Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. ×
GNU is Not Unix

GNU/Hurd Web Server Online 361

Ross Vandegrift writes "Jeff Baily sent an email to the Hurd development list today announcing he has put up a GNU/Hurd webserver. Not much content there, but the fact that it is up is incredible alone. Keep up the good work Hurd! "
This discussion has been archived. No new comments can be posted.

GNU/Hurd Web Server Online

Comments Filter:
  • Don't diss message passing and microkernels until you've tried it. Sure NT sucks, but it would suck even it it was monolithic. (ie. 9x sucks, but has no relation to Linux, even though they are both monolithic, no?) Try programming for BeOS. Messaging is blazingly fast, and the microkernel design speeds up the OS instead of slowing it down.
  • Linux as an OS certainly won't last forever, but in the long run Linus just might be remembered not for writing an OS but for creating a whole new kind of development process, one that isn't going away.

    The whole idea of 'release early, release often', invite patches from everybody, and huge-team development was actually pretty different from the way even gnu worked at the time.

    This is essentially a myth. In The Cathedral and the Bazaar [tuxedo.org], ESR argues the merits of a "bazaar" model by contrasting it with a "cathedral" model of closed software development. The examples of "bazaar" development he presents are Linux and fetchmail [tuxedo.org] to represent the "bazaar," and his example of a "cathedral" project is.... Emacs.

    How's that again? Emacs is not an open source project? It does not invite patches from everybody? It does not incorporate contributions from an army of individual hackers? It has not made all its bugs shallow by offering its source code to millions of eyeballs?

    The plain weirdness of this comparison still leaves me puzzled. What is it supposed to mean, in a paper whose thesis is the fundamental superiority of open source over closed source? That Emacs is essentially a closed-source project? That it has more in common with NT than it does with vi? These notions are absurd, but it is hard to draw a different conclusion.

    The truth is that these development models are quantitatively different, but not qualitatively different. While Linux development is more frenetic than that of FSF mainstays like Emacs or GCC, nightly snapshots and frequent releases are only modest differences in style. They are essentially personality differences; even GNU and BSD projects include nightly snapshots, after all. They don't constitute a sea change in software design.

    While no one invented the open source software ethic, it appears likely that Richard Stallman will get more credit than any other individual, which is IMHO as it should be. Many, many people created projects, but Stallman created the movement. More people became conscious of free software and open source as a philosophy through Project GNU than through any other source, Linux included. Linus didn't create the development process; Stallman did.

  • You seem to imply that microkernel scales better than a monolithic kernel. Yet Solaris, which scales quite nicely up to 64 CPUs, has a monolithic kernel with loadable modules, just like Linux.
  • by kijiki ( 16916 ) on Sunday December 05, 1999 @01:34AM (#1479187) Homepage
    When slashdot descends,
    Experimental box is
    A smoking crater.
  • Rogers @Home isn't upload-limited everywhere.. everybody living around me who has cable has no limitations on upload whatsoever.. Could be because I'm in Canada, but still...
  • The default config if you have a lancity on Rogers is usually as such. I know people in BC and ON who both have the exact same 47KB/s uplink limit -- but yes, service does differ as per locality.
  • Interesting! The FAQ claims 1 million hits or so, so maybe it's out of date. Boosting traffic by 3.5 million hits, my.

    Still doesn't need a quad OC-12, though. Guaranteed.
  • @Home does use DHCP, but they always serve you the same IP. The only reason they use DHCP is that *in case* they have to change your IP (major network reconfiguration -not happened yet) they can do so painlessly.

    I have a dual boot machine (Linux and Win98). I am not using DHCP for Linux, it's all hard coded. Works just fine :)
  • Linux won't really die because there will always be legacy code. Perhaps not as restrictive as 16-bit DOS, but certainly portions of source code will be ported to any new kind of system. It may not be 100% Linux, but I'm sure that we'll see parts of Linux for years after it's "declared dead."
  • Here in BC, (and probably elsewhere too), Rogers has a clause in the @Home license that really makes me uneasy.

    I'm just pulling figures out of my a**, but after the first gigabyte of transfer per month they may charge you with $1.00 CDN/100 kBytes. I've never heard of it being enforced, but I know they could send me a bill for several thousand dollars on a fairly quiet month. Ulp!

    (If anyone can verify this and come up with some proper figures, I'd be much obliged. I can't find my bill anywhere. Secondly, this could probably be thrown out since it hasn't been enforced, but IANAL.)
  • Erm, a point about version numbering:
    Under standard OS releases, version 1.0 is a fully stable version which has been rigorously tested and for which there are no known issues (some projects have chosen to violate this definition, for one reason or another). If this was a 1.0 release, then it probably would have been under heavy testing for some time. Right now HURD is still in heavy developement, and will continue to be for some time to come.

    As another poster pointed out, Linux pre-1.0 was very stable and reliable.
  • You don't understand. A server in this case is not a stand alone program (inetd). It is what you probably think of as a module (but is much more powerful).

    The apps server is (i think) the part of the OS that handles any communication between any user space programs and the traditional OS functions.
  • HURD is not "much like the Linux kernel." It is a microkernel; Linux is a monolithic kernel. How much more different could you get? You could make it closed source, put it in a big box and let people pay to use it. I'd rather prefer it being at least a bit like Linux in that way... =:-) CU, Ventilator
  • most of the time its at least intelligible, if not intellgent, what is this supposed to mean?
  • Well, I dunno. I'm not entirely sure of what HURD exactly is, just that its another OS based on GNU, but it sounds kinda sifty. If you don't feel like running it full time yet, well, maybe the FreeMWare came just in time. Oh, if anyone knows a site / feels like explaining what HURD is that'd rule.

    If you think you know what the hell is going on you're probably full of shit.
  • Yup, it couldn't handle the stampede of slashdotters (sorry couldn't resist)
  • by Zurk ( 37028 ) <zurktech AT gmail DOT com> on Saturday December 04, 1999 @06:29PM (#1479213) Journal
    here's the dump from netcrafts query page :
    Sorry, couldn't determine what the server was for host hurd.zugzug.com on port 80.
    interesting to see that netcraft couldnt determine the server.
  • Whatever the "next Linux" may be, it'll have to be damn good to make me switch over :)
  • If you don't already know this, propietary software isn't the same as commercial software. And ask how the people at Red Hat feed their children.

    You are seeing software as a product to be sold. I see software as knowledge to be learned and experimented with. Yet scientists all over the world are being hired and are feeding thier children. At no point does a scientist say, "Let me propietize my discoveries and forbid anyone sharinng this information."
  • There is nothing insulting about saying GNU/Linux. I have looked at all the GNU software I use and it makes sense to call it GNU/Linux. But if people want to call the system Linux, go ahead. I just fear people will forget about GNU and freedom if we all called it just Linux.

    Please don't speak of GNU, GNU/Linux, or Debian in a derogatory manner.

    Otherwise, I think we have reached an impasse.
  • I fyou read the debates between Andrew Tanenbaum and Linus Torvalds, when Linus was developing Linux with the community, you will see that Linux does not have a microkernel. A microkernel seperates the kernel level services, system calls, interrupt handlers, etc... from the managment systems, like memory managers, file system managers, and drivers a like. To say anything with Linux and Microkernels in the same sentence, one should always include "does not have", or "is easier to implement", etc... There is a big debate because the Linux kernel is so huge because of all the modules that need to be loaded. A microkernel system would be much leaner, I do not know about faster. My time with Minux was short ( a semester ) and my performance tests comparing Minux and Linux would not be fair
  • by Tom Christiansen ( 54829 ) <tchrist@perl.com> on Sunday December 05, 1999 @04:59AM (#1479237) Homepage
    Superficially you'd think there's an element of jealousy there because Stallman almost wrote an OS then someone else got most of the limelight by writing the kernel... but that just doesn't feel like the case
    I realize that emacs is a fine operating system, but some of us still prefer Unix.
  • Yeah. Actually that stands for LSD:

    Lucy in the Sky with Diamonds


    Actually, someone misheard it as "Lucy's getting high with Linus":

    The embarrassing moment of revelation:

    My was it an embarrassing moment. I
    was giving a private concert for Paul
    McCartney and Ringo Starr, and they
    started laughing when I sang the wrong
    lyrics.

    Age when I realized how wrong I had
    been: 23
    -----------------------------

    Me used to be a angry young man
    Me hiding me head in the sand
    You gave me the word
    I finally HURD
    I'm doing the best that I can.

  • Your company can sell all the software you want. I just ask it to be free (as in free speech). If you don't, I will buy my software elsewhere.

    Also you assume (college == get a good job).

    If you can make a living fliping hamburgers than why not? My plan is to get my degree and do exactly that. Or perhaps some other job. But I will still have my free time and perhaps I will do something I love other than making a superfluous amount of money.
  • If you don't like don't use it.
    They have nothing against you.
    You have no God given right to FSF code or any other code, just as the FSF has no God given right to your code.
  • neither are necessary for an OS. GNU tools (compiler, basic shell utils etc etc) are more than sufficient.
  • So I should write my code so you can use it in your proprietary products?
    Ha ha, I don't think so... If that is the case I want to get paid for it.
    See the problem? The GPL is largely responsible for the success of Linux.
  • no. eros is a rather kewl new OS being developed. refer to previous /. stories or search on google.
  • There are at least two major innovations of the FSF. First is a multi-language cross platform compiler. I know of no other that exists. The second is the copyleft and the GPL.

    Calling a system GNU/Linux means that it is the GNU system based on a Linux kernal. I see no insult to that. What greater insult it there in calling a system GNU/Linux than just Linux? I would say it is just the opposite.

    Can I ask to to back up some of your wild assumtions?
  • This is one the reasons free software/open source is so much more like knowledge. Didn't Isaac say that he could reach new heights by standing on the shoulders of giants... Let me see that happen with propietary software :)

    plug:
    I wrote about this very thing in The Rise and Fall of OS Empires [osopinion.com]. It was also a subtle argument against BeOS :)
    /plug
  • Hey! That [slashdot.org]'s interesting. I had not heard of Exokernels before. Anything that increases web server performance by half an order of magnitued is a Good Thing.

    Read more about XOK and ExOS at http://www.pdos.lcs.mit.edu/exo/ [mit.edu] as the man said [slashdot.org].

  • >It seems the advantages of HURD over Linux or FreeBSD, such as they are, are expandability and scalability rather than functionality. I wonder how attractive HURD will be, however, once Linux and FreeBSD have completely multi-threaded, modularized kernels.....

    >Moreover, granted that the FSF goal is to get everyone using their software, I wonder if HURD has anything to convince people, and most importantly industry, to switch.

    >It also seems as if it will take a year or two for HURD to be stable; in a year or two, FreeBSD and Linux are going to be so far ahead in hardware support, stability, and in functionality that HURD would seem to be useless.

    >In sum, HURD seems to be a failed attempt of the FSF to totally conquer the Open Source world, an attempt doomed from the start. In truth, they are about five-ten years too late.

    Actually, you're missing the whole point to Hurd. Hurd is not in competition with Linux, any more than say, OpenBSD is in competition with Linux (which it is not, incidentally).

    Linux is a general purpose kernel based on familiar and well understood technology that tries to be as many things as it can be without compromising any of its existing functionality. It is a monolithic kernel with a lot of modularity built in, but still old technology.

    Hurd is something of a research project aimed at the Next Generation Kernel (or Next Generation Microkernel). Trying to create a viable operating system with as much modularity as possible. Once you have a stable microkernel, you can leave the system up and upgrade vast segments of the operating system that would force even Linux to reboot.

    It is not aimed at displacing Linux any more than Linux is aimed at displacing Windows. Some might say that is Linux's purpose, but others have stated that Linux is simply trying to be the best operating system it can be, with World Domination as a secondary goal. Hurd is trying to be the best kernel it can be.

    And here's an interesting point. There is nothing to say that the two groups can't use each other's code, because they're both under GPL. If the Hurd group is intelligent, they'll try to take advantage of the code base of Linux and not reinvent the wheel implementing functionality already present in Linux, especially for things like hardware support.

    For that matter, the Linux people might decide after a certain point that they can't modularize anymore without going to a microkernel based setup. At which point they join the Hurd project, fork off the Hurd code base, or start from scratch using information gathered from Hurd's work. In any such case Hurd will have fufilled its purpose.

    In short, I think Hurd is an important part of the future of Linux, as part of determining where Linux will go in the future.
  • No, I fear you're misunderstanding a few things. Having had the pleasure of hearing both Richard M. Stallman and Eric S. Raymond (both at my university in the middle of nowhere-land, go figure), I feel somewhat competent to reply to this.

    The first thing to understand, in my opinion, is that the Free Software Foundation, which RMS started have never really claimed any inherent superiority of Open Source programs. RMS's beef is with the philosophy. Free Software is morally superior to proprietary software. To put it bluntly, RMS would use free software (as always, note the ideological schism between Free Software and Open Source. Free Software comes with an ethical theory (one I personally think has a lot going for it), whereas Open Source is far more pragmatic.)

    Now, where was I? Oh, yes. Even if RMS broke with the proprietary model of software development, he kept the model of software development normally found in computer science. And I don't really see that he had any reason not to. The first version of UNIX was basically made by two people (Thompson and Ritchie), the C programming language was developed initially by two (K&R). Whereas the predecessor to UNIX, Multics, a large project with heaps of developers failed miserably and never produced a working system. Not that I'm saying that these had any effect on RMS and the FSF, actually RMS said he had several hesitations about the UNIX system, but in the lack of an alternative he made due with it (the model the FSF strove for was to replace each component of the UNIX system one by one. But note the underlying assumption, that once a component was made, it was basically done, though I should give a disclaimer that that's my interpretation of it, RMS didn't say that explicitly).

    So, now we get to Emacs. You said:

    Emacs is not an open source project? It does not invite patches from everybody? It does not incorporate contributions from an army of individual hackers? It has not made all its bugs shallow by offering its source code to millions of eyeballs?

    To which the answer would be, not really, no.

    Emacs was an editor made by RMS before he founded the FSF. (And in my opinion it's only fair to call it a Free Software project rather than an Open Source.) As to its developments after it was licensed under the GPL, make no mistake, it was very much a cathedral style project. It's developers were an elitist group, and most certainly did not accept patches from anyone. Nobody really thought like that in those days. ERS too said that when he heard about Linux, his first thought was that an OS hacked by several hundred developers (or whoever many it was back in those days) might be fun to play with and look at, but it could not possibly be any good. It would have to be an incredible patchwork which chief success was if it worked at all. To his surprise it not only worked, it was " good!

    So in summation. RMS founded the Free Software theory of ethics (which he should have wholly credit for), but the development process ESR describes as the Bazaar style is new and something completely different from what the GNU project ever envisioned. Linus didn't invent this, I know he said once(I wish I remembered where so I could give the source like a good journalist, but I'm afraid I'll have to leave it at: some interview I read somewhere) that he was surprised when people started turning in patches, his only ambition he'd only expected people to say good or bad. The development style came from nowhere, but make no mistake, it was new.

    Eythain

  • They will use GNU for their operating system.

    Alejo.
  • DOS and Linux were very different, while Hurd and Linux are very similar.

    I think you need to learn a little more about the Hurd before making a statement like that. Start by reading some of the other posts on this thread ;-)

    Daniel
    1. Customers demand free software
    2. Companies meet demand
    3. Developers working for companies develop free software


    That's the way I think it will go.

    I may end up developing propietary software. But that will not prevent me from writing free software. It will also not prevent me from demanding free software. Many people must develop propietary software. There is nothing wrong with that. The revolution has only begun.
  • Linux isn't innovative
    Its a reimplementation of UNIX.
    So whats your point? Whats so important about innovation anyway?
    I'll use whatever is available NOW (if it suits my needs) whether its innovative or not.
    Gcc, gdb and other Gnu software and especially the GPL, where central in making Linux happen. Whether they were "innovative" or not (the GPL certainly was) is not the complete point.
  • The only FSFware I ever use is the compiler.

    That is hard to beleive. GNU is so pervasive in "Linux" operating system that it makes more sense to call it GNU/Linux. Besides, Linux is a kernal, the rest is mostly GNU.

    Take a look at http://www.gnu.org/software sometime. I counted well over a hundred commonly used software packages there sometime in just the GNU section. Including GNU Bash, GNU Window Maker, GNU Emacs, GNU elvis, GNOME, glibc, etc. Are you sure you don't use any of these or other software on the GNU software map?

    GNU is a complete system except for the kernal. That doesn't mean added software makes it not GNU.
  • OK, fine. FSF sucks, GPL sucks harder.
    So why don't you open up Visual C++ on your NT box or some other compilers on the platform of your choice and use them instead?
    The software is only meant to be used by those who find the licensing acceptable.

  • *But* Linux still has many more years of development behind it than HURD.


    Actually, that's not quite true. In fact, I think that the Hurd has been around longer than Linux. The problem as I understand it is that it's an experimental design, and they keep deciding to throw out or rewrite parts of the system, whereas Linux is based on the well-known Unix kernels. Recently there seems to be a significant amount of forward progress, and I wouldn't be surprised if things get a lot better over the next year. Or they might not ;-) Time will tell..

    Daniel
  • What is the point you are making? Notice that propietary programming languages aren't too successful either.
  • You might both be right, depending on what you're each thinking.

    Any POSIX conformant system will have essentially the same API. By that way of thinking, OpenMVS, OpenVMS, Solaris, AIX, Irix, Tru64 Unix, BSD, Linux, and yes, even the crippled NT subsystem [perl.com] seems quite similar. If the Hurd attempts to conform to the POSIX spec, then it would seem very similar to Linux or Solaris or BSD, which also make the same attempt. write(2) and read(2) et alia are pretty much the same everywhere. MS-DOS, on the other hand, seems quite different, because it is not standards conforming.

    Of course, to someone hacking on the kernel source code, you, Daniel, are completely correct. A regular Unix kernel hacker will look for a long time for bdevsw in something running on top of a microkernel and never find it.

    It's really a matter of perspective.

  • He's on rogers, the have static ip's more or less. His uplink is also likely 47KB/s.
  • As opposed to propietary software?
  • You are stealing the work of programmers, who freely gave you access to their code.
    All of these PROGRAMMERS, coders just like you, wished you to follow the terms of the GPL.
    Specifically that you only use the software in the way that they indicated and that you make the source code freely available.
    If you couldn't abide by these terms, why didn't you use something else?
    How would you feel if I treated your work the same way?
  • Emacs was an editor made by RMS before he founded the FSF. (And in my opinion it's only fair to call it a Free Software project rather than an Open Source.) As to its developments after it was licensed under the GPL, make no mistake, it was very much a cathedral style project. It's developers were an elitist group, and most certainly did not accept patches from anyone.

    I'm amazed at how far the disinformation campaign has come.

    I am looking at the documentation for Emacs 20.3. The Acknowledgements chapter thanks over 200 people by name for their contributions to Emacs over the years. I am quite sure that those people were not all principal maintainers of Emacs. Most of them are credited with contributing only one feature, which indicates that they were not part of some rarified Emacs Clique but merely randoms who had good ideas.

    Every piece of evidence I can find indicates that RMS recognized the technical value of free software from the very beginning. From the GNU Manifesto: [mit.edu]

    Once GNU is written.... much wasteful duplication of system programming effort will be avoided.
    I would like to see GNU development supported by gifts from many manufacturers and users, reducing the cost to each.

    From Stallman's interview in BYTE [mit.edu] in 1986:

    Other people may use the GNU system simply because it is technically superior.
    Even just the free support that consists of my fixing bugs people report to me and incorporating that in the next release has given people a good level of support.

    I find it mystifying that Linus is getting credited with inventing these concepts.

    [Linus] was surprised when people started turning in patches, his only ambition he'd only expected people to say good or bad. The development style came from nowhere, but make no mistake, it was new.

    It was not new. Linus knew perfectly well in 1991 that people who liked the software they were using would contribute bug fixes and improvements. This pattern was, indeed, commonplace even then. What surprised him, as he's said before, was the enthusiasm with which people took to his project, not the mere fact that someone would choose to submit a patch.

  • Well, that's hardly surprising... It's probably the only HURD box on the net.

    --

  • Point taken. But no one can reuse the code of either.
  • heres what I got out of it:

    :/= That's a crooked smile guy with a goatee
    1|? Well, in binary, you can have a 1 or a what? A zero of course, so we have this so far:

    :/=0 Which appears to be a fat man with a crooked smile and a goatee.

    57| Well this is, ah, fuck it, i don't know, i'm just trying to get my karma up from -4 by posting stuff, it's not working
  • In sum, HURD seems to be a failed attempt of the FSF to totally conquer the Open Source world, an attempt doomed from the start. In truth, they are about five-ten years too late.

    I doubt the FSF are trying to conquer. In fact they *invented* open source/free software after everything turned propietary. And every Linux-based system uses GNU so they in fact have conquered the open source world as much as Linux has. So I don't think the FSF need to conquer the "open source world". They already have.

    Hurd will hopefully provide a worthy successor to the great kernal Linux is.
  • This [gnu.org] is the official GNU Hurd site. Short summary: It's a microkernel, which means it's harder to develop and possibly slower, but more versatile in what it can do. Think a kernel that's nothing but modules.
  • Since he's on rogers, his uplink is more or less fixed.. Doesn't matter anyway though, far more traffic goes by which could be classified under quake/half-life/whatever servers and warez.. :)
  • Naw, traffic is managed through QOS, most likely using ONAdvantage. Therefore one user can't use all the traffic on a segment himself -- unless of course, no one else is trying to do anything.

    Most rogers users (his MSO) have lancity modems and they have a fixed uplink of around 47KB/s (well, backbone and local problems aside).

  • It's nice to see another OS (kernel) design is still being worked on - people seem to get so caught up in the Linux/Windows war that they forget that there are several other designs that offer benefits of their own.

    I'll certainly give a full HURD distro (say, Debian HURD) a good try should it ever make near-completion. :)
  • I was very careful to say "after everything went propietary". They invented the term "free software".

    Apache, Perl, Python, Sendmail, Tcl, and XFREE86 are just a few of the important projects that are unrelated to the FSF.

    These are free software projects. Some of these sprang from the free software movement. Some of these sprang from the open source movement which came from the free software movement. Some of these just wanted a "you do whatever you want with this just keep my name on it" license.

    Okay. I guess saying that the FSF invented free software is somewhat less than correct but it is close. I am sorry. No need to be hostile.
  • 1 276 ms 28 ms 27 ms aggro-80.lvcm.com [24.234.80.1]
    2 25 ms 30 ms 22 ms edge.lvcm.com [24.234.0.1]
    3 217 ms 70 ms 22 ms H4-0-0.irv-lvg100.gw.eni.net [207.168.88.1]
    4 40 ms 42 ms 42 ms noname.eni [155.229.120.157]
    5 239 ms 29 ms 49 ms bb3.mae-w.home.net [198.32.200.47]
    6 42 ms 42 ms 356 ms c1-pos3-3.snfcca1.home.net [24.7.66.45]
    7 332 ms 57 ms 55 ms c1-pos4-0.sttlwa1.home.net [24.7.66.1]
    8 54 ms 62 ms 317 ms bb1-pos2-0.rdc1.bc.home.net [24.7.72.174]
    9 265 ms 66 ms 65 ms 10.0.186.114
    10 * * * Request timed out.
    11 * * * Request timed out.
  • True, every Linux-based operating system uses some of the FSF's software; viz. the compiler. But some uses far less than you allege. Consider David Parsons's work, for example.

    I will not claim that all Linux based systems use GNU but the vast majority do. All major distributions use vast amounts of GNU software. I am certain that your main system use a lot of GNU software.

    Linux is hardly about the FSF, you know.

    This depends on what you think Linux is about. The FSF is about free software (open source). I would say they are about similar things.
  • But I do give the FSF the credit for the Free Software and Open Source movements as we know them. If it wasn't for the FSF, none of this would be possible.
  • by Daniel ( 1678 ) <`gro.naibed' `ta' `sworrubd'> on Saturday December 04, 1999 @06:53PM (#1479377)
    Periodically, someone comments that warnings should be sent to admins of systems whose systems are linked to from here, just in case the load is too much. But linking to a system running an experimental operating system should probably fall somewhere under premeditated assault.. ;-)

    Daniel
  • Your point about speed is completely specious, I'm afraid.

    The ability of the PentiumII/III architecture to parallelize certain tasks (via having multiple execution units; the Pentium could do this in a limited way, and all modern chips such as SPARC or Alpha or PPC also do this) does not mean your processor can run multiple tasks at once.

    Let me quantify that a bit better: each process (and processes and threads are very close to the same on all Unices, but running NT won't help you in this case) requires its own complete context, consisting of register values on the processor and page tables maintained by the OS. Since the processor only has one set of registers and only one program counter, it can only be executing one thread of instructions at a time. It can parallelize only in a very small local area around the PC. This speeds up execution of the current instruction thread, since you can get other things done (i.e. evaluate other non-dependent instructions) while waiting for a particularly slow instruction to finish. But it doesn't speed up multiple threads.

    In fact, context switches (required to switch threads) are very expensive. You have to save out all the register values, load new ones, and bring up new page tables. In addition, unless the two threads happen to be working on the same chunk of memory, all the cache data will have to be thrown out (a significant slowdown).

    I'm not arguing that Be is not fast, I know it is. And I'm sure it's possible to take excellent advantage of SMP with microkernels. But there's no way to do single-processor threading on your typical processor (of course you are free to build yourself a processor capable of whatever you want). Period. Sorry.
  • Wow. You people are quick with the submit button :)

    No, my main system is a Sparc running OpenBSD. Scant little FSF software there. Just the compiler. The rest is free.

    Very well.

    More importantly, you have a logical fallacy. Just because the Linux kernel is free software, and just because the FSF is about free software, does not mean that Linux is about the FSF, nor vice versa. Aristotle would not be pleased. :-(

    I didn't mean to make a logical implication or any other implication. But if you want a syllogism, then very well.


    FSF is about free software.
    Linux is free software.
    Therefore, FSF is about Linux.


    Is that a fallacy? I hope not. I don't think so. Given that Linux is a subset of free software I think that it is valid. But I will not say and didn't mean to imply that Linux is about the FSF. But I repeat my question: What is Linux about? Linux is a kernal, it can't be about anything. So your fallacy is in implying that it could. :)


  • I wouldn't say that. The BSD system is hardly viable for much more than serving by itself AFAIK. One of the FSF's most significant contributions was probably the GPL. Not to mention that gcc and bash is helpful for BSD as well.
  • It will be a good day when people can choose a kernel, independent of libraries, that suits their need, and recompile it to match their demands. GNU/Hurd is another step closer to this. I can't wait to see Debian's FreeBSD port.

    -----

  • I'll answer a few of the questions raised by various posts on this discussion. (I subscribe to the debian-hurd, help-hurd and bug-hurd mailing lists, and I've tried the Hurd on several occasions, so I more or less know what I'm talking about.)

    Of course, the TCP stack crashed minutes after the site being slashdotted. This isn't really an issue: Hurd is still highly experimental, and its TCP stack (which is merely a copy of the routines from Linux - but done in some haste, I think) is mostly used to make it possible to use the box remotely (the Mach console is a pain for one thing, and X requires some patches to work on Hurd). The interesting thing, however, is that whereas the translator handling the TCP stack crashed (the ``pfinet'' translator), the system didn't. That is, in fact, the whole point about the microkernel architecture.

    The last distribution of the complete GNU system was 0.2 and it is now completely obsolete. The next distribution, Debian GNU/Hurd 0.3 potato, should come out together with the corresponding GNU/Linux distribution, and share some packages with it (the non-binary packages; binary compatibility between Hurd and Linux is a goal for the future and shouldn't be too hard to achieve, but it's not there yet).

    The system now works quite well, and is able to run nearly everything, but it's still far from stable, and miles from being optimized. Filesystem demons are the most important thing to finish, and they are now almost completely stable. More advanced translators like the nfs clients or the ftpfs (allows you to mount ftp directories) are there more to show the power of the translator paradigm than as actual working systems, and they're quite unstable. But, once again, the whole point is that if a filesystem (other than your root filesystem) crashes, the system will typically continue to function correctly anyway.

    The Hurd shares the same libc with Linux, so porting from Linux to Hurd is typically trivial. The major source or problems is that some programs make wrong assumptions about system limits, that are not true on the Hurd. For example OPEN_MAX is 256 on Linux, and is 2So the Hurd certainly won't be ready before a couple more years. But you shouldn't conclude that it never will ``catch up'' with Linux, either. For one thing, most changes made to the hardware drivers of Linux are incorporated verbatim in the GNU-Mach microkernel, so the Hurd team doesn't have to worry (excessively) about all that. Adding filesystems to the Hurd is much easier than on Linux, and debugging them even more so, so there it's also not too much of a worry that the Hurd development team is so small. The problem of the TCP stack remains, and while it should be possible to take some parts from Linux, it will probably be a long time before the Hurd has the same networking capabilities as Linux...

  • By posting this, I shall undoubtedly be subjected to the wrath of the many vocal zealots who make up a majority of the HURD community
    I've never been a zealot for anything, but I understand it's a requirement to being a Real Geek (tm). Where do I sign up to be a HURD zealot? I know next to nothing about HURD, therefore I have all the qualifications!
  • by legoboy ( 39651 ) on Saturday December 04, 1999 @07:03PM (#1479400)
    I can just imagine him surfing the net, reading Slashdot.. He sees a story about him appear at the top of the main page. He has two seconds to say "Oh NO!".

    Suddenly, his net connection is hosed for the next week.

    ------
  • So, the FSF is about all open source software (which you equate to free software)? Is that really true?

    Yes. Absolutely. Freedom to users is the entire reason for the FSF of being.

    There's a lot of free software out there that the FSF doesn't seem to care to have anything to do with. Certainly there's a lot of free software that would very strongly prefer that the FSF have nothing to do with them.

    You are confusing software with developers. The FSF's goal is for their to be more free software and more users of free software. Strategicly, projects that encourage less software or more users of propietary software is not in free software's best interest and the FSF is right for encouraging otherwise.

    BTW: Software doesn't prefer anything :)
  • by Hawke ( 1719 ) <kilpatds@oppositelock.org> on Saturday December 04, 1999 @07:11PM (#1479411) Homepage Journal
    That's a great idea. Don't know how to stress test the TCP stack on your brand new OS?


    Get it mentioned on slashdot.


    Course, that stress-tests your ISP as well. May not be too usefull.

  • by Anonymous Coward
    Everyone seems to be ticked that this server is going slow... I think that's fair. It's still in beta (I'm pretty sure), and even it weren't, it would be a very low version, somewhere around 1.0. Show me a box running linux kernel 1.0 that can withstand the slashdot effect. Hmpf. Aaron
  • That means BSD is good at being useful. GNU/Linux is good at giving users freedom.

    If BSD had all the hype right now instead of GNU/Linux, wouldn't you think that Corel would be propietizing the entire operating system right now? Making incompatible changes for their software to work on it but no other version of BSD to give them a competitive advantage.

    If I at some future point and time use a BSD-derived system, I have no guarentee that I will have source code, freedom to modify, or freedom to redistribute. With GNU and Linux, I have that freedom.

    Copycenter is an interesting concept. Could be used for textbook examples. BTW, what is the difference between the BSD license and public domain?

  • The HURD is a kernel, much like the Linux Kernel.

    When the GNU project was started, it was intended that HURD be the kernel used with the GNU utils to form a UNIX-like operating system.

    However, they left HURD to last (seemingly) and before they could get it finished enough to be useable, the Linux Kernel arrived and was combined with the GNU utils to form "Linux" as we know it today.


  • How's that again? Emacs is not an open source project? It does not invite patches from everybody? It does not incorporate contributions from an army of individual hackers? It has not made all its bugs shallow by offering its source code to millions of eyeballs?


    Emacs development is very closed compared to e.g. the Linux kernel. The pre-tester lists are invitation only, and doesn't contain discussion. Test releases are rare, and placed in execute-only directories on secret ftp sites. There are, to my knowledge, no real developer lists, instead RMS send private mail to people whose help or input he want on particular issues. There are no access to the development code outside an even smaller group than the pre-testers.


    Contrast this with XEmacs development, where the development discussion is public, there are frequent development releases, and anonymous CVS access to the latest sources.


    ESR has been involved with Emacs development, and his characterization is quite on the mark. The Emacs distribution itself is a one man project, however, a community (or bazaar) exsists around the various Emacs Lisp packages.

  • To the user, HURD looks a lot like Un*x. It's a microkernel operating system, which means that there is a small kernel (relatively) who's primary duty is to help various daemons communicate with each other.

    These daemons do all work, and make the system look like a POSIXesque system. The daemons are (theoretically) easier to mix and match then a big monolithic kernel's functions. They also could potentially perform better in an SMP environment (since there's lots of seperate processes).

    The other big thing in the HURD is file translators -- these are programs that are run when a file is opened, read, written to, whatever. So, for instance, you could have a file translator that creates virtual directories or makes a transparent ftp connection (so that, for instance, the file /ftp/ftp.cdrom.com/pub/README would transparently retrieve a file).

    You can read much more at the HURD's website [gnu.org].

  • Larry Wall created patch as free software. After he lost interest in it, the FSF took over maintenance.
  • Some persons are bashing Hurd for not standing up to being Slashdotted.

    I'd like to remind everyone that resisting being Slashdotted does not require a good processor or a good operating system or a lot of memory (that is, unless you are building pages dynamically in a wrong (eg. cgi) way), just a good network connection.

    This box could run any operating system on a fast processor with a lot of memory and the results would be the same if it had the same bandwitdh.

    So don't bash Hurd merely because this box was Slashdotted.

    For those who don't know, Hurd [gnu.org] is GNU [gnu.org]'s kernel. I won't tell you here all the reasons why you should be interested on it, though.

    Alejo

  • by Anonymous Coward on Saturday December 04, 1999 @07:33PM (#1479439)
    Excerpted from www.gnu.org/software/hurd.html [gnu.org]:

    According to Thomas Bushnell, BSG, the primary architect of the Hurd, ```Hurd' stands for `Hird of Unix-Replacing Daemons'. And, then, `Hird' stands for `Hurd of Interfaces Representing Depth'. We have here, to my knowledge, the first software to be named by a pair of mutually recursive acronyms.''

  • It is still free and now also protected by the GPL.
  • If you have a bit of time left, read Open Sources [oreilly.com]. It contains (among others) an interesting article by Richard Stallman and lists up the controversy around monolithic/micro kernels back in the days when Linux was still a very small project. A well-known computer science professor tried to convince everybody else that 'Linux is obsolete' (appendix A).
  • by roystgnr ( 4015 ) <roystgnr@ticam.utexas . e du> on Saturday December 04, 1999 @07:39PM (#1479448) Homepage
    It's my understanding that while the Hurd is aiming for Posix compatibility, and going to use glibc as the standard library, they're going to be much more modular and abstracted (read: cool but possibly slow) under the hood.

    The neatest things I've heard of:

    The whole system is basically cooperating servers running atop the GNU Mach microkernel:

    the idea that every user can build up his own system on top. So, if you want to operate, start compatible servers. It's after all your decision, much like it is your personal decision if you use one desktop system (like Gnome) or Xlib, Athena etc programs together. Latter don't interoperate well (drag n drop etc). As they run in user space, you can tweak the system to your liking, even as a user.

    But there are still some servers that are the base of the Hurd system. Those are the auth, proc, init and password server at least. You don't need to register your process with the proc server (and it won't show up in the output of "ps" if you don't do so), but that the only thing that will give you access to the features of the proc server. Same with auth. If you don use auth, your tasks will have little to none privilegdes.


    Better yet, filesystem support comes from servers, which I believe means that users can have files or filesystems (limited to user permissions) that live off their own servers. Every mounted filesystem is just another new filesystem server added to the pool. No need to make smbmount suid root or put every smb share with the user option into fstab, for instance; any user process would be able to mount arbitrary smb shares in their own directories and make them viewable without being able to circument security. Cooler yet, in theory you could make .signature be randomly generated by a program without using ugly named pipe hacks, you could "cd file.tar.gz" without ugly virtual filesystem libraries, you could implement albods [linuxcare.com] more easily... I can't imagine all the possibilities, but it's fun to try.

    Superficially you'd think there's an element of jealousy there because Stallman almost wrote an OS then someone else got most of the limelight by writing the kernel... but that just doesn't feel like the case.

    There's appropriate mailing lists you can hunt down for deep info, but you can follow the Cliff's Notes of the Debian Hurd work at the debian-hurd Kernel Cousin [linuxcare.com] page.
  • Ouch, the guy is on @Home, where every customer's gateway is on the same /24 subnet with a host number of 1. So, his IP is 24.113.102.239, and his gateway is 24.113.102.1. It is a valid IP, and it isn't returning pings -- there went the entire subnet of @Home customers. :-/

    So, ironically, while most of the discussion here regards the OS on hurd.zugzug.com, a more apt topic would be the OS on the gateway. :)
  • Boyfriend huh? If her name was Shelly, wouldn't that mean they were married? Or related?

    Seriously though - I love that poem.

    -----------

    "You can't shake the Devil's hand and say you're only kidding."

  • by adubey ( 82183 ) on Saturday December 04, 1999 @08:03PM (#1479459)
    Before there was Linux, there was GNU. What is GNU? GNU's not Unix. What is Unix? A castrated Multics. What is multics? An operating system. So GNU's not a castrated operating system.

    Wait. That's not right. No, it is right, but it isn't what you're looking for, right? Right.

    GNU was intended to be a castrated operating system, er, I mean operating system. The difference is that GNU was intended to be free to distribute and modify. When Richard Stallman set about to write the GNU system in 1984, he quickly saw that he couldn't get anywhere without a compiler.

    Or tools.

    Or libraries.

    So the GNU project carried on for 7 years writing compilers, tools, and libraries, always psyching themselves up for writing the key part of the OS - the kernel. But Linus beat them to it. Linux, released under the same "GPL" liscense of the GNU System, starting getting acceptance amoung the so-called open source community (except it wasn't so-called at that time. It started being so-called in 1998. Back then it was so-called something else.).

    But the GNU people never lost their dream of writing not-a-castrated-kernel for their system. And that's what HURD is. I mean HIRD. I mean... well, I know I don't mean herd.

    The HURD kernel, technically, has one big advantage over the Linux kernel - it is microkernel-based.

    As Linus continually points out, anything you can do to make a microkernel OS fast, you can do to a monolithic kernel (like Linux). All this is true - and the monolithic kernel wins because of lower overhead. That's why Linux continually beats out high-end rockin' Solaris on single-processor machines.

    However, the overhead due to microkernels are only constant factors. Kinda like in the old days when you could write in assembler and beat the pants off of any compiler by a constant factor (depending on how good you were, and how bad the compiler was, that constant could be pretty high). But the problem with giving up a high-level language for assembly is the same problem with giving up microkernels for monolithic kernels - giving up the abstraction makes programming harder. (This might be kind of prophetic: these days compilers will beat assembly programmers almost every time)

    Since programming is already so hard (grin), the development of new abstractions has been one of the key factors behind the advancement of computer science. All abstractions (think: high-level languages, structured programming, object-oriented programming, pure-OO programming, lambda-calculus programming) increasingly make programming easier at the expense of decreasing speed.

    *But* Linux still has many more years of development behind it than HURD. Or HIRD. But not herds (some of them have been around for millenia... unfortunate about the Buffalo, though). And Linux still has more coders. So Linux will likely stay ahead of HURD for a while.... (at least until they have a stable release :).

    How does this concern you? To finish this off, I must say that unless you happen to have a 16-processor box lying around, all this won't help. Linux still rules the day on single processor machines, and probably won't be over taken.

    Since you probably don't have a massive SMP machine, everything I've written was a complete waste of time. So there.
  • by Bruce Perens ( 3872 ) <bruce@perens.com> on Saturday December 04, 1999 @08:18PM (#1479465) Homepage Journal
    There is an end to Linux. Some day, we'll be using something else. That something will most likely be Open Source, there's no going back on that one, and it will probably be compatible on some level with what we had before. But it will be a successor to Linux, not Linux. Will it be the Hurd? Maybe. But the point I want to make is that Linux is not forever.

    Thanks

    Bruce

  • I was amused that a Hurd system (the first public webserver?) got slashdotted as well, but when I couldn't connect to it I noticed something even funnier: the connection is an @Home Cable Modem!! Ah, to slashdot an experimental OS connected via a cable modem, thats just mean. :) I wonder what @Home thinks about this one..

  • I hope you gave them a little heads-up before putting this up.

    Something like "It's nice you've got this experimental new server thingy going, 'cause in ten minutes we're going to bury you."

    Or just "Hi, we run Slashdot. Brace yourselves."

    --

  • Nothing is forever, for that matter. However, the fact that Linux is an Open Source OS means, that parts of it and information gained from reading its source code will be used elsewhere, so even if one day it stops being popular, these things will still be around.
  • However, the fact that Linux is an Open Source OS means, that parts of it and information gained from reading its source code will be used elsewhere, so even if one day it stops being popular, these things will still be around.

    This is precisely the power of open-source. Closed-source software generally has little legacy; when the company that produced it goes under, or merely cancels the product, it dies. On the other hand, open-source software doesn't get buried and forgotten based on the failure of other business efforts or the whim of some project manager. Tools, libraries, even individual routines, get stripped out, cleaned off, and reused.

    HURD may or may not be successful. But one interesting thing about its architecture is that it will be much easier to "strip for parts" than the Linux kernel. So after both Linux and HURD are long gone, the latter may wind up with the most impact even if it doesn't achieve Linux's popularity.

    -Ed
  • Except that @Home doesn't use very dynamic addresses too often.

    I'm on @Home, and in the 8 months I've been with them, my IP has changed only once.

    It's against their TOS to have a server up (but everyone does, and they usually let it slide except when it's hurting performance badly). I think this case would qualify as such. Will he get a warning, or cut off for good, I wonder?

    ------
  • I've always envied the "old-timers" who brag about running Linux since 1991. You know -- the people who wax nostalgic about downloading and compiling the 0.9 kernel.

    Now, thanks to Hurd, I have a chance to get in on the ground floor. Maybe in 5 years, you'll read a post from me griping about the glibc9 to glibc10 conversion.

    Thanks, guys!
  • I am having a difficult time believing that Hurd will ever amount to anything. If it is ever to gain a place as a widly used OS, it will have to have major advantages over linux. Linux is several years ahead of Hurd in both stability and features, and i see little reason to believe that this won't contiune. A new OS will only find success if it fills a void, and i see no void for it. Windows filled a void for a GUI for cheap hardware, BSD filled a void for an open sourced OS, and Linux filled one for an open source operating system. If Hurd has something that will make it much better than Linux, then it has a chance, but if not there will be no compelling reason to switch, and it will never be of any real use. I think that RMS and the rest of the FSF should put their egos aside and work on the development of the kernel that is stable and functional. If there is some major advantage to using hurd, please enlighten me.
  • For those of you not familiar with the boyfriend of the writer of "Frankenstein", Mr. Percy Bysse Shelley (sp?) wrote a poem about the futility of all temporal efforts at earthly permanence in his poem "Ozmandias", where the narrator finds the base of an immense statue with an inscription of "Look upon my works, ye mighty, and despair", yet the only thing to be seen are the endless tracks of sand.

    Even though one day GNU/Linux will be as extinct as Minoan Linear A (a language never yet successfully translated in modern times), it was here and it has made and will continue to make (for at least the near future) the world a better place than "Microzmandias" ever did.

    May the penguin continue to evolve and promulgate worthy successors!
  • Linux as an OS certainly won't last forever, but in the long run Linus just might be remembered not for writing an OS but for creating a whole new kind of development process, one that isn't going away.

    The whole idea of 'release early, release often', invite patches from everybody, and huge-team development was actually pretty different from the way even gnu worked at the time. Even now lots of open-source projects don't get it (look at gcc, or even, dare I say it, Mozilla).

    It's funny, I stumbled across Linux (0.9x) because I had heard about Hurd, way back when. I figured I'd use Linux for the six months or so until Hurd got released. It will be great to finally be able to use it.
  • Hurd does have some advantages. As many others have mentioned, it is based on the Mach microkernel. And, as we all know, the microkernel only provides core functionality (i.e. a messaging service, and some other stuff).

    This means all of the other services are implemented and run in user space and mode. Why is this cool? Well, it's easier for you to implement your own if you want. Also, with the way kernels work, it's best to spend as little time as possible in the kernel to avoid priority inversion, and other bits of non-deterministic behavior.

    The drawbacks? Mach is _really_ slow -- the message passing just ain't fast. NT was/is based on a microkernel design, and they had to move several services (to be honest, I'm not sure what -- memory management, networking, and IO, I think) back into the kernel to get reasonable performance.

    There is definitely something to be said for doing things simply. A lot of what Linux does in the kernel isn't fancy like NT or Solaris, but they know they don't have the additional complexity so they can get through the critical sections pretty quickly most times.
  • by SnowZero ( 92219 ) on Saturday December 04, 1999 @09:35PM (#1479507)
    Remember when DOS/Windows had advantages over Linux? Yeah that was only at the *very* beginning, but that was the case for a short time during early development. Well here we are again.

    Hurd has a higher degree of extensibility than traditional kernels. Since each individual feature is easier to add, eventually you may win out if your design is indeed better.

    Supposing that people *only* work on improving things that already exist is short-sited. Everything you use today replaced something before it.
  • HURD is older then linux
  • In FreeBSD, you can have a process back a filesystem. It is called Portal FS. It is really cool.

    FreeBSD has alot of cool filesystems, check them all out. Union FS rocks.
  • by Guy Harris ( 3803 ) <guy@alum.mit.edu> on Saturday December 04, 1999 @11:56PM (#1479516)
    As Linus continually points out, anything you can do to make a microkernel OS fast, you can do to a monolithic kernel (like Linux). All this is true - and the monolithic kernel wins because of lower overhead. That's why Linux continually beats out high-end rockin' Solaris on single-processor machines.

    Given that the Solaris/SunOS 5.x kernel is a monolithic kernel (with loadable modules, but Linux has those as well), not a microkernel plus stuff atop it, how can this be "why Linux continually beats out ... Solaris on single-processor machines"?

  • I was wondering recently if there are any HURD servers out there connected to the net doing real production work - serving web pages, an FTP site, a news or mail server, CVS serving, or anything else. Does anyone know of one? Is anyone willing to subject theirs to the /. effect? The original release of the HURD was a long time ago - 0.2 was released in June of 1997. I would expect that by now there would be a usable production system. I don't intend this as a slam against GNU/FSF/Stallman or the HURD developers. I am just wondering how much momentum the HURD has, and whether people are using it.

    Cheers
    Eric Geyer
  • Why not just mrtg your ethernet adapters. Install UCD-SNMP, install mrtg, query mrtg, get statistics. Here [darkpower.net] is my traffic statistics page for my external ethernet adapter which is connected to a cable modem. I have friends doing the same on rogers.

    Now that you mention it, though, I do wonder if my terayon modem has SNMP MIB's that would allow my provider to measure traffic.. :-).
  • One of the reasons Linux will never be able to replace Mainframe operating systems is the fact that you have to reboot in order to upgrade the kernel.

    Would the microkernel architecture of HURD make it possible to do a running-upgrade of (most of) the kernel? It would at least seem to be possible, if the modules are designed with this in mind.

    Any thoughts?

Is it possible that software is not like anything else, that it is meant to be discarded: that the whole point is to always see it as a soap bubble?

Working...