashdot (and around MIT, home of the FSF) who do in fact exhibit the values you describe in your letter, I don't think that this is true of the Open Source community in general. In fact, one of the whole reasons for the attempt to change the popular "meme" from Free Software to Open Source was to get away from some of the misconceptions you decry.

This has certainly led to some splits in the movement. For example, I am currently in the process of setting up a second "Open Source Summit" (to follow the Freeware Summit I organized last year), and I got heavily flamed by some people who thought I was out of line for inviting the Jini folks, even though their license isn't truly "free," as well as for suggesting an agenda that was looking to explore how much it is in fact licenses and how much it is network effects and other factors that account for the spread of popular open source programs.

But in the end, the people espousing the radical positions you decry in your letter are a small minority. Of the approximately 40 people coming to the summit (and these are about 50/50 split between developers of major Open Source software projects and people from corporations working with Open Source), only 2 or 3 were the source of the flames. And they ended up taking their marbles and going home, realizing their views were not shared by the vast majority of the attendees.

Despite the attempts of Richard Stallman to portray Open Source as a misguided attempt to recast the vision of free software, the Open Source movement is actually a recognition of the fact that a huge percentage of Open Source software developers don't care much about ideology. They care about getting a better job done faster.

If you read Eric Raymond's paper "The Cathedral and the Bazaar" (<http://www.ccil.org/~esr/writings>), you'll see that it's really about the principles of distributed, community-based software engineering, not about free software ideology. (You might also want to look at my special issue of Esther Dyson's Release 1.0 newsletter, at < http://www.edventure.com/release1/1198.html >, which provides a big picture overview of Open Source.)

What's important about Open Source is that it's a recognition that what drives the success of Linux, FreeBSD, Perl, Apache, sendmail, and a host of other hugely successful Open Source software products is not ideology but science — software engineering methods and economic models suited to today's networked world.

From Con Zymaris <[email protected]> :

I suspect you may have received a barrage of email on the publication of Matthew Craighead's letter in the recent ;login: [February 1999]. Here's my take at a response/refutation:

I believe that this letter should not have been published in a journal like ;login:. It belongs in the hodge-podge of idea-and-flame cauldron that is Slashdot. While I have no problems with the notion that Matthew Craighead has a right to his opinions, that doesn't mean they should get picked up and placed as prominently as they were in ;login:, as the sole letter published in the journal. Why? Because the ideas are half-baked, and no better than most of the ideas espoused on Slashdot that the author so fervently decries.

Let me point out a few:

1) "Most Linux fans dislike Microsoft." Likely to be true, but then most serious IT professionals not in the Microsoft-dominated realm also dislike Microsoft. So what?

2) The idea that OSS is a reincarnation of Communism is rubbish. If anything, the current system of either monopolistic or oligopolistic software monarchies is the equivalent to feudal and nondemocratic societies. Why? Because OSS is revitalising an otherwise tired and grey industry. It, through making available the source, is making dozens of startups spring forth as viable system, platform, and infrastructure software/service providers, all with a common, standardized, open-protocol base. This is real competition. As Bill Gates has famously stated, operating systems are a natural monopoly. If so, I want my operating system to not be controlled by one, single, all-powerful vendor. I want many OS source options. This is true competition. The current regime exudes an aura of "all edicts come from the centralized, Seattle-based politburo, for their own financial gain; take-it-or-leave-it." OSS is our (the users') way of "keeping the bastards honest."

3) There is a spread of individuals and groups involved in OSS. Very few are anti-commercial. What they are about is giving users, not corporations, more rights. Free software (I'll use the term interchangeably with OSS, I'm not religious about it) can and should be provided with commercial support, which users can then take or leave. Even the MIT-spawned FSF is perfectly happy with this.

4) Repeating what various anonymous people said on Slashdot is a pointless exercise. So what if someone said something crazy? Slashdot is a (marvelous) discussion platform for geeks, not the planning committee for OSS!

5) Obviously Matthew hasn't noticed that there are many companies and people making good money from supporting OSS. Cygnus, Red Hat, S.u.S.E., etc., are all profitable, and they all pay Linux developers with real money. There's no reason why a good Linux developer needs to be a "starving" Linux developer.

6) The point of the Halloween memos wasn't that Microsoft was unreasonably distressed by OSS; it is that they actually were distressed by OSS at all, which in itself is a major revelation. At that point, it looked like it had not even noticed Linux. Midway through last year, Microsoft was looking like it would have a perpetual reign as the IT industry's self-anointed Emperor; the thousand-year Reich. And along comes a band of "merry-men" (their words) who looked like they had a chance to unsettle this plan. Microsoft had analysed the threat, and found it to be real, and that none of their dirty tricks, FUD, breaking compatibility, proprietary protocols, etc., would work to defeat it. Did you actually read the memos? You will find that the analyst was talking about "breaking" open Internet protocols and the use of patents as their only chance of killing OSS off. Personally, I equate "breaking" Internet protocols as verging on criminal, and here was Microsoft planning this strategy as a fairly humdrum action. The selfishness and gall of this concept defies description. You wonder why people don't like them.

7) Linus Torvalds gets the best of all worlds. He's adored by the professional/technical IT community worldwide, gets to work on writing technically challenging free software OS kernels (his passion), and yet gets paid handsomely by Transmeta to do this and other work. He is not, however, in a position to kick himself for not having made millions off Linux. He himself has stated that the Linux kernel makes up only 1% of the Linux OS, and of this, only 5% is his work. Linux would never had gotten off the launch pad if it had not been developed and released as GPL OSS. The work is the blood, sweat, and tears of a million people worldwide, not one man.

8) OSS advocates do not want to make intellectual property illegal. Almost all, however, want to make software patents illegal. Many far more eloquent observers than I have written on the evils of software patents. Read them.

9) While it is true that Red Hat et al. have a business model that means that they don't have to create Linux from scratch, that is one of the tenets and strengths of OSS, not a weakness. Remember, a good programmer writes good code; a great programmer copies great code, and doesn't reinvent the wheel. This is a primary reason why OSS has mushroomed and developed great systems and apps quickly. Further, Red Hat, S.u.S.E.. and Cygnus have all done their part in bringing out new code and useful extensions to Linux and GNU development tools, such as easier installation, GNOME, X-Free video drivers etc., all released as OSS.

10) You assume that people only develop software for monetary gain. This is an incorrect assumption. While OSS (in its extreme) may not attract the types of people who primarily focus on money, it will attract others who do it for reasons like contributing back to the OSS community, altruists, the talented and curious. Further, your comments are based on the premise that "closed-source" vendors actually make profit from their non-OSS software. This is mostly false. The advantages of OSS are that you can leverage existing code, produce a useful tool quickly, get it out the door for others to solidify and extend, thus cutting out 80% of the effort/cost required to create closed-source "commercial" software. Included in this cost are debugging and beta-test programs, paying for sales/admin/support/marketing staff, paying for marketing/advertising, paying for packaging, paying for printing of manuals, etc. Getting a product to "production" quality is less than 50% of the effort required for commercial vendors. Making profit from closed-source software is not easy.

11) MIT has already become famous amongst the digirati as the birthplace of GNU and FSF. It is one of the "hallowed halls." Enjoy it while you're there.

Lastly, I want to add that while there is a strong notion, particularly amongst the Linux community, of open and fair advocacy, not all users abide by this, and flame mail may be generated. As with all communities, a spectrum of opinion and temperament exists. Don't take the flames from the bad apples to heart, and don't judge the community as a whole on the actions of a few hotheads.

From Prevelakis Vassilis <[email protected]>:

I was appalled to read Matthew Craighead's letter.

I do not think that slashdot.org represents all Open Source proponents. In fact until I read about that particular site in the letter, I was oblivious to its existence.

Now, more to the point, I am afraid Matthew made a logic jump. slashdot IS pro-OSS AND slashdot IS crazy DOES NOT IMPLY pro-OSS IS crazy.

For example, I am sure that there exist Web sites run by fascists or other extremists. If I visit such a site I may also find material about how great America is. Does this imply that anybody who is pro-American agrees with the rest of the material in that site?

I am pro-OSS so my opinion may not count, so let's look at what a staunch proponent of capitalism such as the Economist magazine has to say about OSS ("Computer programming. Hackers rule," 20-Feb-99; see <www.economist.com>).

Open-source programming is more like academic work than business. And just as the disclosure of theories and empirical data usually produces good science, so published code leads to better software. The programmers are motivated not chiefly by money, but by reputation. It is a coup to write "patches" that pass the scrutiny of fellow hackers and get incorporated in the next version of a program. . . . . Yet there are drawbacks. Big software companies have every reason not to go open-source. Hackers might also not be keen to work alongside the likes of IBM and Sun; many are strongly anti-commercial. There is also the danger of forking when a group falls apart and incompatible versions of a program emerge as has happened to one operating system, BSD Unix, when personality conflicts led to splits. Few managers will bet their companies on the product support they receive in news groups on the Internet.

Hence the importance of the commercial fringe to open-source software. Numerous service companies, such as Caldera, Red Hat and S.u.S.E, have built a business out of making Linux easier to install. Eric Allman, the benevolent dictator of Sendmail, has set up a company that supports the open-source development of the program, while selling a commercial version and services to support it. . . . It is too early to say whether such approaches will work. But open-source is here to stay. Perhaps the software industry will eventually look a bit like a highway. The infrastructure (operating systems, networking technologies) will be largely a public good, while services (support, training) and specialised applications are for sale. Just don't expect Bill Gates to like the idea.

I find myself in full agreement with that article. Besides, capitalist dogma also states that the market should be left to decide what is best. So let's wait until the "horrifyingly large number of Linux fanatics" at MIT graduate and join companies and then we'll see whether their adoration for OSS is commercially justified.

As far as the tone of the letter is concerned, I think that if people disagree with any concept, they can express arguments against it. Calling the other side communists or other names does not help the discussion.

From Yiorgos Adamopoulos <[email protected]>:

It is clear to me that Matt has not worked to earn his living as a CS/CEng professional. What Matt seems to be missing is that people support OSS because it is better than corporate software (in many cases, not all, not even most; I have yet to see something that competes with Excel, for example).

What also Matt seems to miss is that in any ideology (and OSS is an ideal) there exist fanatics (like the ones he saw in the slashdot posts). Working for "love and fame" requires a total change of the day-to-day model that the world has. In such a world clothing/eating/housing is a solved problem. I for one do not believe that we are going to have such a world ever (the ancient Greeks didn't make it and it was easier then ;-).

Matt, your motives in life show in your letter: You are into money and fame. Well, simply put, not all humanoids are. You seem to miss the fact that if the other side "wins" the "race" there will be a total change in the society, so the logic of "starving," "no TV," and "no radio" do not apply ("no TV" is a benefit to the society anyway).

Ask yourself how Linus Torvalds makes a living today — if he needs more — and enjoy your university years, because they will undoubtedly be the best in your life. Also learn to give to others, you will always get more back (that is what OSS is really about IMNSHO).

From Marty Leisner <[email protected]>:

Matthew Craighead's letter bothers me. While lots of people says outrageous things about OSS (or Free Software as it used to be called), lots of people say outrageous things about everything — from politicians to televanglists to the guy on the street corner — you have to separate the wheat from the chaff in unmoderated discourse and quote people who have useful opinions.

While many consider Richard Stallman's views to be somewhat socialist, I don't recall him espousing anti-capitalist views. While he called for a software tax (which I don't support) his attitude is very simple: "I consider that the golden rule requires that if I like a program I must share it with other people who like it." He later says, "By working on and using GNU rather than proprietary programs, we can be hospitable to everyone and obey the law" (GNU Manifesto, 1985, <http://www.fsf.org/gnu/manifesto.html>).

It turns out a small percentage of the world's software industry write "off the shelf" applications. Most software work is inhouse labor. What I find maddening is that proprietary applications rarely provide the source code. From working with software for two decades, I know that having the source code (it helps if it's logical and good quality) can often make sense out of a knotty problem. If the problem isn't commonplace, support is impossible remotely, the only advice to give (after "Is the computer plugged in?") is "Maybe reinstall the application." I wonder if the cost to everyone of managing proprietary applications exceeds the cost of writing the proprietary applications. I found it humorous (or pathetic) when several days were lost at work due to a mysterious Word virus.

With binary applications, the only way to debug "problems" is via tools like strace or strings (I've seen countless instances where set-uid-non-root applications cannot read files which exist — due to permissions). They generally give useless diagnostics (of the order "Cannot read important file").

Having good quality source code is often a much more productive way to use a computer. If you regularly do something which the software doesn't properly support, having source code allows the source code to be changed. And if you can't change it, you can hire someone to do it. Commercial off-the-shelf software vendors rarely do work for hire . . . it's not in their model.

Richard Stallman has a model of selling support and service for a profit. The actual software should be free. I'm not sure I go for that, but if I buy an application, why can't I have the software for a nominal fee (let's say several times the price of the application). Whenever I ask a vendor, "How much is the source code," they're mystified. Oftentimes they quote a price several hundred times the price of the product. Other times they say, "It isn't for sale" (it's an attitude the baffles me . . . I can understand them insisting on NDA, but "not for sale?").

Also, if Linus Torvalds developed Linux with the idea of selling it, it wouldn't have become a usable system. I understand Linus is doing very well financially, and he has substantially more than 15 minutes of fame.

I'm also baffled by developers (for example, the xforms library), who make binaries freely available without providing source code. I find it hard to justify spending time with software (which invariably has bugs) without source code — so if I want to I can fix a problem or at least understand it.

OSS and Linux
From Steven Lembark <[email protected]>:

In the last issue of ;login: [April 1999], Rik Farrow wondered (mused?) whether commercial distributions of Linux might prove its bane. The example cited is HP's arcane shadow password system. The (rather accurate) description was: "Yikes!" Now for some good news: (a) passwords have nothing to do with the "Linux" kernel itself and (b) the problem can be fixed because if it's Linux then you have the source. Point (a) matters: so long as HP doesn't botch the kernel too badly everything on top of it can be fixed.

Here is an alternative outcome:

HP distributes their system, using their shadow scheme. The password setup drives people crazy. In order to make the scheme work HP hacked the fgetpwentq(3) and friends of pwd.h to handle their lookups. With all these people not liking the hacked libc distributed with HP's Linux there is a market for a "clean" version. So, someone starts with libc from GNU, makes it work with HP's Linux, compiles the utilities with it, and distributes a shadow-in-a-box-like package. Let's say it only costs $500 and they sell only 500 copies worldwide: $250,000 isn't chickenfeed.

Open-source software may be the way to keep hardware vendors honest. They can't botch the software too badly without someone else fixing it. They can't even screw up Linux too badly because the source is available to fix. If they can all agree on Linux then we might actually have hardware companies in the hardware business and software companies in the software business!

The one worm in this apple would be another "distribution" war, à la window-manager battles. So long as Linux stays open and keeps evolving, however, we have a chance to get a reasonably standard OS. And fixes.

Software Patents
From Simon J. Gerraty <[email protected]> :

I saw Cynthia Deno's article in the recent ;login: [April 1999] and just thought I'd point out that much of the software industry (myself included) thinks that the best improvement that could be made by the US Patent Office would be to abolish software patents altogether.

The issues mentioned in ;login: regarding the difficulties surrounding software patents are good arguments in favor of the above position.

Of course that would upset a lot of lawyers . . . .