Welcome to Slashdot Advertisement 
   
    faq
    code
    awards
    slashNET
    older stuff
    rob's page
    submit story
    book reviews
    advertising
    supporters
    past polls
    about
    BSI
   Feature:The Linux Standard Base System
   Linux Contributed by CmdrTaco on Tuesday May 19, @10:35
   from the stuff-to-read dept.
   Bruce Perens sent us a bunch of info about the status of the Linux
   Standard Base System. I'm running it as a feature to get some feedback
   on it. The LSB is definately one of the most important issues facing
   Linux these days; especially if we want it to continue to grow. Bruce
   also says "We would like to discuss the following proposal in BOF
   sessions at both Linux Expo and Linux Kongress. Nothing about the
   proposal is cast in concrete until after those meetings."
   
   The following comes from Bruce Perens
   
   PROJECT PROPOSAL AND CALL FOR PARTICIPATION: THE LINUX STANDARD BASE
   (LSB) project (V1.5)
   
   The Linux(R) Operating System's rapid adoption by millions of computer
   users everywhere is direct recognition of the quality of the software
   and success of the freely distributable software development model. In
   order to ensure that large software application programs, from
   binary-only tools sold by the largest software companies, to freely
   distributable desktop environments built cooperatively over the net,
   run smoothly on as many Linux-based computers as possible - the Linux
   Standard Base (LSB) Project is an attempt to define the common core of
   components that can be expected to be found in any "Linux" system.
   
   The signers of this proposal are most of the leading commercial Linux
   distributions, board members of Linux International, and key personnel
   like Linus Torvalds, the creator of Linux. We propose a set of goals
   and the organization for this project, and invite all other Linux
   distributions to join us in planning the project and carrying it out.
   
   The "base system" is the set of programs, libraries and files that are
   essential to every Linux system. These objects and their related file
   formats play a supporting role for every application. Examples of this
   include (but are probably not limited to) the C library, the format
   and placement of system files, and other necessary interfaces. Linux
   distributions traditionally do not distinguish themselves on these
   interfaces, they distinguish themselves in other categories, such as
   the applications on their system, quality and ease of installation,
   and quality and ease of systems administration as well as support for
   users. Linux distributions should maintain the base system
   collectively, as the kernel is maintained, rather than individually.
   
   The Linux Standard Base project will provide a vendor-neutral
   standard, backed by source code, upon which to build Linux
   distributions, much as the Linux kernel project provides a single
   kernel that is shared by all distributions. This standard base will be
   distributed as a reference platform from which Linux distributions may
   be derived and which application producers may use for testing, but it
   will _never_ be targeted to be an end-user solution in itself, as that
   is the role of the Linux distributions that incorporate the standard.
   
   The application of the standard will be that any program that runs
   successfully on the reference platform can be expected to run on all
   Linux systems. If they don't, the distribution creator must either fix
   a problem with their own distribution, or convince us that there's a
   bug in the sample distribution which violates the standards. This is
   not intended to prohibit distributions from making their own
   extensions to the base system, or even to use different source code
   from what is supplied in the reference platform - it's only meant to
   provide a common set of features that will be known to exist on every
   Linux system which ISVs can depend on.
   
   Participation in the base standard will assure the distributions of
   compatibility with each other for the set of applications that depend
   only on the files and libraries in the reference platform. As time
   passes, the standard will expand to include most of the files and
   libraries upon which a commercial application might depend.
   
   The Linux Standard Base System will be 100% compliant with the Open
   Source Definition. This assures all distributions that they can derive
   from it without concern over licensing problems for themselves or
   their users. Development will be carried out in the public, with
   anonymous access to the CVS archive and the developer mailing lists.
   The core group will be a mix of high-quality developers from the Linux
   community and the staff of commercial distributions, with an
   organization similar to the tremendously successful Linux kernel
   development team. Attention will be paid to standards such as POSIX
   and the FHS (the successor to the Linux Filesystem Standard). However,
   the project goes far beyond the utility of these standards, because
   rather than produce only paper documents, it will provide a complete
   implementation of the standard, ready to be integrated into Linux
   distributions or used as a reference platform for application
   developers. This will provide the Linux distributions with improved
   time-to-market, lower cost, and much less duplication of effort than a
   paper standard which is defined to fully take into account side
   effects, undocumented issues, etc.
   
   We propose Bruce Perens as the project leader. He has the experience
   of having run Debian for several years, during which he did extensive
   work on Debian's base system. He is a board member of Linux
   International, president of Software in the Public Interest, and a
   member of the 86open steering committee. He is the principal author of
   the "Open Source Definition" and its predecessor, the "Debian Free
   Software Guidelines". He has been a Unix systems programmer since
   1981, and is currently employed by Pixar Animation Studios, makers of
   "Toy Story".
   
   The Linux Standard Base System will implement some of the goals of the
   86open project, which proposes to establish an interoperability
   standard for all Unix-like operating systems.
   
   We, the undersigned, endorse this proposal, and ask that other
   distributions and ISVs also join us to help further define this
   proposal and then to help implement it:
   
   Linus Torvalds, Creator of Linux
   Jon A. Hall, Executive Director, Linux International
   Bruce Perens, Director Linux International, proposed Project Leader
   Ransom H. Love, Director Linux International, General Manager,
   OpenLinux Division, Caldera, Inc.
   Roland Dyroff, Director Linux International, S.u.S.E. Linux
   Mark Bolzern, Director Linux International, President Linux Mall and
   WorkGroup Solutions, Inc.
   Phil Hughes, Director Linux International, Publisher, Linux Journal
   Larry Augustin, Director Linux International, President VA Research
   Kit Cosper, Director Linux International, President Linux Hardware
   Solutions, Inc.
   Garry M. Paxinos, Director Linux International, Vice President Metro
   Link Incoporated.
   Cliff Miller, Director Linux International, President, Pacific HiTech
   (TurboLinux)
   Ted Cook, Director Linux International, President, Enhanced Software
   Technologies.
   Tom Lang, Director Linux International, President, H & L Software
   Eric S. Raymond, open-source evangelist and author of "The Cathedral
   and the Bazaar"
   Sam Ockman, President, Penguin Computing, Chairman, LINC: the
   International Linux Conference and Exposition
   
   Non-Linux Supporters:
   Jordan Hubbard, FreeBSD project.
   
   Linux(R) is a registered trademark of Linus Torvalds in the United
   States and other countries.
   
   Jon Hall, Executive Director of Linux International, said "The success
   of this project will mean a tremendous boost for Linux.
   Across-the-board compatibility of applications will be a tremendous
   boon for end-users, and will help Linux assume a leadership position
   in the operating-systems market."
   
   "A Linux standard supported in the spirit of Open Source software is a
   major step in removing the last obstacle to the main stream adoption
   of Linux by key hardware and software providers in the industry. This
   standard will facilitate an increased quantity and quality of business
   solutions for Linux. Caldera salutes Linux International's efforts and
   lends its support to promoting this much needed step in the evolution
   and adoption of Linux." - Ransom H. Love, General Manager, OpenLinux
   Division, Caldera, Inc.
   
   "The Linux Standard Base (LSB) project will help to insure that
   software that runs on Linux from one vendor will also run on another.
   It provides a common meeting ground for all looking to support a truly
   open solution to computing. Progress is only possible through
   communication, and communication is possible through concepts in
   common and words that express these concepts. Consider LSB to be a
   dictionary for basic Linux." - Mark Bolzern, Director Linux
   International, President of Linux Mall and WorkGroup Solutions, Inc.
   
   "Over the past six years, Linux has grown from a college student's
   idea to a serious commercial success. This project assures users that
   their choice of Linux will be supported across vendors as well as
   across hardware platforms." - Phil Hughes, Publisher, Linux Journal
   
   "It's one of the great advantages of Linux that when users have
   problems like compatibility someone like Bruce Perens can step forward
   and solve it for them. Nowhere else do you find that kind of response.
   It's that kind of problem solving support that made InfoWorld give its
   Best Technical Support award to the Linux community. VA Research will
   support the project through the donation of hardware resources and
   network bandwidth." - Larry Augustin, President VA Research
   
   "The fact that the Linux community is actively addressing the issues
   of binary compatibility between various distributions indicates the
   transition of Linux from a hobbyist's operating system to that of a
   commercially viable environment," said Garry M. Paxinos, Director
   Linux International and Vice President of Metro Link Incorporated.
   
   "The vitality of the Linux movement comes from the diversity,
   cooperation and openness of the Linux community. The LSB Project
   furthers the cooperative spirit by providing a set of common building
   blocks that draw together the various Linux flavors. It will be
   welcomed by Linux users and developers alike." - Cliff Miller,
   President of Pacific HiTech (TurboLinux)
   
   "The Linux Standard Base project, coupled with the outstanding growth
   of Linux in the marketplace, presents a compelling invitation to
   software developers to port their wares to Linux. As an ISV who
   entered the Linux market early on, I am proud to be associated with
   this project. EST's success in the Linux marketplace has far exceeded
   our original expectations. and I invite other ISV's to jump on the
   Linux bandwagon ahead of their competition." - Ted Cook, President,
   Enhanced Software Technologies, Inc.
   
   "The value of this project, both for the Linux world at large and OS's
   like FreeBSD who would have far fewer applications available were it
   not for their ability to run Linux binaries, is beyond question. A
   single standard for ISVs to follow will allow them to reach the
   maximum number of Linux and FreeBSD users with the least amount of
   effort and is in everyone's best interest. Bruce Perens has our full
   and enthusiastic support." - Jordan Hubbard, FreeBSD project leader.
   <<Ask Slashdot:Trouble After KDE Install | Reply | Flattened | POSIX
   on top of MacOS >>
   
   
     Related Links
   Linux
   Debian
   Bruce Perens
   Bruce Perens
   Related Articles 
   More by CmdrTaco 
   [INLINE]
   
     Features
   Ed has written an editorial addressing the key issues concerning the
   Various Open Source Licenses.
   
   Harry McKee writes "Report on a company developing a telepathic
   "clapper" light switch and a "conscious" Q-chip. link"
   
   Slashdot is participating in Distributed.Net's Decryption Projects.
   Read who's helping, and why. Check out Our Current RC5 Rank.
   [INLINE]
   
     Quick Links
   CDnow If you follow this link and buy something, I get free CDs. This
   makes me happy.
   
   Follow this link to the LinuxMall and buy all sorts of cool Linux
   stuff. If you do, I get beer money.
   
   Nerds, Unix, and Virtual Parenting A short story by Rob Malda.
   [INLINE]
   
   Good news for developers
   by Matt Kimball (mkimball@xmission.com) on Tuesday May 19, @10:49
   http://www.mkimball.org/
   It's about time we standardized on things like glibc and pthreads. It
   will make life much easier on developers if we can know that all
   systems meeting the standard have a particular major version of those
   libraries in a standard place.
   [Reply to this comment]
   
   Re: Good news for users too. by Sandor Lengyel on Tuesday May 19,
       @11:00
       
   Redhat?
   by Tanner Lovelace (lovelace@nospam.acm.org) on Tuesday May 19, @10:54
   Sounds great, but what about RedHat? I don't see them on the list.
   [Reply to this comment]
   
   Re: Redhat? by Erik Troan on Tuesday May 19, @02:21
       
   Good news for Users
   by Arne Hueggenberg (zalaster@nef.wh.uni-dortmund.de) on Tuesday May
   19, @10:56
   I think that, sould this priject go forward, wich i surely hope it
   will, it would benefit users tremendously.
   No more headaches cause the configure script coming with the xyz
   tarball expacts an library somwhere else than your distributor placed
   it and the like.
   I think this will go a long way to making linux more user friendly.
   [Reply to this comment]
   
   standard base system
   by Korey () on Tuesday May 19, @10:56
   The question that comes to my mind is how extensive this base system
   will be. The larger this base is, the more hd space will be needed to
   run it.
   For example, you could run a RH 4.1 or 4.2 distribution with 60 megs
   of hd space. Glibc2 in RH 5.0 makes that hard to do. Is Xwindows a
   standard part of a distribution?
   I can see certain tools being required in a core system, such as vi,
   tar, mount and so forth, but there will be some disagreement on things
   like the init format, where RH & Debian differ greatly.
   I'm just throwing out bait for discussion.
   [Reply to this comment]
   
   Re: standard base system by BadlandZ on Tuesday May 19, @11:21
       
   Re: standard base system by Jens on Tuesday May 19, @12:31
       
   Re: standard base system by BadlandZ on Tuesday May 19, @12:50
       
   Re: standard base system by Jens Owen on Tuesday May 19, @03:22
       
   Re: standard base system by Victor R. Rivarola on Tuesday May 19,
   @01:18
   
   Re: standard base system by Matthew Mc Clement on Tuesday May 19,
   @11:35
   
   Keep the base small by Fringe Ryder on Tuesday May 19, @01:32
   
   odd... that list didn't include
   by Andy Kahn (kahn@zk3.dec.com) on Tuesday May 19, @11:00
   the list of endorsers/supporters for this proposal didn't include
   anyone from Redhat nor Debian.
   [Reply to this comment]
   
   Re: odd... that list didn't include by BadlandZ on Tuesday May 19,
       @11:23
       
   Re: odd... that list didn't include by Steve McIntyre on Tuesday May
   19, @11:25
   
   This is what I've been waiting for...
   by Jay Gramlich (gramlich@hotmail.com) on Tuesday May 19, @11:03
   I hope this gets off the ground and RedHat jumps in too. In my
   opinion, this is one of the major
   advantages of the BSD's and can do nothing but help Linux.
   [Reply to this comment]
   
   Standards
   by BadlandZ (BadlandZ@cgsa.chem.und.nodak.edu) on Tuesday May 19,
   @11:17
   http://cgsa.chem.und.nodak.edu/~current/badlandz
   I read throught a copy of that last Wednesday, thought it was a good
   idea, but there are some details that need to be worked out.
   Would be great to be able to know stuff is there, know libs are where
   thier suppose to be, be able to count on distributions being as
   complete and clean as Linus has kept the kernel. It's about time
   structure and distributions caught up to how well Linus has done with
   the kernel.
   [Reply to this comment]
   
   Linux filesystem standard?
   by Nelson Minar (nelson@media.mit.edu) on Tuesday May 19, @11:26
   http://www.media.mit.edu/~nelson/
   How is this different than the Linux filesystem standard? The goal
   seems similar. Is this more about functionality? Will it be based on
   the filesystem standard?
   [Reply to this comment]
   
   Re: Linux filesystem standard? by Ray Dassen on Tuesday May 19, @11:40
       
   Making a case to change file system
   by BadlandZ (BadlandZ@cgsa.chem.und.nodak.edu) on Tuesday May 19,
   @11:50
   http://cgsa.chem.und.nodak.edu/~current/badlandz
   Here is what I was thinking. Moving /home back into /usr/home for some
   disk
   management reasons, like IRIX and BSD have it. LINUX was ment for a
   PC, and
   therefore it is quite frequently run on home systems with only one or
   two users,
   so this could be very important.
   Why? Well, K, I have some systems with small disks, and I like to keep
   /home and /usr on the same partition, so that I can download, and
   compile programs in /home and then install them in /usr without
   running out of drive space. I know,
   drives are cheap, but low cost of hard drives is not an excuse for
   mismanagement.
   To put /home and /usr on the same partition now, I have to make one
   big / partition that everything goes on, even root ("/"), and that
   sucks for a couple reasons.
   Disk access time is fastest on the first partitions, and therefore,
   the /swap partition should be as close to the first partition as
   possable.
   So, ideally, partitioning for a lot of people would be:
   First Partition "/" root
   Second Partition "/swap" where it's fast
   Next partition "/usr" to load applications from
   Now, if "home" was in "usr" you could stop partitioning there, OR,
   have the choice to add a "/usr/home" partition (along with /var and
   /tmp, or whatever someone chooses).
   But when "home" is just "/home" then unless you make a "/home"
   partition, it goes in root ("/"), which makes it harder to combine
   "/home" and "/usr" and keep the "/swap" down at a one of the first
   partitions to make access fast.
   I know this can be done by simply making a symbolic link, and I have
   done so before. But, I would think that more people would benifit from
   this too, and after explaining why I did it soo many times, I thought
   "why isn't it the other way to start with"
   [Reply to this comment]
   
   Re: Making a case to change file system by Bob Tinsley on Tuesday May
       19, @12:07
       
   Re: Making a case to change file system by BadlandZ on Tuesday May 19,
       @12:40
       
   Re: Making a case to change file system by doug on Tuesday May 19,
       @01:15
       
   Placement of file systems by Ben Kosse on Tuesday May 19, @03:26
       
   Re: Making a case to change file system by Andrew Mobbs on Tuesday May
   19, @12:16
   Re: Making a case to change file system by Daniel Mashao on Tuesday
       May 19, @12:39
       
   Re: Making a case to change file system by Andrew Mobbs on Tuesday May
       19, @12:57
       
   Re: Making a case to change file system by BadlandZ on Tuesday May 19,
   @01:23
   Re: Making a case to change file system by Nathan Rice on Tuesday May
       19, @02:43
       
   Re: Making a case to change file system by BadlandZ on Tuesday May 19,
       @03:14
       
   Re: Making a case to change file system by BadlandZ on Tuesday May 19,
   @03:15
   
   Re: Making a case to change file system by RF on Tuesday May 19,
   @12:27
   Re: Making a case to change file system by BadlandZ on Tuesday May 19,
       @12:43
       
   Welcome to the big time.
   by Art Cancro () on Tuesday May 19, @12:01
   http://uncnsrd.mt-kisco.ny.us
   Yup ... this is it ... the final obstacle to overcome in making life
   easy for commercial software vendors who wish to make their
   applications available on Linux.
   
   On other operating systems, you've got a pretty good idea of what to
   expect wrt placement of various system files and API's. Consider how
   well RPM works for RedHat-specific installs. Consider how brainless a
   (gasp!) Windows application install is: you run the installer and it
   knows where your system files are, and it knows how to set up icons on
   the desktop.
   
   This is exactly where we need to be with Linux. If developers can
   provide install programs that will completely set up a program,
   including all files and settings, Linux can make a large leap forward
   towards becoming an operating system suitable for end users. I for one
   would love to see apps labelled "Ready to install under Linux base
   system 1.1 or higher..." and such.
   
   Right now I'm trying to get together a binary distribution for my
   Citadel/UX BBS software and am running into lots of trouble trying to
   get it to figure out which libc the user has, whether pthreads is
   properly installed, etc. etc. Having a working reference platform
   available would make this much easier.
   [Reply to this comment]
   
   Re: Welcome to the big time. by Nick Moffitt on Tuesday May 19, @12:10
       
   Re: Welcome to the big time. by Steve Baker on Tuesday May 19, @01:07
       
   Re: Welcome to the big time. by Patrick Michael Kane on Tuesday May
   19, @12:31
   
   Re: Welcome to the big time. by RF on Tuesday May 19, @12:35
   Re: Welcome to the big time. by Sandor Lengyel on Tuesday May 19,
       @12:49
       
   Re: Welcome to the big time. by Danny on Tuesday May 19, @12:54
   Re: Welcome to the big time. by Brion Vibber on Tuesday May 19, @02:23
       
   Re: Welcome to the big time. by Victor R. Rivarola on Tuesday May 19,
   @01:37
   
   Re: Welcome to the big time. by BadlandZ on Tuesday May 19, @01:48
   
   Re: Welcome to the big time. by Sam Lantinga on Tuesday May 19, @03:40
   
   startup scripts
   by Thayne Harbaugh (thayne@plug.org) on Tuesday May 19, @12:18
   www.plug.org
   Hmmmm, I wonder how they will deal with the startup scripts. Some
   prefer BSD style rc scripts while others prefer Sys V style runlevel
   start/stop scripts.
   I don't see these as being compatible, yet they are essential to
   properly configuring daemons and services.
   These are a distinguishing feature between distributions.
   [Reply to this comment]
   
   Re: startup scripts by Victor R. Rivarola on Tuesday May 19, @12:48
       
   Classes of Base System?
   by SEGV (mlepage@cgocable.net) on Tuesday May 19, @12:57
   http://www.cgocable.net/~mlepage/
   How about having classes of base systems? A class A base system would
   be the bare minimum, class B would include class A but have a little
   more, same with class C which would also have X Windows.
   Then you could say this program will run on a class A Linux system, or
   a class C Linux system with this extra library, and so on.
   [Reply to this comment]
   
   Re: Classes of Base System? by BadlandZ on Tuesday May 19, @01:30
       
   Why isn't Red Hat on the list?
   by Steve Shreeve (shreeve@uci.edu) on Tuesday May 19, @01:10
   http://www.med.uci.edu/~shreeve/
   The list mentions somewhere between 15-20 individuals, but I didn't
   spot anyone from Red Hat on there. Since the great majority of "users"
   of Linux seem to be running Red Hat, why aren't they included on the
   list? If someone from Red Hat *is* on there and I missed it, then my
   bad...
   [Reply to this comment]
   
   Party Pooping
   by Michael Bacarella (defile@nyct.net) on Tuesday May 19, @01:19
   Much as I hate to be the Party Pooper (tm), I see
   a few slight problems with this scheme (as others
   have already pointed out).
   Linux's diversity is a strong point, if I can buy
   a Linux distribution that comes on 4 disks and
   install it on a 386 w/ 4 megs of ram and a 40
   meg hard drive, that's a strong point IMO.
   (Although DR-DOS would probably be a better
   candidate, but, shutup :)
   I like the hierarchial system proposed above.
   class a, class b, class c, etc system.
   If a vendor fails to do so, what are we going
   to do to them? Shun their systems? Declare them
   heretics aiming to compromise the advancement
   of Linux!!@#$! ?
   heh
   Also, someone mentioned reading about this
   a time back, if a paper has already been written
   on how to structure the system, I would
   definately like to read it. (As a developer, I
   wouldn't mind a head start here)
   (A home page wouldn't be such a far fetched
   idea, now would it? :)
   /me places a "LSB compliant" stamp on his page
   with a link to the main page.
   I hope I sounded at least somewhat serious here.
   --Michael 'glDeFile' Bacarella
   [Reply to this comment]
   
   Dependencies and configuration
   by ewjc (ewjc@iname.com) on Tuesday May 19, @01:31
   Within a software package, we may see a lot of Makefiles as well as
   configure scripts to detect the dependencies between system
   libraries/utilities and the program source. After
   compilation/installation, we are supposed to get a working system,
   right? In reality, this is not the case. The main problem, as I
   claimed many times before, is the lack of system configuration
   information detection in many programs. Many programs (especially
   related to networking stuffs) make a lot of assumptions and shrug a
   lot of post-installation responsibilities. These kind of problems will
   haunt the free software movement if we don't address the issue now.
   The assignment of programs/file systems/configurations are really
   trivial compared with the dependency problem. An ftpd program needs
   certain pre-conditions and post-conditions to work satisfactorily;
   cron/reminder have important configuration steps to work out before
   starting your own daemon. Some programs really depend on
   the other programs to run correctly. The Redhat rpm files and debian
   packages put the sugar coating on the mess but are indeed assuming
   that you install the system from scratch. Getting all these things to
   work properly is a system administrator's nightmare and you guys
   expect a layman to have a Linux home machine?
   Before the GNU autoconfigure/automake, cross-platform compilation was
   practically a boring and tricky work. Then we can compile programs to
   "fit" into the system at the *library* level. It's probably necessary
   to start a project on autoconfiguring a given package to fit into the
   given system. Yes, getting packages like GIMP to work perfectly is
   trivial since it's practically self-contained. All the messy things in
   networking, multiple-compiler, multiple-library management can be
   quite hard and potentially prohibits system integrators from getting
   things done quickly.
   From the very beginning (Linux's pre-X11 age), I always hand-craft my
   system. I never install redhat/slackware/debian packages because I
   don't trust them. I do sincerely hope that certain smart perl scripts
   will be able to autoconfig the installation of packages safely and
   properly regardless how the current system is composed.
   [Reply to this comment]
   
   need "lesstiff" now more than ever!
   by BadlandZ (BadlandZ@cgsa.chem.und.nodak.edu) on Tuesday May 19,
   @01:51
   http://cgsa.chem.und.nodak.edu/~current/badlandz
   I was just thinking, if we have a standard set of Open Source Libs and
   stuff, "lesstiff" is going to be really important. Motif is expensive
   (to me), and compiling stuff like Mozilla without it leaves you seeing
   just how much work is left to be done on lesstiff.
   I sure wish lesstiff worked as well as motif. But, there isn't much I
   can do about it, since I am primarly an end-user and (very) small time
   systems administrator (and only as a hobby, since I am a Chemist by
   day, not a programmer).
   [Reply to this comment]
   
   Yipee!
   by David McCanney (D.McCanney@udcf.gla.ac.uk) on Tuesday May 19,
   @02:11
   http://www.gla.ac.uk/Clubs/GLUG/
   Yes! I've been waiting for this standard for quite a while. It won't
   be long 'till commercial UNIX vendors are saying that they're 'Linux
   99' compliant... :-)
   
   There might be some disagreement over init scripts but in the end,
   Linus et al. should just choose one method and be done with it.
   [Reply to this comment]
   
   Very good
   by Stefan Wille (wille@netlife.de) on Tuesday May 19, @02:29
   This is very good, I hope Red Hat and Debian
   will join the effort.
   [Reply to this comment]
   
   So, it's public now. Make it PUBLIC
   by BadlandZ (BadlandZ@cgsa.chem.und.nodak.edu) on Tuesday May 19,
   @02:38
   http://cgsa.chem.und.nodak.edu/~current/badlandz
   
   PROJECT PROPOSAL AND CALL FOR PARTICIPATION: THE LINUX STANDARD BASE
   (LSB) project (V1.5)
   
   Judging from the title, and the fact that it's now on SlashDot (which
   means I guess now it's public news), I really think this thing should
   be posted on some other sites like www.linux.org, www.linuxhq.com,
   freshmeat.net, etc... And, does it have a mailing list yet? I know it
   has been floating around the "closed developer" lists for some
   distributions for a while, but now that it's public, maybe there
   should be a list for it?
   [Reply to this comment]
   
   Configuration probing/collection
   by Christopher B. Smith (cbsmith@envise.com) on Tuesday May 19, @02:39
   http://www.envise.com/
   I'm not so sure that locking us all in to a particular filesystem
   configuration is such a bright idea. It will definitely slow down
   deployment of new software/technologies. You'll notice that
   NT/Win95/Macintosh have even less of a standard than Linux currently
   does. So why does it work?
   Well, the trick is that they don't ASSUME location of files or what
   libraries are installed. Installation programs check for where files
   are, and should they be unable to find them, they go ahead and install
   them. Someone should write a standardized installer that handles all
   of this (Internet based loading and installing of other software would
   be most cool!). Think about Perl, which installs wonderfully on almost
   any system. It has scripts which check for all kinds of things, and
   then makes the appropriate adjustments. Window's applications do the
   same thing.
   Finally, it of course would be nice to have a text file somewhere that
   stores configuration information kind of like home Imake works with
   Xwindows apps. Anyway, it's a dream, but somebody
   really should build a full function installer like
   that.
   [Reply to this comment]
   
   Re: Configuration probing/collection by BadlandZ on Tuesday May 19,
       @02:46
       
   Re: Configuration probing/collection by Anonymous Coward on Tuesday
   May 19, @04:16
   
   Comments about the Linux Standard Base proposal
   by Frank LaMonica (frankl@precisioninsight.com) on Tuesday May 19,
   @02:41
   http://www.precisioninsight.com
   The LSB idea is both necessary and timely if Linux is to continue its
   emergence into the role of a pervasive and respected operating system.
   My only caution is that "the devil is in the details" and we should,
   as Mr. Perens suggests: "...discuss the following proposal in BOF
   sessions at both Linux Expo and Linux Kongress.", and remember, again
   as Mr. Perens says, that: "Nothing about the proposal is cast in
   concrete until after those meetings."
   Because the proposal is still simply that - a proposal - it is much
   too early to expect endorsements from any quarter. The most we can and
   should do is to strongly endorse the concept of an LSB, to sincerely
   thank those people who are donating their time to further the idea,
   and to express our own commitment to contributing to the very large
   task that lies ahead. The LSB team must create a realization of the
   concept that is beneficial and acceptable to users of the operating
   system, ISV's, and to all of the major distributors of Linux. That's a
   big job and it's a job that will demand a lot of "give and take" from
   all those involved. Let's not break out the Champagne yet; too many
   good ideas burn out before they're given a chance to mature.
   [Reply to this comment]
   
   Multi tiered std
   by chris clark (clarkcd@hotmail.com) on Tuesday May 19, @02:44
   I'm very much in favor of a multi-tiered std.
   Caldera , for instance has a spot in the install
   script which offers several types of install and
   gives approx installed size. I'd like to see many levels, from
   ridiculously tiny (I dunno how tiny, how about
   a headless box without x for instance), on up.
   (as a side comment, I put RedHat on my laptop recently
   and picked "everything " to install, and was
   surprised to find that it didn't bother to
   verify that there was enough free space avail before
   doing the install. That seems like an obvious oversight.)
   [Reply to this comment]
   
   Ideas from other OS's
   by BadlandZ (BadlandZ@cgsa.chem.und.nodak.edu) on Tuesday May 19,
   @02:48
   http://cgsa.chem.und.nodak.edu/~current/badlandz
   I do like how I can "cat /proc/cpuinfo" in LINUX, but I also like
   "hinv" in IRIX. I was wondering if there was some POSIX standard LINUX
   may move twards or some LINUX standard they may creat in reguard to
   "hardware awareness" of the system.
   People have been talking about "software/appication/library"
   awareness, but what about "hardware awareness"?
   [Reply to this comment]
   
   FreeLinux
   by Rob Cabacungan (spassky@idt.net) on Tuesday May 19, @03:32
   The FreeLinux project seems to have very similar goals, though not so
   carefully elucidated. They are developing a baseline, self-hosting
   glibc-2.0.9x based Linux system for many architectures.
   
   [Reply to this comment]
   
   It could be quite simple
   by Sam Lantinga (slouken@devolution.com) on Tuesday May 19, @03:51
   http://www.devolution.com/~slouken/
   It seems to me that the base Linux system could be quite simple:
   net-tools, GNU fileutils, bash, one or both styles of init, etc. The
   beauty of what I see being proposed is that it makes official what
   most of us take for granted already.
   The bells and whistles like Qt, Python, sz/rz, etc. would be detected
   normally. I see this as an extension to the Linux Filesystem Standard,
   including a set of base libraries (C, math, threads, etc) and programs
   that can be relied upon. This doesn't necessarily stifle distributions
   (though it could if it gets way out of hand ;-)
   My 0.02
   [Reply to this comment]
   
   But...???
   by Jon Bobson (bobson@freecar.com) on Tuesday May 19, @03:53
   Isn't SEUL (http://www.seul.org) doing the same thing now? What
   happens when there's a bunch of different people trying to set a
   standard?
   [Reply to this comment]
   
   a great, quick start...
   by ziggy (ziggy@panix.com) on Tuesday May 19, @03:55
   The layered/hierarchical approach discussed above solves a lot of
   problems. One thing that seems to be lost is that there are going to
   be wizards who know what they're doing, and LSB seems to be geared at
   getting Linux to the masses. Existing projects like Beowulf supported
   by Extreme Linux will probably remain outside this sphere for a while,
   as will the Linux mini-router project (but still might be based on a
   specific baseline). As would palm pilot linux, the linux mini-router,
   and so forth - and that's OK.
   
   What's left are a huge number of people who want a reliable setup that
   removes the drudgery or a simple setup that minimizes the wizardry
   needed (Grandpa's Linux).
   
   The other half of the bargain, once features like X/Netscape/Apache
   get shunted into the appropriate tiers, is the upgrade problem.
   
   Having something like LSB-ClassA-Mar98 that's easily upgradeable to
   LSB-ClassA-Dec98, LSB-ClassB-Jun98 or LSB-ClassC-Jan99 would be
   wonderful. Each blessed LSB would contain a well-known list of
   components (kernel 2.0.34, glibc, gcc 2.8.1) that could be counted on
   to be there based on the blessing of that configuration. Upgrading a
   box would become more trivial and less vendor-specific (going from
   Debian LSB Class A 1998 to RedHat Class C 2000).
   
   Of course, you could deviate at your discretion, but that deviation
   implies you know what you're doing in the first place (this would
   break your system's "blessing").
   
   Furthermore, using such a "thumbprint" will help track down security
   related issues. Assuming that a baseline Class A configuration is
   intended to be used as a PC, the standard configuration would lock
   down things like Apache, NFS exports, mail relays, etc. rather than
   the current one-size-fits-all distribution philosophies.
   
   One other benefit is that s/w distributions could have recommended
   idiot checks tied to the baseline of your system - i.e. don't bother
   downloading GNUstep if you don't have X (e.g. class B or better).
   [Reply to this comment]
   
   Criteria, and prioritization
   by Will Page (willpage@rocketmail.com) on Tuesday May 19, @03:59
   Five criteria... in this order:
   Stability
   Support (continued development, and documentation)
   Efficiency/speed
   Size
   Quality of UI
   First of all, a Linux box should be solid. Above all else, it should
   work. If anyone disagrees with me on this, I'll sell you a copy of
   Win95, real cheap.
   Second of all, it should be well documented and well supported, for
   obvious reasons. I'll bet even the members of the Most High Order of
   Linux Gurus must RTFM for new software. If you can't use it, and you
   can't get help, it's useless to you.
   Third is efficiency/speed. How slow is too slow? How fast is fast
   enough? Well, we'll just have to take the best we can get (See #1 and
   #2, first).
   Fourth is size. Gotta kill bloat before it starts. Nobody likes bloat.
   If you can't run Linux on a 386sx with a 40MB hard drive, and run it
   as a high load web server, what good is it? *grin*
   Fifth is quality of UI. Not necessarily ease of use. A quality UI
   makes itself appealing to both newbies and gurus. IMHO, Pine has a
   nice, quality UI. Well organized, powerful, and quite useable for both
   newbie and power user. And what good are whiz-bang features, if you
   can't figure out how to use them?
   [Reply to this comment]
   
   Is X and some X toolkit included?
   by Erik Hensema (hensema@xs4all.nl) on Tuesday May 19, @04:05
   This standard is great, I've been thinking about such a standard
   myself too.
   Does this standard include X support and most importantly, a X
   toolkit? We really need that...
   [Reply to this comment]
   
   Re: Is X and some X toolkit included? by BOredAtWork on Tuesday May
       19, @04:18
       
   Hmm...
   by BOredAtWork () on Tuesday May 19, @04:11
   Hmm... how did Eric Raymond get on this list? C&B was a decent paper,
   but nothing truely earth-shattering. I just served to prove linus's
   development model works - which the kernel shows, and showed many
   years ago. Fetchmail is good - but GIMP is better. The man hacked out
   one hell of a mail utility - but others have hacked apps that are just
   as important. To be fair, we'd need gimp developers, vim developer,
   etc. The key should be _continued_ support of free software, etc. In
   this respect, I think Raster would be a better choice (still unfit,
   but a better choice).
   Redhat says they're interested, but where's stampede in all this?
   www.stampede.org - new flavor of linux, looking very promising. I
   think they should be included.
   As to what to include, I think SMALL is good. Just a standard location
   for X (still make it optional to install, but have each setup prog
   install to the same place), vim, glibc2, and system configuration
   files. Where I want to install netscape, ghostscript, gimp, gtk,
   gnome, E, window managers, etc., etc., etc., should be my business -
   if I want to install them at all. Please don't let this project get
   carried away and make a bloated '95-ish 50-60 meg installation any
   type of standard.
   [Reply to this comment]
   
   I rather like Slackware
   by Will Page (willpage@rocketmail.com) on Tuesday May 19, @04:26
   How about model it after Slackware's scheme?
   Software can just state it requires the A (base) and such-and-such
   standard.
   For example... You want to try out the latest Doom-like game. You
   don't know if you have everything you need to run it. So you check,
   and find out that it requires:
   A (base)
   Y (games)
   and for multiplayer action:
   N (network)
   Wouldn't that be so much easier?
   And as far as versioning...
   A.1998.3.10
   Y.1998.3.10
   and for multiplayer action:
   N.1998.3.10c
   The LSB version 1998.3.10
   and to designate bugfix releases since that official release,
   1998.3.10a, 1998.3.10b, etc.
   The biggest drawback I can see with that scheme, is with disksets that
   get updates more often than 26 times between LSB releases. For
   example... networking fixes. They seem to be semi-weekly.
   Anyhow, just in general, I find the slackware organization quite
   workable for this situation.
   [Reply to this comment]
   
     ____________________ Search
   
   
     There are two times when a man doesnt understand a woman -- before
                       marriage and after marriage. 
                                      
   All trademarks and copyrights on this page are owned by their
   respective companies. Everything else is  1998 Rob Malda.
   
     [ home | awards | supporters | rob's homepage | contribute story |
      older articles | advertising | past polls | about | faq | BSI ]