Slashdot Mirror
FreeBSD at COMDEX
From: Brett Glass
Subject: FreeBSD at COMDEX
To: chat@FreeBSD.org
Just got back from COMDEX, where the response to FreeBSD was very, very different compared to previous years. Here are some random observations:
With few exceptions, all of the open source UNIX products and companies were relegated to a separate, "co-located" show: Linux Business Expo, in the Hilton. This had both good and bad effects. On the plus side, it gave open source a separate forum in which to strut its stuff (albeit with the Linux name hung on it). On the minus side, it segregated virtually all of the open source activity away from the mainstream. (Except for Linus, all of the keynote speakers for the Linux show were on a separate track and in smaller venues.) Companies which exhibited only in the Hilton didn't get as much attention as they would have on the main floor -- even if they had been crammed into one of the tiny "sheep stalls" which Microsoft uses to make ISVs seem small and insignificant. And those which had the financial wherewithal to exhibit in both places seemed unwilling to mention their open source activities on the main floor, where it was "Windows, Windows, Windows" all the way.
FreeBSD got a small, but not insignificant, amount of attention. Red Hat CEO Robert Young even mentioned it in his keynote -- a pleasant surprise.
Walnut Creek had a daemon "hostess" in the booth for the first time. ("You mean they haven't ALWAYS had one?" asked my wife, who was surprised that it hadn't been done before -- especially in Vegas. I suggested that a chorus line of female daemons -- remember the "Devil Girls" in Schmidt and Jones' classic musical "Celebration?" -- might be even more Vegas-like.)
Two fellows from the NetBSD project, including Charles Hannum, were at a booth elsewhere on the floor selling CDs. They didn't seem to be getting as much interest or recognition as they deserved, alas. The timing of the show was bad for the OpenBSD project, which is currently struggling like crazy to close a bunch of open issues so that it can ship Version 2.6. Perhaps this is why I saw no mention of OpenBSD on the show floor.
I noted that Digi was displaying some new serial hardware in the Red Hat booth, and asked them about BSD drivers. They said that they didn't have them, but "why don't you just port them from Linux?" (I tried to explain to them that the GPL, which is designed to monkey-wrench exactly such activities, precluded this; alas, they seemed not to understand the licensing issues. I plan to be in touch with them about getting "raw" technical specs, as I need a driver for a Digi 56K modem/channelized T1 board.)
The reps from Borland/Inprise -- whose booth was directly across from Walnut Creek's -- told me that they now had a Linux command-line compiler for Borland Pascal/Delphi. (This is a fantastic Pascal dialect which I'd love to use for UNIX projects. The GPLed "Free Pascal" simply can't compete in terms of code quality.) Unfortunately, despite the fact that recompiling and relinking a command-line compiler for BSD is nearly trivial, their PR people claimed that they weren't considering an implementation for FreeBSD. (This sounds like a company that's ripe for a bit of advocacy; there is NO reason why there should not be Delphi compilers for ALL of the BSDs.)
Hardware and software vendors on the main floors of COMDEX were, alas, focusing on Windows and NT. Few had driver support for any non-Microsoft operating system, and they seemed to be annoyed by the question -- as if they'd been asked quite a few times and didn't have a good answer. (Others denied ever having been asked for drivers for ANY other OS -- even Linux -- even though it's highly unlikely that this would be true.) I noted that the inkjet printer manufacturers were especially adamant about calling their printers "Windows printers," and claiming that it was impossible to run them from any other OS. Laptop vendors, when asked if their modems were "WinModems" (which I often call "lobotomodems" because they lack sufficient intelligence to work without MAJOR help from the host CPU), often couldn't provide an answer.
In general, the hardware vendors -- even more than the software vendors -- seemed to wish that all of this UNIX stuff would just disappear and leave them happily dependent upon Microsoft in a one-OS world.
The most extreme case of this of this phenomenon occurred when we wandered into the booth of a robotics vendor called Robix. We are working on a project for a client which will involve some robotics, and thought at first that this vendor's toolkit -- which contained a computer interface and enough servos and parts to build a complex manipulator -- might be just the thing. But when we inquired, we discovered that the included software, which ran the interface, was specific to -- you guessed it! -- Windows. Since "rolling your own" is the essence of robotics, we politely asked if we could obtain some sample code so we could adapt it to run under UNIX -- or, if not, the specifications for the interface so we could write something ourselves. We even offered to share the code we developed.
But instead of welcoming our interest, the owner of the company snapped in response: "We had enough trouble developing this for Windows, and we're not going to go through the sweat and tears to rewrite it for something else! Go away!" He scowled, turned his back and refused to talk to us further.
Our remark must have touched a nerve that had already been frayed by previous encounters at the show, and it was rather sad. We literally had our checkbook ready, but this one fellow was willing to throw away $500 of on-the-spot business (and that would just have been the initial order!) to avoid so much as thinking about supporting an alternative OS.
Another disturbing trend was that many of the embedded systems vendors seemed to be going with NT and failing to acknowledge its continued lack of fitness for mission critical applications. One vendor which had built a PBX around NT admitted, under duress, that to keep their system even semi-reliable they had to threaten to void the warranty if ANY other application was installed on the system. (I asked them whether they were concerned about the system blue-screening due to network activity, and told them so. The vendor seemed not to fathom the notion that NT could be crashed via a network. Duh.) Other companies had tape libraries and similar systems -- many of them likely to be mission-critical -- attached to NT boxes. Scary.
About the only exception I could find to this trend (at least on the main floor) was Maxtor. The company's MaxAttach dedicated file servers (a product line which they acquired when they bought Creative Design Solutions) have FreeBSD inside, and they're very proud of that. (They don't use Samba for SMB support; instead, they've written their own SMB server which seems fairly impressive. I didn't get all of the technical details, but their rep suggested that they may be doing some things in kernel space to increase performance.) Maxtor believes that FreeBSD will make their servers far more stable and reliable under load than Linux-based solutions such as the Cobalt RAQ.
All in all, it seems to me that FreeBSD, and BSD UNIX in general, need a LOT more promoting and a lot more vendor support -- on the main floor, not just in the Linux "ghetto." My personal approach, were I Walnut Creek, would have been to go for a booth on the main floor at the Sands and share a smaller booth with the NetBSD folks in the Linux pavilion. It's important that FreeBSD not preach only to the converted. It should not be seen as a "niche within a niche," but rather as moving toward the mainstream.
--Brett Glass
I was just wondering whether anyone knows what happened to all those. BSD no longer seems to come with them by default. And at least on the open flavor of the same, I see no /usr/ports/lang/pascal directory.
It's a shame that the Linux/BSD crowd didn't get any space on the main floor. I think separating them out as happened would do more harm than good. While having a separate area for the Unix folks does enable people to go right there when they want unix type things, it does absolutely nothing for advocacy.
Take, for example, a manager who, for some unknown reason, wants to convince the higher-ups to install only windows on their machines. This manager takes one of them to the main floor of comdex and says "Look - Only windows machines here. And all the hardware's for windows too". Even if the higher-ups know there's a bunch of unix gurus somewhere else, it's certainly going to make an impression on them that there's not a trace of open source on the main floor.
-Denor
Unfourtunately, this is just one of the many areas in which mass-cluelnessness seems to be winning over rational thought. It's very nearly on the same level on which entities like the Catholic Church operated before the 1700's. (please, try not to make this a religious thread guys! :) )
The masses didn't have knowledge of reading/writing, and because of this, many things were well beyond their understanding. Under this condition, they had no choice but to trust their authorities to tell them what was right and wrong, and out of this came a strongly dictatorial ruling. (beyond this, my knowledge of that period of history is limited... anyone have corrections? Suggestions for how the age of enlightenment could relate to us fixing misconceptions in manfacturer's heads?)
Likewise, the mainstream OS manufacturer has managed to beat the party line into _OUR_ hardware manufacturers. They seem to believe that Windows is the only system worth programming for. At the same time, start-up programmers may assume that "*nix CAN'T be easier to program for then Windows! and (quote)We had enough trouble developing this for Windows, and we're not going to go through the sweat and tears to rewrite it for something else! Go away!"(/quote)
Now, through brute force, Linux is breaking into the mainstream... but it's still not addressing the root problem. The manufacturers, for one reason or another, do not understand how easy it is to support other platforms, especially platforms with a long history of standardization. Short of us (open-source/open-hardware advocates) starting to build our own hardware, we're going to have to start seriously putting it to the manufacturers... "open specs! open source drivers! rational thought!"
"You know, Hobbes, some days even my lucky rocketship underpants don't help" -- Calvin
As in the cases mentioned in the parent article, companies make proprietary hardware with a proprietary interface that only runs with their proprietary binary which only runs on Microsoft systems. The proprietor feels that its his property, and he's done what he wants with it to recoup is investment. I understand that. But the network effect strikes again, and we're screwed.
If no vendor could create hardware with complete closed specs and Microsoft-only code, this would help a lot. But I can't see that the current MS-DOJ case could require that, since it's not a penalty to be imposed on the Evil Empire. It's all the myriad little companies nursing off their teat that are doing this to us. I don't know what kind of mechanism to fix this there could be that wouldn't be too overreaching.
Do you have any idea how expensive it is to have even a small booth on the main floor at Comdex? A booth 'next to Microsoft' would cost multiple millions of dollars. Is it worth it? I think not - the crowd at Comdex is generally pretty ignorant and thirty seconds in a 'Linux booth' on the main floor isn't going to change that. I thought the Linux pavilion was OK. I am sure the quality of attendees in that area was much higher than on the main floor. As for forcing vendors to support Linux - it's not just a technical issue but it's something that impacts an entire vendor product lineup from code to documentation, support, SKU's and pricing. The only way this will happen is if they see a large installed base of Linux out there, and those Linux users are happy spending $$$ on vendor Linux products rather than using 'free' alternatives. Complaining to people on the show floor does little or nothing, except make the experience worse for the poor schmucks who have to do it. For a start, most of the unlucky people working the show floor are either contractors or low level marketing and sales staff. Companies rarely solicit their input in product plans. You would be better off asking their tech support or emailing their product management people to support Linux as they are much more likely to have input into the product.
You don't have to like it, but the simple fact of the matter is that the GPL is a serious impediment to a lot of good, honest open source work. You can't blame the BSD people for wanting to make a free Unix. In my ever so humble opinion, the LGPL addresses most of the freeness problems in the GPL, and really should be used a lot more. However, I still suspect that even that's not free enough for BSD, and I'm not going to argue with them, because they're trying to do as much good for the world as they can. There are more axes of dissent here than meets the eye.
When I noticed it first, it was aesthetically very pleasing to see that you have managed to dynamically link those two BSD source trees.
I am not sure about the implications of this approach, yet. Made me wonder if that is the way a future united BSD source tree will look like. Not one single organization, but a tree distributed over 3 or more focal sites, that specialize in some aspect of the grand system.
If you truly believe that this alleged POSIX compliance is in the least bit useable, then please compile up trn or perl or nvi on that system. It's part of a sick and twisted joke, and the joke is on the American taxpayer, too. Read Heinz's article.
Here is my impression of the GPL/BSD license issue:
/also/ wants to be free. The goal: Give the source code to anybody. Corporations are just as much consumers as end users, and they can use the source however they want, including integrating it in a product of which they will not open the source.
/more/ restrictive, in that it isn't then giving to "everybody".
Both parties want code to be "free" (liberated), and both believe their license does that. Both believe the other license is more restrictive than theirs because it fails to be "free" in a way theirs does.
The GPL seems to advocate End User/Consumer "freedom". The goal: Get the source code to the people. GPL cannot discriminate corporations, and like anybody else they must give access to any modifications they distribute.
The BSD
However, the GPL sees that BSD does not strive for its goal because corporations may then choose to NOT redistribute that source, and that the end users may never get access to the customized corporation code. BSD sees that GPL, by not allowing corporations to incorporate and withold GPLed source, is
There are two stratas of consumers then: the end user, and the corporation. Each license tries to get their stuff to the most people, but the distribution accross these stratas are just different.
Well, that's my blurry impression of it anyway. If I'm innaccurate in some way it's not because I'm a troll...It seems to me that both licenses want to do "good", however there are differing approaches to the end result of "good".
It's 10 PM. Do you know if you're un-American?
The words which you deemed an insult were in fact reflective of sentiments that were anything but.
You then continued to write:
What's happening is that we're not talking about the same thing. Yes, you're right that the BSDL and the AL permit binary distributions. The AL goes further than the BSDL in what it says about making the source for that binary distribution available.That really wasn't the thrust behind my comment. What I was trying to express was that if you put any of these OSI-style licences on a bit of code, then that code will always have that licence. You can't just take a bit of AL'd code (or whichever licence) and throw away the old licence and re-license it as you feel like. The originally licensed code stands, and it stands forever--or at least until the owner himself releases it under an alternate licence. That means that the original code is not going to "go away" or be "taken over". You can't do that with the original code. It's got a free licence on it, and that's that. The difference is that free licences other than the GPL allow you to license your own software that uses the original stuff in any way that you care to. Even if you do so, the original remains inviolate. Nothing can happen to it.
That's what I meant when I said that the other free licences make sure that the code "stays free". Surely you must see that they do this. But you're talking about something else: code that wasn't in the original. Yes, you're right, the other licences make no claims upon that code the way the GPL does. But this hardly changes the original code.
(Yes, some of that was redundant.)
Microsoft's old strategy of "embrace and extend"--which in fact is often "embrace, extend, and extinguish"--is going to be with us no matter what we do. Look at the whole MS-HTML fiasco. This was an open standard. That didn't stop Microsoft from using it to screw the world into reliance upon them through Microsoft-only extensions. Do you really think getting a copy of their exact code to handle this crud would make any difference? I don't think it would. I don't see that licensing could make any difference here. Even if you define in the standard that extensions would make the result no longer be standard [whatever], as for example, I have heard said about XML, I can't see this stopping the Microsoft juggernaut from attempting to give you a "better" version. They'd say, well sure, MS-ML isn't XML (or whatever), it's better, and it's fully compatible with simple XML (or whatever). Think about POSIX, again.
So I think your fears about "embrace and extend" are well-founded, but your apparent conclusion that the GPL would adequately address this issue, and do so in a way that other free licenses would not, seems incorrect.
My colorful use of "viral fingers" is simply that. I grow weary of the times that I see the term "FUD" used to brand an opinion with which the antagonist disagrees. That's what I think you're doing, and I don't think it's fair. I've been completely up-front about why I consider the FSF to be dangerously dishonest in the whole GPL issue. I have repeatedly requested that they stop spinning stories and twisting common definitions. Their doing so has without a doubt tricked at least some people into misunderstanding how the GPL works, why it's there, and what its ramifications entail. I believe that this is an intentional deception, a cheap word game they will never admit to, but from which they clearly benefit. It would not take more than a few small fixes in the surrounding literature to clarify matters for honesty's sake. No licence changes. Just spin changes so as not to misrepresent what's happening.They don't do that. Therefore, they don't mind that it tricks people. Hard to see how using "sneaky" is inappropriate there.
As for "viral", this, too, has a long tale behind it, and it was hardly I who first made the observation and coined the term. I don't care that the term should discomfit the FSF. There are clearly ways in which the term is descriptive of the action. Yes, it has a negative connotation. Yes, I intended to use a term with a negative connotation. I did so, quite simply, because to my mind the term fits. I did not do so to cause people to fear something they did not understand, nor to be uncertain about the reality of the matter, nor to doubt for some nebulous and amorphous reason the intentions of the parties involved.
I think, perhaps, we are not understanding the word "infect" to mean the same thing, because I cannot see how a reasonable man could say what you have just said if his understanding of the words and the effects involved were the same as I myself hold. I note with some amusement that many GPL advocates disagree with your statement. Their standard retort is that it doesn't stop you from doing that; it merely imposes conditions upon your use of the resulting work. This is deceptive sophistry on their part. I agree with you here, and commend you for not falling into their webs of deceit.Nevertheless, this prevention that hardly seems sporting, not does it? I can barely think of a potent disincentive to code reuse. And code reuse frees programmers from needlessly reinventing the wheel.
I think either we are actually in profound but subtle agreement, or else we have severely disparate notions of what constitutes "complete and total ownership of your code". If I have complete and total ownership of my code, then I as owner may do whatsoever I please with that code. Alas, the FSF would have you believe otherwise.One can of course avoid this through linking. The FSF really hates this idea, and they routinely cluck enough about its legal viability that fair-minded people everywhere are uncertain about the true effects. But in the client-server days of RPC, CGI, DLLs, CORBA, OLE, COM, mobile agents, and other segmented forms of computation, it becomes increasingly obvious that the GPL cannot possibly be as infective/effective at reaching across those boundaries as the FSF wishes it were. That's how I made removed the virus from all GPL'd libraries and made them LGPL'd. Even Bruce Perens has confessed here in this forum that such separations are going to happen in the computing models we see today, and that the GPL does not address them. My recollection is that he called them "loopholes".
I agree with you that my code code is not affected. The FSF, however, most vociferously disagrees with both of us. They claim that the GPL means that I cannot distribute my own code under my own terms. If I can't do as I will with my own code, then it is hardly unaffected, nor am I its complete and total owner. That's there position. Mine is that I am complete and total owner, and am consequently free to do whatever I please with that which is mine. It is, well, peculiar at best, that copyright law should extend its notion of a derived work in this way. The FSF recognizes no notion of proportion in their figmentational notions of what constitutes a derived work, staying completely boolean in their thinking. The courts have never been so binary.In any event, that's not what the issue is. The issue is that the FSF feels that their code affects my code, that their licence on their code spreads to everything their code touches, meaning my code, and that this process continues in perpetuity, without regard to dilution or proportionality. It is this very complete silliness that they espouse which has engendered the notion of the GPL being an infectious virus. If you prefer another way of thinking about how silly it is, consider it as an application of digital homeopathy.
Surely you don't really believe that, do you? Free licences will do that, but not the GPL. The GPL affects other code in a way that a free licence does not. This is my entire point. Yes, it forces. But it forces something else. It forces what happens to something other than the original code. It allows the author of one work to restrict what happens to the work of an entirely different author.Can you imagine how silly it would be if a book were published whose copyright included a restriction that the book could not be used by black people, or in a public library, or placed on the same shelf as a book by a different author? Imagine if a song were published under a restriction that it could not be played a station that also played a song by a competing musician, and that any other songs played by that station fell under the same restriction as the first song? Really, it's completely silly.
I've always found "use" restrictions very strange. See my previous paragraph. And I find it nothing short of mendacious that the FSF should claim their restrictions are anything but that. Yours is a very common sentiment, and I can certainly respect your feelings on this matter. A lot of people feel this way. That's why you so many licensing terms that in effect grant unlimited non-commercial use, but that for commercial use, you must contact the author to make other arrangements. I can hardly fault them too strongly for this, because I do hear what they're feeling. They've done their work for free, and they don't want people to get some benefit out their work which they themselves are not getting.According to the FSF, this is not free source, open software, or anything else you care to call it, because it's got anti-commercial restrictions. And then they tell you that their restrictions which bar anyone from using their software in what any businessman would be call a commercial sense (traditional fee-for-licence schemes) is not anti-commercial. Who but George Orwell could be so proud of the boldness of this spin job?
It would be more honest of the FSF to get out of the business of word games and related spin. But they are, fundamentally, a politico-economic foundation, not a technical one. They wish that all software were GPL'd, because they could thereby impose their morality upon others by the--what word do you want me to say here other than the completely honest "infective nature" or "viral nature"--perhaps "collateral damages", then--of the GPL.
I must confess that I have on a few occasions in the past, and doubtless several in the future, contemplated places where I would dearly delight in seeing the GPL installed and enforced. Oh, you do not know how sorely tempted I have been! One example is with Microsoft's operating systems products, because if they were court-ordered to slap the GPL on their OSes, that would be a likely end to their strategy of putatively "integrating" into the OS any application area that they care to monopolize.
But you know what? This is a personal weakness of mine, and I must overcome the urge. Tempting though it may sound, I must reject the temptation. I must. That's because it is fundamentally immoral to coerce others to behave in accord with your own sense of morality. It doesn't matter whether your morality happens to be the best around, or even the best there can ever be. Coercion is by its very nature by definition immoral. As with the paradox of having your cake and eating it too, morally you simply cannot enforce your own morality on others without sacrificing that very moral high ground which you would claim to occupy. Without free will, there can be no morality at all.