| DistroWatch Weekly
|DistroWatch Weekly, Issue 58, 19 July 2004
Welcome to this year's 29th edition of DistroWatch Weekly. Thanks to all who sent well wishes to me (Robert) during my illness. Now on to more interesting topics below.
GRUB - The Second Coming
In order to boot Linux, you need a boot loader (note that not every OS needs or has one - MS-DOS and Windows 95/98/Me, for example, never had a boot loader). These days, when people tend to install more than one OS on their hard drive(s), one should ideally have a "multiboot" loader (one which can install multiple OSs). In years past, people would actually pay money for such a thing (anybody remember System Commander which sold for US$99?).
Nowadays, we have several good free open source alternatives to choose from. From the earliest days of Linux, there was LILO (LInux LOader), but more recently GRUB (GRand Unified Bootloader) has gained popularity. As for which one is better, it's sort of like the war between vi and emacs - good luck on trying to get a room full of geeks to agree on this one. Many Linux distros avoid the debate by offering both LILO and GRUB, allowing users to choose which one to install.
GRUB was originally the creation of Erich Stefan Boleyn. Due to some other commitments, Erich couldn't continuing maintaining GRUB. The current maintainer is Gordon Matzigkeit, and GRUB has morphed into a GNU project with many new features added.
Good as GRUB is (and it is really good), it's about to get a major shot of steroids. A new, improved more powerful GRUB dubbed "GRUB 2" is in the works. The original GNU GRUB has been renamed "GRUB Legacy". GRUB 2 offers the following benefits over GRUB Legacy:
- Create a compact core image. This will make Stage 1.5 unnecessary.
- Add support for dynamic loading to the core image, so that it can
extend itself at the run time rather than at the build time.
- Add real memory management, to make GNU GRUB more extensible.
- Add a more object-oriented framework for file systems, files, devices,
drives, terminals, commands, and OS loaders.
- Add support for internationalization. This includes support for
non-ASCII character code, message catalogs like
fonts, graphics console, and so on.
- Add an actual localization, based on the above internationalization
support. We will target on Japanese as the first step.
- Segregate code specific to i386-pc from generic code, to make
GNU GRUB portable.
- Add support for cross-platform installation.
- Develop additional software packages which will help our project and
hopefully other projects.
More details can be found on the GNU GRUB home page. Also see the "GRUB To The Rescue" article under the new "Tips, Tricks and Hints" column.
A New BSD Rises
Practically every week we see the birth of a new Linux distro (and often the death of an existing distro as well). Yes, Linux distros come and go with the wind. Many of these distros were the creation of a single starving student, and in some cases never get used by more than one person. True, a few distros are the work of a team of paid professional programmers who intend to take the world by storm with their new product, but that's the exception rather than the rule.
It's a rather different situation in the BSD world. This may, in part, be due to the fact that there are still far fewer BSD users than Linux users in existence, but another factor is no doubt the existence of Linux From Scratch which allows any home hobbyist to slap together a Linux distro relatively quickly. Perhaps in the future we'll be seeing a BSD From Scratch, but for now there isn't one.
In view of the above, it's quite a major (rare) event when a new BSD comes into existence. Last week we witnessed the arrival of DragonFly BSD 1.0 (and this week a 1.0a bug-fixed version). The project is the brainchild of Matt Dillon and some devoted followers. DragonFly is a fork of FreeBSD 4.x - apparently the DragonFly developers were not happy with the direction FreeBSD 5.x was taking. Specifically, they wanted to implement SMP in a different way from FreeBSD 5.x, and they've also taken out a fair bit of the Perl dependent stuff, plus they thought they could do a better job of writing an installer.
Everyone who's tried DragonFly agrees that it's fast, but also far from complete despite the "1.0" (or "1.0a") designation. The DragonFly team will have their work cut for them trying to equal and improve upon FreeBSD, which by all accounts is no slouch of an operating system.
Also this week, the folks over at ekkoBSD decided to throw in the towel. EkkoBSD was a recent fork of OpenBSD, and the developers had some good ideas, one of which was to make the system more user-friendly. However, as they soon discovered, creating a major operating system is not trivial, and all the more so when the developers are basically volunteers doing the work in their spare time.
Which begs the question - does the world really need another BSD? Given that the BSDs don't command a very large share of the world's OS market, wouldn't it be better for developers to pool their resources rather than going off in different directions? The Unix market has been afflicted with this issue for several decades, and some say that Linux (which is relatively young compared to Unix) already suffers from "distroitis". Perhaps less is really more? Then again, competition is generally a good thing - isn't it? An interesting point for debate.
Along similar lines to the above story, some have wondered whether or not Linux should be "standardized" in some way or other to avoid the fragmentation that turned Unix into a synonym for chaos. Some may recall United Linux, announced May 30, 2002. Undoubtedly the most ambitious attempt at getting Linux developers to cooperate, United Linux was an alliance between Conectiva, SuSE, SCO and Turbolinux. As things turned out, SCO had a change of management and became anything but an ally. The remaining three players managed to turn out version 1.0, but there are no plans for 2.0. To look back at articles like this one, it seems almost laughable that UL was taken so seriously.
Less ambitious than UL, but ultimately more influential, has been the Linux Standard Base. In the early days of Linux development, distros were so non-standardized that developers were giving different names to the same files, and putting them into different directories. The LSB (which relies on voluntary compliance) has been fairly influential. However, in a recent article posted at Newsforge, writer Jay Lyman questioned whether or not the LSB could survive.
There are several significant things about the Linux Standards Base that the author didn't mention in the article. The LSB is the offspring of the Free Standards Group, an organization whose membership is dominated by corporations. That in itself is hardly scandalous, but more troubling is that the Free Standards Group has been reluctant to take input from projects outside the member companies influence. In particular, Debian users feel particularly miffed that the LSB specifies that software packages should be delivered in RPM package format. This is despite the fact that Debian's "deb" format is both older and - in the opinion of many - better. Debian developers have basically rejected this "guidance," and ironically the "deb" format has been picking up converts and could very well wind up pushing RPM to the fringe. On the other hand, some have argued that this is a non-issue since the RPM format specified by the LSB is well-supported by the Debian "alien" program. However, not everyone is totally convinced, and the question remains, "Is the LSB going to survive?".
Internet Explorer Living On Borrowed Time?
Just when you thought that Microsoft might finally be getting its act together in terms of security, someone comes along and pours cold water on the whole idea.
It's no secret that many of the security holes in Windows are a direct result of Microsoft's decision to co-mingle Internet Explorer's code base with that of the operating system. This is the very opposite of the traditional modular approach of Unix programming. The deeper you integrate a program into the operating system, the likelier that a programming bug will have far-reaching ramifications that no one foresaw. It also means that fixing one simple bug could trigger a whole new slew of bugs. Thus when a security exploit is discovered in Internet Explorer, Microsoft's programmers (who by all accounts are actually quite good at their job) are forced to spend a huge amount of time testing to make sure that their patch won't create a worse problem than it solves.
In view of the above, one would expect that Microsoft would be expending an effort to make Internet Explorer more modular, but according to this story published at OSNews, the exact opposite seems to be taking place. The author, Roberto J. Dohnert, participated in an online chat with the Internet Explorer development team, and was told that IE would be integrated even more tightly into future versions of Windows. This is, no doubt, great news for developers of anti-virus software and computer security consultants. Indeed, one can almost hear the champagne bottles being uncorked at the corporate headquarters of major anti-virus vendors. As someone once said, the more things change, the more they stay the same.
However, Mr. Dohnert took his arguement one step further and suggested that IE's continuing security problems might lead to it being abandoned totally by users. Personally, I doubt that. I have many Windows-using friends who have suffered repeated virus attacks - their email address books have been hijacked, their machine was made into a spam zombie, and their hard drive became a repository of warez, all without their knowledge. And what did they do about it when they finally discovered the problem? Nothing. As long as they can boot, surf the web and send an email to grandma, everything is fine. Downloading and installing a security patch is just too much trouble.
As for those of us who don't even have Windows installed, this is not the time to become smug. Security exploits have indeed plagued non-Microsoft browsers. Other than keeping your system up-to-date, the best protection is running a good firewall. If you're not running a server, you probably don't need any incoming ports to be open, so lock them down tight. Complacency can be your worst enemy.
|Released Last Week
Yoper V2 has been released: "The
Yoper team is pleased to announce the public release of Yoper V2.
Combining the elements of raw speed with the renowned stability of
Linux, the V2 is the fastest out-of-the-box Linux system in the world.
Seamless updating and software integration provided by Apt/Synaptic.
Leading edge Technology enhancements using Linux kernel 2.6.7,
bolstered with performance patches, innovative pre-linking, win4lin
support, VMware support, VMware module integration, NVIDIA 3D support,
ReiserFS4, secure shell file system, CDROM supermount, and Ndiswrapper
for Windows binary driver integration." The full announcement. Download: yos-i686-2.0.0-9.iso (644MB).
DragonFly BSD 1.0
DragonFly BSD is out: "One
year after starting the project as a fork off the FreeBSD-4.x tree, the
DragonFly Team is pleased to announce our 1.0 release! We've made
remarkable progress in our first year. We have replaced nearly all of
the core threading, process, interrupt, and network infrastructure with
DragonFly native subsystems. We have our own MP-friendly slab
allocator, a Light Weight Kernel Threading (LWKT) system that is
separate from the dynamic userland scheduler, a fine-grained system
timer abstraction for kernel use...." Find the full announcement on dragonflybsd.org. Download: dfly-1.0REL.iso.gz (78.6MB); also available via BitTorrent. Update: Release updated to 1.0a (78.6MB) to fix a serious fdisk/slice issue with the installer. An xdelta patch is also available for people who have downloaded the original 1.0REL iso.
Puppy Linux 0.9.1
Puppy Linux 0.9.1 has been released. These are some excerpts from the release notes: Mozilla
has been upgraded to version 1.7, with Skipstone as the GUI front end.
Skipstone displays as version 0.8.4 however it is actually a hybrid of
0.8.4 and 0.9.3. The problem with Skipstone crashing when flipping
tab-views is now fixed, however there are still some minor problems
with tabbed-windows view. GTK applications in Puppy can now have
anti-aliased fonts, courtesy of the libgdkxft package, and this is
applied on a per-application basis. Font anti-aliasing is applied to
Skipstone, Dillo (web browser), Amaya (HTML editor) and Ted (word
processor). Download: cd-puppy.iso (49.0MB).
Turbolinux 10 F
The English language edition of Turbolinux 10 F is now available: Based
on Turbolinux 10 Desktop, Turbolinux 10 F... is the latest desktop
operating solution from Turbolinux that brings Multimedia to your
fingertips. 10 F extends functionality to support a variety of
streaming video, audio and multimedia content delivered over broadband.
10F... Fun, Fast, Future, Freedom, Flexible. Please find your 'F'.
Turbolinux 10 F is the first Linux distributions that brings Windows
and Real streaming video, DVD movies and Apple iPod support to Linux,
legally. To find out more, please read the original press release, product information page and data sheet (in PDF format). Turbolinux 10 F is available for purchase online (US$69.00).
Damn Small Linux 0.7.2
Damn Small Linux 0.7.2 has been released. From the changelog: "Added
myDSLgui, a click-and-run system for extensions; added Lua scripting
language and Lua sockets; replaced Scite with Beaver; replaced nvi with
vim; changed user from damnsmall to dsl (also took 'damn' out of the
boot process); upgraded busybox; simplified filetool.lst usage - always
edit at home; updated Firefox Flash plugin in Firefox download script;
actually shrunk the ISO by nearly 1MB from 0.7.1." Download: dsl-0.7.2.iso (46.9MB).
redWall Firewall 0.5.4c
This is a bug fix release of redWall Firewall 0.5.4: "A
'major' bug in MySQL (related to the environment on the CD) has been
fixed in this release (again ;-) ... please upgrade any 0.5.4 release
prior to 0.5.4c if you need MySQL support! Changelog: downgraded MySQL
to Version 4.0.18; modified mkisofs options which should fix reported
problems with booting the CD on certain PCs." Read more on the distribution's news page. Download: redwall-0.5.4c.iso.tar.gz (148MB).
Linux 2004-r4 has been released. This is not to be confused with the
recent OnebaseGo 2.0 (Live CD) release. Two major new features are the
new Net-Installer and OLM 3. "Being a net-based
installer it provides high level of flexibility in selection of
packages and mode of installation. Also you directly install the latest
available packs from OL-apps." The full scoop on this release is available here.
It should be noted that the Net-Installer is only meant for new users
of Onebase. The new release is available for purchase or as a free download (106 MB).
OnebaseGo 2.0 (Live CD)
OnebaseGo (Live CD) version 2.0 has been released. "This
release comes with numerous package updates, improved EPS and Docking.
EPS - eXtended package store, a new feature which was introduced in 2.0
preview1 allows users to access additional software via this." The announcement for this release can be found here.
The developers encourage users to purchase the product from their store
in order to support the growth of this project, but it also available
as a free download (457 MB).
Development and unannounced releases
|Web Site News
Linux On The Road
Robert Storey fell ill for almost a full week, slowing things down at Distrowatch for several days. Nothing more will be said here about that hiatus, but geeks are advised to get more exercise. Pushing around a mouse while getting your brain fried by a CRT monitor (even though protected by an aluminium foil hat) is not healthiest pastime. If the foregoing describes your condition, now is a good time to wake up and smell the coffee (preferably decaffeinated).
Ladislav lent a (big) helping hand from an Internet cafe in Ferrara, Italy. Not surprisingly, all the machines in the Internet cafe were running Windows, but Ladislav managed to sneak in a Knoppix CD. The interesting thing is the staff didn't even notice the difference - perhaps we Linux enthusiasts can launch a stealth campaign, slipping in Knoppix disks where it's least expected. Anyway, Ladislav's next stop was Geneva, Switzerland, and plans call for hitting Austria next. We should be hearing interesting tidbits about Linux-on-the-road from Ladislav as he winds his way through Europe.
DistroWatch database summary
- Number of Linux distributions in the database: 315
- Number of BSD distributions in the database: 7
- Number of discontinued distributions: 32
- Number of distributions on the waiting list: 84
|Tips, Tricks and Hints
GRUB To The Rescue
I received a brief but urgent message from a friend (named Burt) a few days ago. And it was this message that inspired the main topic of this week's Distrowatch Weekly News.
Burt is new to Linux and admits to being bewildered at times, but he's very enthusiastic and eager to learn. He's been downloading and testing various distros, still looking for the "perfect" one. Over the past few months he's probably had more Linux installs than hot meals, and he keeps coming back for more. I admire his perseverance.
One minor hazard of doing so many installs is that your MBR (master boot record) gets repeatedly overwritten by the installation programs. Some distros install LILO, some install GRUB. If you mess with the BSDs, you'll wind up with something else, and Windows will overwrite your MBR without even asking. This is exactly what happened to my friend Burt. He installed Mepis, liked it, then tried something else that rendered GRUB inaccessible. Now he wanted to know if there was a way to get it back.
Fortunately, there is. What you really need (and should prepare in advance) is a GRUB boot disk. Normally, this would be a floppy, and if you search around the Internet you are sure to find a bootable floppy image somewhere. However, it's my opinion that floppies are on their way out, and indeed I have a couple of laptops that don't even have a floppy drive. Obviously, a bootable GRUB cdrom would be more useful. You could make one yourself, but easier would be to download one, and I happen to know that you can find one here.
If you boot this GRUB CD, you should wind up with a plain text mode prompt that looks like this:
Admittedly, this does not look very informative, but if you've gotten this far then you're already half way towards rescuing your system. All you've got to do now is know what to type at the prompt - the big question is how to find out?
What you should have done in advance was to look into the GRUB configuration file on your hard disk, which is /boot/grub/menu.lst (and note it's "l" not "1" - they really should have named this file "menu.list" rather than "menu.lst", but they didn't ask me). Anyway, look through this file (and perhaps print it out on paper). You'll be looking for a section that resembles this:
title Knoppix 3.3
kernel /boot/vmlinuz root=/dev/hda3 ro hdc=scsi hdd=scsi
What appears in your GRUB configuration file will no doubt be slightly different, but should be similiar. Now all you've got to do at the GRUB prompt is to manually type the above configuration information. However, you do NOT need to type the line that begins with "title", nor do you need to type "savedefault". You will need to type out the other lines. It is possible you won't have a line beginning with "initrd" (initial ram disk) - not every Linux distro makes use of this feature for booting. Assuming you typed everything correctly, after you type "boot" and hit enter, your machine should boot into Linux as it normally does.
Once you have booted and logged in as root, you can type this to restore your MBR:
That should do it. Note that the above will only work if you have "grub-install" already installed, which is usually the case. Otherwise, you'll need to download and install it.
The foregoing was only meant as a brief introduction in repairing your GRUB install. A more thorough treatment of GRUB can be found in this excellent article from Linux Journal.
That's all for this week.
2 • DW Weekly (by Life at 2004-07-19 03:29:43 GMT) |
A nice edition of DW weekly, kudos to Robert and Ladislav.
I have one point to add regarding the standardisation of Linux, as far as I am concerned, although there would be some perceived benefits, I think the Linux community should not be held back with constraints such as those proposed. Yes, it may make the operating system more likely to succeed on the mass desktop market, but at what cost?
I am not a Debian user, but I share the viewpoint given out by many of them, Linux is about freedom of software, and the varied nature of Linux installs also has the added benefit of promoting varied setups and for the most part, escapes the monoculture trap Windows has fallen into, this could be a major assest to Linux security concerns should it become a major player in the desktop market.
Without starting packagewarfare here, I really dislike RPM's from my experience with them under Mandrake and Fedora, they have left me with countless headaches in the past. I really cannot see a way that a standardised package format could be agreed upon between the major players, and I certainly don't believe it will come in the form of RPM's if it happens at all. Besides, variation with package management is a major factor that helps define one distribution's differences from another, the more you try to merge aspects such as this, the less reason there is to make new distributions and implement new and innovative ideas. As you mentioned, competetion is a good thing, when people stop competing with their products, they stop innovating, and that harms the consumer, the producer and the industry as a whole in my opinion.
3 • Life's comments (by eedok at 2004-07-19 06:51:48 GMT)
Becoming more standard creates more competition, not destroying it, because that way there'd be less distro-specific utilities, and more focus on the way the utilities are used, and it also creates less problems for people who make software without a specific distro in mind. Standardisation is a good thing.
4 • The number of different Linux Distros. (by mpwmedia at 2004-07-19 07:28:39 GMT)
The diversity of Linux is surely one of it's strengths. That way you can get a distro that will be good for the job you have in mind for it and your level of experience. Could it also be a weakness though that might hinder the breakthrough of Linux into the 'mainstream' as people not only have to choose between Linux and Windows but, assuming they choose the former, then have to choose between a bewildering array of versions?
5 • LSB & RPM (by Penguin on 2004-07-19 08:11:39 GMT)
It is quite obvious that RPM doesn't help standardizing Linux package management. It hasn't done it in the past and it will hardly do it in the future. LSB should drop the need to conform to one package type (RPM) from its guidelines, and concentrate on more relevant things.
Sometimes LSB also seems like it is too much promoting a few big distros only. That is hardly a way of getting the support of all the rest of the Linux world (nor a succesful attempt to throw the non-LSB-conforming competitors out of market).
LSB has potential though, of course. But I just hope that a few over ambitious and greedy marketers and businessmen (of certain big distributions) don't make its potential nonexistent in the long run?
6 • Standardization is good (by Steve at 2004-07-19 08:16:17 GMT)
Standardization is a very good thing. For example, just think if everyone thought their own HTML-like language was the best so everyone made their websites in non-standard languages... web browsers would either have to support many languages or you'd only be able to view a limited portion of the WWW. We don't have to deal with that because HTML is the standard. A standard base for all Linux distros is a good idea.
7 • bootable grub floppy (by mrbass at 2004-07-19 09:00:40 GMT)
best is GAG http://gag.sourceforge.net/....you can install it from the free http://www.ultimatebootcd.com/ makes it super simple to install and install your multiple OS's on.
8 • On standardization (by Avdibeg at 2004-07-19 09:16:58 GMT)
As many other readers, I agree that the Linux community desperately needs some kind of standardization. I mean, look at the Windows - one can say a lot about its security, stability, speed etc., but its one major strenghth is one standard, which helped to its great accessibility and usability.
Linux needs one standard - now!
9 • About bootloaders (by ageox at 2004-07-19 09:17:31 GMT)
XOSL is a graphical boot manager released by Geurt Vos under GPL. XOSL can dual boot over 30 various operating systems on a single disk.
It's not strictly a substitute to a bootloader, which is still required for each OS installation, but takes all the setup work for dual (or more) booting onto its shoulders in a very easy way. It's graphical, has mouse support and inbuilt setup. It also has quick and easy support for partition hiding (ever tried to boot Windows from hdb?)
I warmly recommend a look - link below.
10 • On standardization (by Avdibeg at 2004-07-19 09:17:58 GMT)
As many other readers, I agree that the Linux community desperately needs some kind of standardization. I mean, look at the Windows - one can say a lot about its security, stability, speed etc., but its one major strenghth is one standard, which helped to its great accessibility and usability.
Linux needs one standard - now!
11 • another freebsd (by sundar on 2004-07-19 10:25:17 GMT)
We have seen how Linux distributions competed each other and became more mature in few years.
I guess similar competition is needed in Freebsd to bringout the best. So far, as a newbie, who toys with different distros, I like Freebsd can be more newbie friendly. After seeing the ease of installation and custom packages in Linux distros, I guess it can be done in Freebsd also. I like to see bsd distro's coming out like Linux.
One important new Freebsd distro to watch is DragonFlyBsd. It strive for advanced features without compromise. Congradulations for those DragonFlyBSD developers!
12 • Distroitis (by Aussie on 2004-07-19 11:13:12 GMT)
As much as I don't view Distroitis as a bad thing as I, like most here, are OS whores who like to see whatever's new on the scene.
Not a good thing for the overall Linux community imho.
I'd prefer less Linux distro's and for the efforts to be put into the top 50 distro's - not spead insanly over 600 odd...
13 • Standardisation (by Uncle Fester at 2004-07-19 11:26:56 GMT)
Linux does not need (nor ever has needed) standardisation. While it's beneficial for different distros to be able to cross-pollinate, it's not a necessary criterion for success of the OS. What *is* needed is good documentation and information across the distros (big up Distrowatch for its role here) and continued development activity by users (hopefully this can be relied on for the foreseeable future - at least until a better solution than Linux comes along). One of Linux's greatest strengths lies in its ability to evolve by occasional forking and survival of the fittest. Standardisation is actually likely to weaken it in this area.
So long as there are interested hobbyist coders, Linux will continue to grow. For Linux to go mainstream, using Linux must be easy but that's more about catering to 'Joe Sixpack' than deciding on an arbitrary single standard. Windows scores highly not because of de facto standardisation but because it is designed to be usable by computer illiterates. Between that and some hardcore marketing, Windows has become the 'standard' OS for general desktop use.
There is a difference between an emergent standard and an arbitrarily imposed standard. As software matures, standardisation should emerge as the natural result of competition between forks. In any case, variety is important as it maintains flexibility. To impose a standard artificially risks forcing development to take the wrong branch of a fork and discourages developer participation. It could be the last thing Linux needs.
14 • Slack (by Gstar on 2004-07-19 11:30:31 GMT)
I tried SuSe 9.1 recently but I was disapointed to find out it has old version of gnome. I couldn't find any repository with 2.6.1 version so I gave up.
A few days later I noticed this nice Dropline Gnome system which is easy to install and update Gnome on lsackware. The only problem is that I'm sort of newbie and I don't want to wreck my bussines installation of Win xp.
Is there a distro based on slack with nice installer.
I really want to use nice Dropline Gnome
15 • Re: Slack (by Vectrox on 2004-07-19 12:27:22 GMT)
Slackware isnt difficult to install. You should try it first on VMware.
16 • Linux Standards Base (by Tim at 2004-07-19 12:40:01 GMT)
By going off on RPMs, the article missed the point of having standards for Linux. The Linux file system needs standardization so that developers can write programs that will work with all complying distros. How is a program to find a file it needs if every distro places the file in a different place? Anyway, RPM is just a SUGGESTION of the LSB for a package manager. Settle down people.
17 • why linux should be one... (by maceto on 2004-07-19 12:47:44 GMT)
security and not inventing the wheel 2-3 times, that is a bug on debian could be a bug on fedora etc.. a security patch can come out faster on one distro vs another... get the point..
18 • Many Distros.... (by Lord-Storm on 2004-07-19 12:56:42 GMT)
At least the kernal hasnt been fork-ed though some times I wish it did.. Personaly Linux is starting to become unfriendly with windows NTFS. Suse & ark Linux destroys my SP4 kernal. Shame...
Not many BSD's Out there but a few would be good. Too many is sometimes a bad thing when there are so many development distros Shouting for help. Aries linux formaly JAMD has grinded to a slow down due to developers leaving.
Patents have destroyed FC2 (Free and non Free dont play nice.)
Im sorry but im not going to re-rip my 2000 songs from CD all again.
19 • RPM as a standard (by koorek at 2004-07-19 13:13:15 GMT)
IMHO there's no difference in pakcage format (there're just archives with some extra info about package). The important thing is how these packages are managed. E.g. PLD Linux Distribution has outstanding poldek (there's also apt-rpm) so upgrade, installation with dependency check is normal way to manage my distro. Other thing is how these packages are prepared. PLD Linux (IMHO the best rpm based distro - but not for newbies) has really good set of packages. There're prepared much better than in Mandrake, Fedora or SuSE.
PS. Sorry if I made some mistakes, my english skill is not good.
20 • Easy to install slack with dropline (by Joel Ebel at 2004-07-19 14:27:51 GMT)
Gstar, I can't say I've tried it, but Vector linux is a slackware based distro with Dropline Gnome included. It might be worth a try if that's what you're looking for.
21 • Many Distros.... (by Lord-Storm on 2004-07-19 12:56:42 GMT) (by kmp on 2004-07-19 14:37:52 GMT)
Answer to your Fedore Core 2 mp3 problems:
22 • Re: GRUB To The Rescue (by Geoff on 2004-07-19 16:23:29 GMT)
Here's a simpler method to restore your MBR that has worked for me. Some distros (e.g., Redhat) have a rescue mode on the first install CD. Boot using the rescue mode or boot using a live CD such as Knoppix. Your main Linux installation may already be mounted (possibly under /mnt/sysimage for Redhat rescue mode). If not you will need to mount it. Then do:
23 • English, Slackware, and fonts (by prairiedock at 2004-07-19 18:10:34 GMT)
Robert, ouch! "Which begs the question, - does the world need another BSD?" should read: "Which raises the question ..."
"Begs the question" is properly used only in the context of discussing or criticizing an argument, and means "to (tacitly) assume the conclusion in one or more of the steps of the argument." This internet misusage is getting to be as bad as spelling "lose" as "loose."
As for Slackware (review of 7/19), I don't care to see any more reviews of Slackware as a desktop distro which do not address the difficulty of installing and getting working properly anti-aliased truetype fonts. IIRC, your own excellent review of Slackware 9.1 mentions but does not solve this problem. Good font rendering is essential, not optional, for anything which has pretensions to be a desktop distro.
If any wannabe reviewer is out there and reads this, do us all a favor. Install (if necessary) and launch Abiword, nedit, gvim, kate, thunderbird, firefox, Konqueror, gimp, ghostscript (among others). If all of these work properly with good anti-aliased truetype fonts, then write a how-to or a review which shows us how you did it. While you're at it, if you have a laptop, show us how you got your wireless card, dvd/cdrw drive, printer, and flashdrive working.
If someone solves all these problems, and it turns out not to be too complicated (e.g., you DONT have to recompile freetype.), then I'll be on board soon after. Reviews or articles which fail to address these questions are misleading and mostly a waste of the reader's time.
24 • BSDs (by anon at 2004-07-19 19:20:43 GMT)
I think it would be important to clarify a distinction when comparing Dragonfly vs. the BSDs and LinuxDistroX vs the Linux Distros in that there is a significant difference in the implemenation at the kernel level that the Dragonfly developers want to pursue. As far as Linux goes, *all* the distros use the same kernel, the difference is largely packaging (RPMs, apt ...) or some niche (Newbie distro, Live CD distro ...).
I don't see much "forking" of Linux at the kernel level. You can patch it with performance optimizations for certain applications, but I don't know of any living fork like Dragonfly.
It's also interesting to see a particular application of OS theory in production as opposed to merely reading about it in a textbook.
Others closer to this project may have some additions/corrections to my comments above.
25 • Don't know where else to put this (by Anonymous on 2004-07-19 20:29:13 GMT)
The link to the DeLi linux homepage seems to be dead.
26 • LSB et al (by Mike on 2004-07-19 23:00:33 GMT)
The FHS or Filesystem Hierarchy Standard is essential to the interoperation of POSIX systems and is widely accepted by distro and software developers alike.
There are two levels of software compatability: source and binary. POSIX, the single UNIX specification, FHS and so on provide source level compatability which makes it easy(ish) for programs written for i386 Linux to compile on *BSD, Mac OS X, Solaris and so on. Even on Windows, although Microsoft's lack of support for POSIX means Windows software is a bitch to compile on UNIX.
The LSB is about binary compatability - something that can be a little tricky. It means software written for i386 Fedora should run on i386 Debian, for example. The reason for specifying package management is so that software providers can simply provide LSB-RPMs for standard architectures, which could then be installed on any distro, with dependencies and unistalling dealt with.
In practice, no-one is providing LSB-RPMs. Instead, big projects like Mozilla and OpenOffice.org provide binary tarballs or installer programs that guide you through installation, much like Window's Install-Shield Wizard (that used to give me so much joy ). They avoid having many uncommon dependencies so they run out-of-the-box on most systems. The LSB does help, by specifying standard interfaces to shared libraries - potentially sparing us from the notourious problems of glibc version changes.
Personally, I think package management could be standardized but hasn't been yet because no-one has found the best solution.
As for those who say Linux shouldn't be standardized - it always was. How else do all these programs dance together without breaking faster than Windows 95? Until now we have had de-facto standards. Mostly, the LSB merely formalizes those. Where it goes beyong that remit, people just ignore it and do what they like. That's free software for you!
27 • DragonFlyBSD (by Haldir on 2004-07-19 23:15:13 GMT)
Pretty poor start. Their first release had a bug that allowed the installer to wipe out partitions on the hard drive that weren't supposed to be wiped.
28 • RE: why linux should be one... (by Peter Damoc at 2004-07-20 10:27:59 GMT)
"inventing the wheel 2-3 times" are you kidding? the wheel gets reinvented a lot more :) if it would only have been 2 or 3 times that would have been great....
In my view... people should start over :) and create a layered system for GUIs because GUIs are the problem.
a nice standard renderer,
a tookit with basic widgets,
a set of mega-widgets (think Gecko and friends)
a glue layer for putting the widgets together (python would be nice)
Apps created with sound MVC architecture....
This way an interface would become nothing more than a script (think small downloads) and updates to the core widgets would benefit every single app.
as for the package management... I would like to see people advocate the benefits of current package management systems over ZeroInstall.
29 • Re:System Commander (by fizzol on 2004-07-20 11:22:12 GMT)
>anybody remember System Commander which sold for US$99?
I still use System Commander. I find it very reliable and use it in preference to any other boot loader.
30 • Standardization is needed... (by CharlieJ at 2004-07-20 19:44:36 GMT)
As a long-time and experienced MS user who is trying to go to Linux, I agree with the standardization proponents.
I don't think anyone wants Linux to end up a Windows clone. However, standardization of core filenames, locations and installation components would be a great way to gain popularity amongst those who have only known Windows. It would lend familiarity to the process of switching to Linux -- and that would go along way towards gaining ground with "mainstream"
Right now, the options are TOO widely varied. For me, I've put off installing ANY distro of Linux because there are sooooo many to choose from -- I have no idea where to start. Yeah, I know -- pick one you like and use it. But when there are 50+ popular distros, there is no good way to choose ONE to try. I have no desire to install ten and then choose one either. The ONLY way I've narrowed the field of candidates is through forums and the info on Distrowatch.com.
Back to standards -- I truly believe standardization would help programmers realize their overlap -- thus resulting in a pooling of efforts, better distros and a smaller number to seriously consider -- which is a GOOD thing from where I sit. ;-)
31 • LSB & RPM - 2 (by Penguin on 2004-07-20 22:37:09 GMT)
Yeah, of course, Linux standards are a good thing and very much needed - especially when the main goal is what benefits users. Business goals in standardization are ok too, as long as they don't favor a few big players only, and don't restrict innovation, competition and development.
As to RPM, it is true that RPM is an ok package format in itself, but IMHO far from being the perfect Linux package standard for years to come. Also both Debian and Slackware package types are even older than RPM, AFAIK.
The LSB does also mention that a distribution "may use a different packaging format for its own packages". But - it is also mentioned (LSB 2.0) that "a future version of the LSB may require RPM".
I also doubt how credible the following LSB statement is: "Supplying an RPM format package is encouraged because it makes systems easier to manage. " For example, managing third party RPM packages in Redhat or SUSE, with all the potental dependency problems, can be a real pain. What does the LSB specifications REALLY say about solving the notorious RPM & dependency hell issues?
Ask yourself why LSB certification costs so much money? Ask yourself what distributions are LSB certified? Ask yourself who decides what the LSB specifications are?
LSB can and could do great things to promote Linux development - but unfortunatley I see some problems there too.
32 • RE:English, Slackware, and fonts (by Mephisto on 2004-07-21 00:11:55 GMT)
In reference to True Type Fonts add:
to your xorg.conf? Abiword, firefox, knoqueror, and gimp can all use TTF if they are installed and. If further tweaking is needed create a local.conf on /etc/fonts
I own a laptop (2 actually) Wireless works, DVD/CDRW works, usb flash drive works, printer works. for that matter 1394 works, as does CF and SD. The only things I had to do is for one of the laptops install the IPW2100 driver (no problems) and configure CUPS. I am not saying that Slackware is for everybody, but if you don't mind reading the documentation it is not exactly difficult.
If you had stuck to the TTF argument I might have agreed more or less but I had not problems whatsoever with everything else you mentioned.
33 • Progeny (by Mike on 2004-07-21 10:02:12 GMT)
Progeny Componentized Linux, which is based on Debian, has achieved LSB certification and they have provided patches to Debian stable and testing to make them LSB compliant.
So Debian can be LSB compliant. In fact there's only one file in Progeny's http://archive.progeny.com/progeny/lsb/patches/sarge/">patch directory for testing.
34 • Oops, sorry. (by Mike on 2004-07-21 10:03:19 GMT)
OK that didn't work :(
35 • I just checked the www.distrowatch.com for dns problems. (by Anonymous on 2004-07-21 16:39:16 GMT)
;; Now linting www.distrowatch.com from www.domtools.com http://www.domtools.com/dlint/nph-dlint.cgi?zone=www.distrowatch.com
;; Checking serial numbers per nameserver
;; 104196140 ns.owlriver.com.
;; 104196140 ns.herrold.com.
;; All nameservers agree on the serial number.
;; Now caching whole zone (this could take a minute)
;; trying nameserver ns.herrold.com.
ERROR: no A records found.
;; no subzones found below www.distrowatch.com., so no recursion will take place.
;; dlint of www.distrowatch.com. run ending with errors.
;; run ending: Wed Jul 21 09:14:40 MST 2004
======== www.dnsreport.com checking www.distrowatch.com
WARN SOA Serial Number WARNING: Your SOA serial number is: 104196140. That is OK, but the recommended format (per RFC1912 2.2) is YYYYMMDDnn, where 'nn' is the revision. For example, if you are making the 3rd change on 02 May 2000, you would use 2000050203. This number must be incremented every time you make a DNS change.
FAIL SOA REFRESH value WARNING: Your SOA REFRESH interval is : 86400 seconds. This seems very high. You should consider decreasing this value to about 3600-7200 seconds. RFC1912 2.2 recommends a value between 1200 to 43200 seconds (20 minutes to 12 hours, with the longer time periods used for very slow Internet connections), although some registrars may limit you to 10000 seconds or higher, and if you are using DNS NOTIFY the refresh value is not as important (RIPE recommend 86400 seconds if using DNS NOTIFY). This value determines how often secondary/slave nameservers check with the master for updates. A value that is too high will cause DNS changes to be in limbo for a long time.
36 • BSD from scratch (by wouter at 2004-07-21 23:53:11 GMT)
It is quite common for more advanced BSD users to make world, as it's called. I guess this is a bit comparable to Linux From Scratch, in that it potentially rebuilds large parts of the system. Most popular (commercial) Linux distributions are very binary, with complex dependencies, and rebuilding a system is not an easy task - if it's at all possible. It's difficult to customise the software, if needed.
If distributions like Gentoo (mind you, I've never tried that one myself yet) with similar build systems and ports like BSD has, would have been popular from the start, then we would probably have less linux distributions, too. But complex package systems and for some distributions 'commercial influences' (semi-open development process or code, no easy way to locate source and any specific patches) made tinkering with some linux distributions very hard.
37 • Reply to Slackware post (by wouter on 2004-07-21 23:59:32 GMT)
prairiedock: Slackware has always been something special, where you have to do some things yourself. Don't blame Slackware, blame yourself for making (apparently) a wrong choice. Different distro, different audience. There are many point-and-click distro's, and adding TTF fonts really isn't that hard.
Linux for everyone, but not every Linux is for everyone.
Number of Comments: 37
|• Issue 585 (2014-11-17): openSUSE 13.2, PC-BSD's "roles", MATE + Compiz on Mint, cleaning package cache|
|• Issue 584 (2014-11-10): OpenMandriva 2014.1, Debian freeze, trickle, systemd and boot times|
|• Issue 583 (2014-11-03): Ubuntu 14.10, ownCloud, Kylin interview, The Book of PF, Elive's commercial ways|
|• Issue 582 (2014-10-27): GhostBSD 4.0, Tumbleweed and Factory merge, systemd and fork of Debian|
|• Issue 581 (2014-10-20): SparkyLinux 3.5, Fedora's graphics stack, Debian and systemd, OpenBSD 5.6|
|• Issue 580 (2014-10-13): Rolling releases, Arch as best distro, GNOME on Wayland, MINIX 3.3.0|
|• Issue 579 (2014-10-06): PC-BSD 10.0.3, Debian's Jessie freeze, setting up home server|
|• Issue 578 (2014-09-29): Calculate 14, Debian's default desktop, Shellshock vulnerability, practical Tiny Core|
|• Issue 577 (2014-09-22): SymphonyOS 14.1, FreeBSD drops pkg_add, MINIX on ARM, GNU screen|
|• Issue 576 (2014-09-15): PCLinuxOS 2014.08, Mint's documentation, Debian's hardware database, CDE|
|• Issue 575 (2014-09-08): Porteus 3.0.1, Fedora's blivet-gui, Red Hat's Docker, systemd|
|• Issue 574 (2014-09-01): Ubuntu Kylin 14.04, Haiku and Linux kernel, Wayland support, Lumina, Bash completion|
|• Issue 573 (2014-08-25): SolydXK 201407, VPN gateway with FreeBSD, Ubuntu MATE, Raspbian, trusting binary packages|
|• Issue 572 (2014-08-18): ZFSguru 10.1, Fedora's Flock, beta installer for "Jessie", Ubuntu Core, rolling releases|
|• Issue 571 (2014-08-11): HandyLinux 1.6, LMDE update, default desktop in "Jessie", running out of disk space|
|• Issue 570 (2014-08-04): Neptune 4, Kubuntu's KDE Plasma 5, FreeBSD and UEFI, Linux servers|
|• Issue 569 (2014-07-28): Deepin 2014, Ask Fedora, Gentoo and LibreSSL, encrypted package downloads|
|• Issue 568 (2014-07-21): Antergos 2014.06.24, Mint based on Debian stable, upgrading CentOS, BinaryTides|
|• Issue 567 (2014-07-14): Manjaro 0.8.10, PC-BSD jails, Debian and glibc, Fedora's DNF, Xiki and Opera 24|
|• Issue 566 (2014-07-07): LXLE 14.04, OpenBSD's SimpleDE, openSUSE artwork, home security basics|
|• Issue 565 (2014-06-30): Chakra 2014.05, Fedora on BeagleBone, Matthew Miller interview, e-book readers|
|• Issue 564 (2014-06-23): Antergos 2014.05.26 and Q4OS 0.5.11, Debian LTS and glibc, Fedora DNF|
|• Issue 563 (2014-06-16): Mint 17, CentOS 7 pre-release, Debian MATE, accessing encrypted content|
|• Issue 562 (2014-06-09): GoboLinux 015, Gentoo interview, Fedora leader change, climagic tricks|
|• Issue 561 (2014-06-02): OpenMandriva 2014.0, Debian GNU/Hurd, Lubuntu and LXQt, Final Term, TrueCrypt|
|• Issue 560 (2014-05-26): KaOS 2014.04, Wayland and KDE 5 on Fedora, distros with commercial support, DenyHosts|
|• Issue 559 (2014-05-19): VortexBox 2.3, LTS-only Linux Mint, FreeBSD 11 ambitions, KDE 5 beta|
|• Issue 558 (2014-05-12): RHEL 7 Workstation impressions, LXQt and Lumina, Haiku interview|
|• Issue 557 (2014-05-05): Xubuntu 14.04, Ubuntu 14.10 roadmap, Fedora Workstation, ownCloud|
|• Issue 556 (2014-04-28): Ubuntu 14.04, LibreSSL, Lumina desktop, Deepin interview|
|• Issue 555 (2014-04-21): Robolinux 7.4.2, Ubuntu release day stats, Debian security, Porteus update|
|• Issue 554 (2014-04-14): Review of FreeNAS, OpenSSL bug, Fedora.next, Robolinux Stealth VM, measuring memory|
|• Issue 553 (2014-04-07): Puppy 5.7 "Slacko", end of Ubuntu One, file encryption with GPG|
|• Issue 552 (2014-03-31): Tanglu 1.0, Ubuntu GNOME LTS, SliTaz for ARM|
|• Issue 551 (2014-03-24): Linux Mint "Debian" 201403, call for end to proprietary firmware, LVM|
|• Issue 550 (2014-03-17): Review of NixOS 13.10, Lubuntu seeking feedback, Android-x86 4.4-rc1 impressions|
|• Issue 549 (2014-03-10): ClearOS 6.5 and UCS 3.2, Gentoo interview, Ubuntu app contest, Into the Core|
|• Issue 548 (2014-03-03): Review of Mageia 4, FreeBSD console driver, filtering web content, Pitivi fundraiser|
|• Issue 547 (2014-02-24): Chakra 2014.02, Ubuntu privacy, preventing unwanted remote logins|
|• Issue 546 (2014-02-17): Review of PC-BSD 10.0, Red Flag closure, Ubuntu and systemd, SlackE18, Fedora book review|
|• Issue 545 (2014-02-10): Impressions of FreeBSD 10.0, Debian votes systemd, Ubuntu file manager, server security|
|• Issue 544 (2014-02-03): Netrunner 13.12, openSUSE future, Ubuntu Touch in emulator, running commands in multiple places|
|• Issue 543 (2014-01-27): Review of Korora 20, FreeBSD 10.0, DNF, ZFS rescue CD, Bridge Linux interview|
|• Issue 542 (2014-01-20): QupZilla, Ubuntu with MATE, Arch on Raspberry Pi, best applications|
|• Issue 541 (2014-01-13): openSUSE 13.1 and Zentyal 3.3, CentOS joins Red Hat, Bodhi on Chromebooks|
|• Issue 540 (2014-01-06): SMS 2.0.6 and SME Server 8.0, Hawaii desktop, PHR statistics 2013, more on multi-part archives|
|• Issue 539 (2013-12-23): Centrych 12.04.3, Fedora 20 and its spins, dividing archives across multiple discs|
|• Issue 538 (2013-12-16): Mint 16 review, RHEL and CentOS 7 plans, SteamOS, Windows XP replacement suggestions|
|• Issue 537 (2013-12-09): OpenMandriva 2013.0, Gentoo developer interview, project Neon, Linux Mint and security|
|• Issue 536 (2013-12-02): Impressions of openSUSE 13.1, Ubuntu Touch, FreeBSD 10 delay, troubleshooting OS lock-ups|
|• Full list of all issues|