‘System administration command. This simple backup utility accesses ext2 and ext3 file devices directly, quickly backing up files without affecting file access times. files may be specified as a mount point or as a list of files and directories to back up. While you can use this on a mounted system, dump may write corrupted information to the backup when the kernel has written only part of its cached information. Dump maintains a record of which files it has saved in /etc/dumpdates, and will perform incremental backups after creating an initial full backup. Use the restore command to restore a dump backup.
Options
-a

Write until end-of-media. Default behavior when writing to tape drives.

-A file

Create a table of contents for the archive in the specified file.

-bblocksize

Block size in kilobytes to use in dumped records. By default, it is 10, or 32 when dumping to a tape with a density greater than 6250BPI.

-Bblocks

Specify number of blocks to write per volume.

-c

Treat target as a 1700-foot-long cartridge tape drive with 8000 bpi. Override end-of-media detection.

-d density

Specify tape density.

-D file

Write dump information to file instead of /etc/dumpdates.

-E file

Exclude inodes specified in file.

-f files

Write backup volumes to the specified files or devices. Use – to write to standard output. Separate multiple files with a comma. Use host:file or user@host:file to write to a networked host using either the rmt program or the program specified by the RMT environment variable.

-F script

Run script at the end of each volume other than the last. dump will pass the current device and volume number to the script. The script should return 0 to continue, 1 to prompt for a new tape, or any other exit value to abort the dump. The script will run with the processes real user and group ID.

-i inodes

Specify a comma-separated list of inodes to skip.

-I n

Ignore the first n read errors. dump ignores 32 read errors by default. Specify 0 to ignore all errors. You may need to do this when dumping a mounted filesystem.

-j[level]

Compress each block using the bzlib library at the specified compression level. By default dump uses level 2 compression.

-k

Use Kerberos authentication when writing to a remote system.

-L label

Write the specified volume label into the dump header.

-m

Save only metadata when backing up changed but not modified files.

-M

Create a multivolume backup. Treat any filename provided with -f as a prefix.

-n

Use wall to notify members of group operator when prompting for information.

-q

Abort the backup instead of prompting for information when operator input is required.

-Q file

Create Quick Access information in the specified file for use by restore.

-s n

Write only n feet of tape in a single volume. Prompt for a new tape upon reaching this limit.

-S

Calculate and print the amount of space required to perform the backup, then exit.

-T date’

‘Last year Vector Linux came in four flavors. The list has now been expanded to seven different variations on the distribution: Standard, Deluxe, SOHO, Live, Light, Mini, and Light Live. SOHO, with KDE as the default desktop and all the most popular applications, is the full featured version. Standard is based on the Xfce4 desktop and provides superior speed and performance. Both are freely downloadable. Deluxe, available for purchase, is Standard plus a second CD with additional applications including KDE and OpenOffice. Live, as the name implies, is a live CD version of Standard. Light is a paired down, extremely lightweight version designed to run on older systems with as little as 64MB of RAM. In reality it will run with less than that. Light is built around either a JWM or Fluxbox desktop and lightweight applications. Mini is a further reduced version of Light that fits on a 5cm/3″ mini CD and requires only 1.1GB of disk space. Finally, Light Live is, as you’d expect, the live CD version of Light. So far only new Standard and Deluxe versions have been released but the others, all in various stages of development and testing, can already be sampled. This review will stick strictly with the Standard version from here on out.

My main box for testing Vector Linux 5.9 Standard my aging general purpose laptop, a five year old Toshiba Satellite 1805-S204 (1 GHz Celeron processor, 512MB RAM). Though this system has adequate memory for any current Linux distro it’s sufficiently challenged in terms of processing power for KDE to be sluggish and for Gnome to be noticeably slower than Xfce4 in most distros.

Vector Linux is almost to the point where it can seriously considered by almost any user, not just someone experienced with Linux, as most things do work as they should out of the virtual box. Some issues still require manually editing configuration files. I had hoped that by this point VL would be as user friendly as any distro out there but it isn’t there yet.

Installation and Configuration

Vector Linux 5.9 Standard is available for download as a single iso image. I’ve used both supported installation methods: a conventional installation booted from CD-ROM and a hosted installation booted from another Linux distribution already running on the system. There are two scripts provided for hosted installs: one which runs from an iso image on a mounted filesystem and one which runs from a CD-ROM that isn’t bootable. Directions and all the tools needed for a hosted installation from DOS or Windows are also provided. Installation across a network and automated installations, such as Red Hat’s kickstart, are not supported.

Vector Linux still uses an old fashioned text-based installer similar to Slackware, which is fine as far as I am concerned. A new, snazzy graphical installer is still promised for Vector Linux 6. I didn’t test to see if the installer now checks for adequate disk space, a problem with the VL 5.8 Standard installer. You need at least 3.2GB for a reasonably complete installation.

Disk partitioning is handled smoothly in VL 5.9. ext3, reiserfs, jfs, xfs, and trusty old ext2 are all supported. A list of common mount points is provided but the ability to define a non-standard mount point for a partition is lacking. Much like Slackware, Vector Linux still uses lilo and only lilo for the bootloader. Grub is available in the repository and can be installed after installation.

Video hardware detection remains relatively poor compared to other distributions and has not improved since 5.8. For example when I booted up to the GUI I was once again left with a small display in the middle of my screen surrounded by lots of black space. No amount of fiddling with X configuration in vasm (the GUI configuration tool) can make it work the way it should. It turns out the Monitor section of the /etc/X11/xorg.conf file doesn’t have proper horizontal and vertical refresh rates set and you are warned about this possibility by the installer. In addition, some options required by my laptop screen aren’t set. This problem has shown up for other users of Toshiba laptops with similar chipsets as well. Most major distributions simply don’t have problems like this in 2008.

In addition the installer doesn’t setup the system to load the kernel module needed to fully support my laptop at boot. I had to manually add:

modprobe toshiba

to my /etc/rc.d/rc.modules file to correct this. On a plain vanilla desktop system this wouldn’t be an issue, of course, but I suspect other laptop users, not just those of us with Toshiba machines, will need to do some tweaking by hand to get their laptops to be 100% functional under Vector Linux.

Printing and wireless networking aren’t handled by the installer and have to be configured after the system boots up for the first time. Expect to go into vasm or vasmCC to choose what services to start at boot as well.

One unique option in the Vector Linux installer is the ability to chose between HAL and VL-Hot for managing removable media. HAL, used by most distributions, continually polls the hardware and can have a performance impact on slower machines. VL-Hot is triggered by udev events and offers faster performance. It also uses rather long mount point paths which can be annoying if you work at the command line a lot. VL-Hot uses a second desktop icon for unmounting removable media which might be confusing to some. This is typical of many of the innovations found in Vector Linux: power users will love the flexibility and newcomers will wonder what it all means. I think most newcomers to Linux will find HAL easier to deal with.

Seamonkey is the default browser in Vector Linux. Alternate browsers (Firefox, Opera, and Dillo) are offered as options during installation. OpenOffice is not included in the iso but it is available from the repository after installation. AbiWord and Gnumeric are installed by default and Scribus, a desktop publisher, is an installer option. GIMP is also no longer included on the iso but rather is part of the new Multimedia Bonus Disc.

The installer and the documentation are in English, period. No other languages are supported despite huge progress in other areas of internationalization and localization.

Changes Since Vector Linux 5.8

There have been a huge number of changes in the latest release, both in applications and in configuration tools. Xfce 4.4.2 is the default desktop but VL has returned to the practice of offering alternative, lightweight desktops during the initial installation. This was dropped in 5.8. The alternate choices are JWM 2.0.1 and Fluxbox 1.0.0. KDE 3.5.9 is available in the repository as are a nice selection of lightweight window managers. If you choose Xfce you’ll find a much larger selection of panel applets available compared to 5.8. It’s still not the full selection from Xfce Goodies but I suspect most users will find what they want. Thunar is now the only fine manager installed by default but a number of others, including Xfe and pcmanfm, are available in the repository.

The biggest steps forward are the tools to manage and configure the system. vasm, the graphical system configuration tool, still isn’t the prettiest out there but it has a lot of functionality and works flawlessly. It’s now joined by vasmCC. The new Control Center is a fully graphical, pretty, and very functional tool for system configuration with most but not quite all the functionality of vasm. It’s reasonably easy to use and generally quite intuitive. OK, it would be more obvious for newcomers if an icon in the Network section was labeled “Network Configuration” instead of “Netconf”, but that’s a minor quibble. Once you get to below the second level of icons things look suspiciously like vasm but that’s to be expected and everything still works flawlessly.

WiFi-Radar now allows changing configuration settings within the GUI with a new preferences tab. You no longer a need to manually edit a configuration file to setup a wireless network interface. In addition, the semi-graphical alternative, vwifi, also works flawlessly. gslapt, the graphical package manager, now correctly displays dependencies for most packages. All in all configuring Vector Linux is now easier than in any previous release in most cases.

Under the hood Vector Linux now runs on a 2.6.22.14 kernel. An updated 2.6.24.3 kernel is currently in testing. It’s worked flawlessly for me so far but I did have to compile an updated madwifi driver to work with it. That should be included in the final package. Users who stick with the default kernel are well advised to install the novmsplice module to close a significant security vulnerability which came to light after VL 5.9 was released.

Kernel level support for popular laptops works as expected. Tools for my laptop, such as toshset, are in the repository.

Multimedia applications including XMMS, Xine, and optionally MPlayer have all been updated. Canada, the home of Vector Linux, has no equivalent to the DMCA so all the libraries and Win32 codecs are installed by default. Vector Linux is ready to play your mp3 files and DVDs right out of the (virtual) box. People in the United States will need to remove some packages after installation in order to comply with the law.

k3b, the KDE tool for burning CDs and DVDs, is the nicest of its kind for any platform and is now available as an option during install. KDE does not have to be installed to use k3b in VL, but a number of required KDE libraries are automatically included if you choose this option. The alternative is xfburn, a rather simplistic and limited CD burner designed to integrate with Xfce. Graveman, which I found to be horribly buggy, is no longer included.

The number of games included in the iso image have been greatly reduced. The repository does contain a large selection of games that more than makes up for this.

One of the most interesting new tools included in Vector Linux 5.9 is vpackager, a remarkably easy to use graphical tool to build packages from source. I suspect a lot of relatively new Linux users who looked at compiling software from source code as a daunting task will be surprised at just how easy it can be with vpackager. vpackager is also designed to work with CruxPorts4Slax but unfortunately that functionality isn’t quite ready from prime time just yet. When everything works properly there will be a truly easy way to install software directly from source. A How-To for wriiting ports for CruxPorts4Slack is included in the Vector Linux Wiki.

Running Vector Linux 5.9 Standard

Unlike Ubuntu (or Xubuntu or Kubuntu) the VL installer doesn’t leave you with a stripped down system to build on. It takes the approach used by larger distributions (i.e.: Fedora, Suse, Mandriva) and gives you a system with a fairly comprehensive set of applications immediately available after installation and configuration.

One of the main reasons I keep coming back to Vector Linux after trying other distributions is because it clearly is the fastest distribution with a decent set of applications and features for my older hardware Once you get past installation and configuration I still give Vector Linux a slight edge over most other distributions in terms of a user friendly, well thought out desktop.

I did run into a few bugs. As I previously reported Vetcor Linux 5.9 installs a full set of fonts for X by default but only enables the TrueType and OpenType fonts. Installing a package called fontfix, currently in the testing repository, resolved this problem.

Another annoying but minor issue is a display problem in AbiWord. Vector Linux developers chose to go with the latest bleeding edge development version, 2.5.2, rather than the latest stable release. 2.5.2 generally works but I often get a floating part of the toolbar about halfway down my screen. It’s always on top and doesn’t go away when I change instances of AbiWord or even switch to another application window. It only disappears when I close the instance of AbiWord that caused the problem. I’ve also had some interesting (OK, weird) fonts and spacing magically appear in documents.

The included version of Opera is also bleeding edge, a preview release of 9.50. It mostly works well but I’ve run into some issues with Javascript and a few web forms.

The Vector Linux 5.9 repositories include far more packages than ever before. Unfortunately there is currently a large backlog of packages in the testing repository that haven’t made it to extra just yet. Users may need to enable the testing repository to find applications they want and then disable it again before doing a system upgrade. While the number of packaged applications for Vector Linux continues to grow rapidly it still falls far short of what is available for distributions like Debian, Mandriva, Ubuntu, or Fedora. VL 5.9 is based on Slackware 12 and any Slackware packages, including those from third party sources like linuxpackages.net or slacky.eu, should work on Vector with one caveat: dependency checking may not work correctly. It’s pretty easy to end up in dependency hell if you enable third party repositories.

Multimedia and graphics software, as well as software of interest to musicians, are not going to be an issue. The selection for Vector Linux is now truly up to par thanks to the new Multimedia Bonus Disc, which deserves a review of its own.

Security Concerns and Package Management

In my review of Vector Linux 5.8 Standard I devoted a lot of space to package management, specifically the fact that for the first time in a distribution derivative of Slackware packages could be handled much the same way as they can in Debian based distributions using slapt-get and gslapt. gslapt is very similar in look and feel to synaptic and works in much the same way. slapt-get is similar to apt-get though the command syntax is different. Other distributions, i.e.: Wolvix 1.1.0, have followed suit.

The good news is that throughout the lifecycle of Vector Linux 5.8 and now 5.9 doing a:

slapt-get –update
slapt-get –upgrade

does precisely what you’d expect it to: cleanly upgrade your system by finding all the latest patches and updates and installing them. The bad news is that in Vector Linux 5.9 this functionality has been disabled from gslapt, the graphical alternative. Writing in the Vector Linux forum, one user who calls herself GrannyGeek put it this way:

I’m not glad at all that the Mark All Upgrades and View Marked have been disabled. This is nannyism at its worst. I *like* to be able to see what upgrades are available in a way that’s less laborious than going through the whole lengthy list of packages. Marking upgrades and viewing what’s marked was a convenient way to do this.

Disabling that feature was a really bad idea because it makes keeping a system secure difficult for a Linux newcomer. There is now no automated method within the GUI to determine what patches are out there and install them.

It’s worse than that. While the Vector Linux developers have been lightning fast at getting new packages out to close security vulnerabilities since the release of 5.9, there is no set method to inform users that a patch is available or even necessary. A security section exists within the forum but it isn’t used. This wasn’t a big issue in VL 5.8 because one quick check of gslapt would show you what’s out there. That simply isn’t true any longer for users who aren’t knowledgeable and comfortable on the command line.

Internationalization and Localization

Support for languages other than English has always been an area where Vector Linux was weak compared to other distributions. There has been tremendous improvement in VL 5.9 in terms of providing the tools necessary for supporting a large variety of languages. What is still lacking are graphical configuration tools to allow a newcomer to change language, locale, and keyboard settings for the system default, individual users, or on a session by session bases.

A wide variety of international font sets for languages that use non-Latin character sets are now installed by default. One one hand I’m not sure how many Vector Linux users are likely to need the Syriac or Ethiopic fonts which are installed. On the other hand Vector Linux is probably the only distro where probably every web page in the world will likely be displayed correctly out of the box. Whether a user can read what’s displayed is another matter. I should note that the full set of Cyrillic fonts for X is installed by default but not enabled. The fontfix package I referred to earlier does remedy this. In addition the selection of fonts for a given set of non-Latin glyphs is generally small. Additional font packages are beginning to show up in the repository.

KDE i18n packages are in the main repository. fribidi, the software at the core of support for languages written right to left such as Arabic, Hebrew, and Thai, is installed by default. The Xfce4 keyboard switching applet is also included. That applet doesn’t alter the /etc/X11/xorg.conf file, though, so if you don’t have KDE installed you still have to manually edit that file to add support for additional keyboard layouts other than your chosen default.

A full set of aspell dictionaries are now in the repository as are language packs for Mozilla Firefox, Thunderbird, and Sunbird. This is a huge step forward. Strangely langauge packs for Seamonkey, the default browser in VL, aren’t there yet. Open Office internationalization packages and dictionaries are also still missing.

If your goal is a truly localized system, not just a system running in English with support for other languages, you still need to do significant work to make it happen in Vector Linux Standard. Most all packages are now built with whatever translations are available. However, the default display manager is a rebuilt, trimmed down version of kdm. As previously noted language switching at login is not supported. If you want to change the default language or locale you have to do it at the command line and/or by editing appropriate configuration files.

Conclusions

Many of the issues I complained about in my previous review of Vector Linux 5.8 were fixed in 5.9. The installer still has some issues with hardware detection on my laptop. I had to do only minor configuration tweaking after installation and I was up and running. Someone with well supported desktop hardware might find the Vector Linux “just works” for them. Internationalization and localization are vastly improved in VL 5.9 but there still is a long way to go to make it user friendly and some packages are still lacking.

Once you have Vector Linux installed, configured, and customized to your needs it’s an absolute pleasure to run. It’s fast and sleek and well thought out. Experienced Linux users and those willing to roll up their sleeves and learn will likely find Vector Linux to be one of the best desktop distributions out there. Newcomers and anyone else who just wants things to work after a simple installation may still find that getting Vector Linux 5.9 is still an exercise in frustration if there is anything that isn’t vanilla in their system. Those whose primary language is something other than English and aren’t bothered by the lack of GUI tools may find that Vector Linux 5.9 finally meets there needs. Newcomers to Linux and anyone uncomfortable with manually editing configuration files will find that getting VL to work for them in their own language is more than they bargained for.

Vector Linux 5.9 is their best release yet. It just isn’t for everyone. Vector Linux was about 90% of the way to being about the best distro out there a year ago. While tremendous progress has been made in some areas in other areas, like keeping a system upgraded and secure, VL actually took a step backwards with this release. With a little more attention to the needs of newcomers Vector Linux could become the next big Linux distribution.

I have yet to find a perfect Linux distribution and I have never written a review where I didn’t find something to complain about. Even with the issues I’ve raised Vector Linux is still my favorite distribution for the desktop or laptop.’

‘A few weeks ago a new version of RSSTail came out. After reading the RSSTail page I found out that you can use RSSTail as a plugin to a fantastic application called multitaill. That is when I realized that multitail is actually written by the same person.

If you rely on tail, or have ever used tail you’ll probably enjoy this one. As the name implies it is a pager like tail except that it supports for more than one stream. Now, I know some of you have probably used some wizardry do get your favorite pager to show multiple streams, either in the same output or by using screen / splitvt but multitail really simplifies the process. It has all kinds of settings (such as color) and allows you to manage your tail’d files on the fly and even save the current view for later use. It does have a little bit of a learning curve but I think it is well worth it, especially if one of your hobbies includes log watching.’

Run Levels in a Nutshell
Filed under: Quick Linux Tutorials — TheLinuxBlog.com at 9:03 am on Wednesday, March 11, 2009

Run levels in Linux are a great thing. Basically, a run level is by definition a configuration for a group of processes. The run levels and default run level is specified in /etc/inittab. Most Linux systems these days, with exception of a few boot into run level 5 which is generally a graphical user interface such as KDM or GDM. The others boot into run level 3 most servers will boot into this run level which is multi-user with networking but no X, and is many users preference.

To define what run level your system boots into by default you would edit the /etc/inittab file and edit the line similar to:

id:5:initdefault:

This is run level 5, if you wanted to switch to command line you’d change the 5 to 3 and vice versa.

If your not ready to make the jump yet but would like to check it out, you can (as root) use the command telinit to tell init to change run level. If you are in run level 5, try (be prepared to lose everything in X, as it will kill everything for you)

telinit 3

If you are doing maintenance, you may want to switch to level 1 which is single user mode. Level 2 on Fedora is the same as 3 except it doesn’t have NFS support.

Level 0 is halt and run level 6 is reboot which are the best ones to accidentally set as a default run level (trust me on this one.) For more information on the different run levels check out the man pages

“This is a small distribution designed for i386, derived from SuSE 6.4. The base version contains only that needed to run Linux. The package is small and requires approx 7MB. The purpose of this TinyLinux is to have a small and clear system that can be used as starting point for further developments. Its use is not limited to desktop systems. It can be suitable for embedded systems. All binaries, directories and configuration files are taken from SuSE”

“You got a new hard drive and attach it to your laptop/desktop but before you can use it, you need to partition and format the hard drive. This can be done with a series of easy steps. This is a small and easy tutorial which will show you how to use fdisk, one of very famous utility for disk partitioning.

Fdisk is the utility which we will be using for this example. Though there are other options also like parted/gparted and use of them depends totally on personal preference. For Fdisk, you need to provide the device name as an argument. For that the possible options could be:

/dev/hda
/dev/hdb
/dev/sda
/dev/sdb

Where

/dev/hd[a-h] stands for IDE disks
/dev/sd[a-p] stands for SCSI disks
/dev/ed[a-d] stands for ESDI disks
/dev/xd[ab] stands for XT disks.

Hard disks can be divided into one or more logical disks called partitions. The partition info is stored in the partition table which is present in the first 512 bytes of the disk.

By default, a hard drive can be divided into 4 partitions but you can have much larger number of logical partitions by dividing one of the primary partition. And only one of the partition can be sub-divided into logical partitions.” geekride.com

“The Linux Standard Base, or LSB, is a joint project by several Linux distributions under the organizational structure of the Linux Foundation (merger of the Free Standards Group and the Open Source Development Labs) to standardize the software system structure, including the filesystem hierarchy, used with Linux operating system. The LSB is based on the POSIX specification, the Single UNIX Specification, and several other open standards, but extends them in certain areas.

According to the LSB:

The goal of the LSB is to develop and promote a set of open standards that will increase compatibility among Linux distributions and enable software applications to run on any compliant system even in binary form. In addition, the LSB will help coordinate efforts to recruit software vendors to port and write products for Linux Operating System.”

“A Linux distribution (also called GNU/Linux distribution by some vendors and users) is a member of the family of Unix-like software distributions built on top of the Linux kernel. Such distributions (often called distros for short) consist of a large collection of software applications such as word processors, spreadsheets, media players and database applications. The operating system will consist of the Linux kernel and, usually, a set of libraries and utilities from the GNU project, with graphics support from the X Window System. Distributions optimized for size may not contain X, and tend to use more compact alternatives to the GNU utilities such as Busybox, uClibc or dietlibc. There are currently over six hundred Linux distributions. Over three hundred of those are in active development, constantly being revised and improved. Linux distributions are software development systems that are constantly revised and improved.”

“Some concern has arisen about the suitability of object-oriented classes in LGPL’d software being inherited by non-(L)GPL code. Generally, these concerns are unfounded, and clarification is given on the official GNU website:

The LGPL contains no special provisions for inheritance, because none are needed. Inheritance creates derivative works in the same way as traditional linking, and the LGPL permits this type of derivative work in the same way as it permits ordinary function calls”

“The main difference between the GPL and the LGPL is that the latter can be linked to (in the case of a library, ‘used by’) a non-(L)GPLed program, regardless of whether it is free software or proprietary software.[1] This non-(L)GPLed program can then be distributed under any chosen terms if it is not a derivative work. If it is a derivative work, then the terms must allow “modification for the customer’s own use and reverse engineering for debugging such modifications.” Whether a work that uses an LGPL program is a derivative work or not is a legal issue. A standalone executable that dynamically links to a library is generally accepted as not being a derivative work (in LGPL). It would be considered a “work that uses the library” and paragraph 5 of the LGPL applies.

A program that contains no derivative of any portion of the Library, but is designed to work with the Library by being compiled or linked with it, is called a “work that uses the Library”. Such a work, in isolation, is not a derivative work of the Library, and therefore falls outside the scope of this License.”

“The GNU Lesser General Public License (formerly the GNU Library General Public License) or LGPL is a free software license published by the Free Software Foundation (FSF). It was designed as a compromise between the strong-copyleft GNU General Public License or GPL and permissive licenses such as the BSD licenses and the MIT License. The GNU Library General Public License (as the LGPL was originally named) was published in 1991, and was the version number 2 for parity with GPL version 2. The LGPL was revised in minor ways in the 2.1 point release, published in 1999, when it was renamed the GNU Lesser General Public License. Version 3 of the LGPL was published in 2007 as a list of additional permissions applied to GPL version 3.”

“The GNU General Public License (GNU GPL or simply GPL) is a widely used free software license, originally written by Richard Stallman for the GNU project.

The GPL is an example of a powerful copyleft license that requires derived works to be available under the same copyleft. Under this philosophy, the GPL grants the recipients of a computer program the rights of the free software definition and uses copyleft to ensure the freedoms are preserved, even when the work is changed or added to. This is in distinction to permissive free software licenses, of which the BSD licenses are the standard examples.”

“Linus Torvalds, a computer science student at the University of Helsinki in his early twenties, took his first course on Unix and C in the Fall of 1990. In the Spring of 1991 Linus was running Minix (a small Unix-like operating system designed for teaching) at home on his new 386. What was to become Linux started in the Summer of 1991 as a basic protected mode system that evolved from a Hello World program into a terminal program. By October 1991 Linux 0.02 was announced to the world. In two years, through the hard work of Linus and many other people, Linux, currently at version 0.99, has become an extremely useful and popular operating system. The comp.os.linux.* hierarchy is among USENETs busiest and there are several companies selling Linux and providing professional support. All this in such a short time, yet Linux is available for free, and development has almost entirely been done by volunteers. Meta interviewed Linus via E-mail to probe his mind about Linuxs future and the environment it is developed in. The results follow…

Meta Magazine (Mike Linksvayer): Do you agree that without the net to facilitate collaboration and a base of preexisting free software (e.g., the GNU tools), Linux would not be nearly as developed as it is?

Linus Torvalds: No question about it. Without net access, the project would never have even gotten off the ground; having access to gcc and the other GNU tools was very important. I was also able to get in contact with some people like Bruce Evans (author of the Minix-386 patches and the 16-bit assembler that is still used to assemble the Linux 16-bit startup code), and we had some interesting discussions by E-mail. Aside from getting me started, net access also kept the development going and accelerating: up to about version 0.12 or so, I wrote most of the code myself, but in the current kernel, only about 50% of the code is mine or very closely related to code written by me. The SCSI drivers, the networking code and the new floating-point emulator code is completely written by others.

Even when people havent sent in patches or new code, just the fact that Ive had access to a lot of testers has meant a lot for Linux development; theyve found bugs I wouldnt have noticed myself, and have suggested features that I might not have otherwise cared for, but that have turned out to be very useful indeed. One extreme example is the memory manager: I originally implemented demand-loading and swapping to disk mostly because people who used the early versions of Linux thought it might be useful.

Meta: Probably the most frequent complaint about Linux is the lack of certain applications. With the net, free development tools and a free Unix with source code all in place, what do you think are the prospects for free end-user applications being developed in a similar decentralized manner?

Linus: While Linux has a very reasonable development environment and a lot of programmers that would potentially be able to write a good word processor or spreadsheet or whatever, there are some problems which make me doubtful that it will happen soon. Right now, I think there is a better chance of getting a word processor by being binary compatible with Windows or some real 386 Unix (both of which are being worked on). The programs that have made it through a decentralized network development have usually had a few things in common:

(a) Somebody (usually one person) wrotevery similar, and may one day merge, the basic program to the state where it was already usable. The net community then takes over and refines and fixes problems, resulting in a much better program than the original, but the important part is to get it started (and channeling the development some way). The net works a bit like a committee: youll need a few dedicated persons who do most of the stuff or nothing will get done.

(b) You need to have a project that many programmers feel is interesting: this does not seem to be the case with a lot of the application programs. A program like a word processor has no glamour: it may be the program that most users would want to see, and most programmers would agree that its not a simple thing to write, but I also think they find it a bit boring.

I think its entirely possible that the Linux community (or some other group of net.persons) will get a good word processor going, but while having net access helps some parts of development a lot, its certainly not enough in itself.

Meta: Could you comment on the effort to make Linux binary compatible with real Unixes and speculate on the effect Linux is having on the Unix market, especially on Coherent and lower-tier System V vendors?

Linus: This one is hard for me to really say much about: I havent been in contact with any real i386-Unix users, and have only once seen a Xenix system being run on a friends machine (that one was converted to Linux, but that doesnt really count when I know him personally). I have gotten various mails and seen some newsgroup messages about persons who have switched over already or would like to switch over once Linux is able to run commercial binaries, but at least so far, I doubt Linux has dented the real unix market very much. Coherent might have a bit more problems competing with Linux. While Coherent is commercial, it doesnt carry the same real Unix stamp as SCO and the other major PC Unix providers, so a potential Coherent user is also likely to chose Linux, if he has access to it. And the superior performance and features of Linux may well be (and has been in many cases) reason enough to chose Linux despite the reportedly good documentation and support of Coherent.

Being binary-compatible with SVR3 and SVR4 might change the picture a lot: it would make it possible to reasonably easily mix Linux machines into an existing machine park, and would make Linux much more viable in some situations. The current kernel can load ELF binaries, and COFF support is available as patches. The actual binary code emulation is still being worked on, but there seems to be no major obstacles. If the Wine project (running Windows binaries under Linux and X11) also works out, the picture changes again.

Meta: What is your opinion of 386BSD?

Linus: Actually, I have never even checked 386BSD out; when I started on Linux it wast available (although Bill Jolitz series on it in Dr. Dobbs Journal had started and were interesting), and when 386BSD finally came out, Linux was already in a state where it was so usable that I never really thought about switching. If 386BSD had been available when I started on Linux, Linux would probably never had happened.

I also have very limited computer resources (right now I have 160MB of disk spacethe original Linux development was done in 40MB), so I havent tried to set up 386BSD just to see what the competition does. This means that I have only followed the 386BSD discussion and development from the side. As far as I can tell, its a good port of BSD that is plagued by some problems (mostly non-technical).

One of the major problems with 386BSD seems to be the lack of co-ordination: that may sound weird coming from the Linux background, but in fact the 386BSD project seems to suffer from a lot of people working on the same thing due to the long release cycle (I think there are three different and incompatible keyboard/console drivers for 386BSD). A long release cycle is the way to go in a controlled environment (i.e., commercial development), but I think it hurts the free development that results from a lot of unconnected persons having access to sources and doing lots of modification. The NetBSD project may be a step in the right direction, but I think 386BSD has been hurt by the way it has been developed.

Note that others that know more about the actual 386BSD development may disagree and think the Linux releases have been very chaotic (which is also true, but differently). Also, 386BSD has had different starting points and different goals, so any real comparison may not really be valid. In any case, I usually ignore Linux/386BSD comparisons: Ive not let any 386BSD considerations change the way I work, but just done things the way I want them done and hoping it works out. I have gotten a few mails like were considering changing over to 386BSD, as Linux doesnt do… but I refuse to be blackmailed by things like that. Ive also gotten mails from people who have changed the other way, so its obviously a matter of taste.

Meta: Some people, particularly Peter MacDonald of SLS, have been criticized for trying to make money on free software. What is your opinion of this?

Linus: The people who criticize Peter are usually persons who have written none of the kernel, or even user-level code, and I hope Peter (and others) just ignore the monetary issues raised by some. Peter not only has written code for Linux (he worked on the original pty and VC code, which was adapted by me, and he is still making suggestions and patches to the kernel), but the SLS release has been of immense value for the Linux community. SLS has its share of problems (which also get criticized), but there is no question about the fact that it was one of the things that made Linux really available for Joe User.

The fact that others make money by selling Linux is something that I find mostly amusing, and something which does my ego no end of good. Frankly, I wouldnt want to bother personally, so if somebody else does it, it doesnt hurt me. Its also quite legal by the copyright, and so far I havent seen any major developer stand up and say he doesnt like his code being sold, so I dont see the problem.

Meta: There seems to be a perception that Linux is very tied to the 386 architecture and would be very hard to port to others, yet there are apparently projects underway to port Linux to the Amiga and MIPS. Whats the real story?”gondwanaland.com

Let’s say that you’ve just decided to rename a variable from foo to fooOne. In Vim, hit Esc for command mode, then use this command:

:%s/foo/fooOne/g

% means that the operation should be carried out throughout the whole document. The important part is s/foo/fooOne/, which means “replace every instance of ‘foo’ with ‘fooOne’”. The final g means “global”; without this you’ll just replace the first instance on every line, but with it, you replace every occurrence.

To use this search-and-replace pattern in Emacs, hit M-x then type replace-string RET foo RET fooOne.

However, while this non-regexp operation would replace foo with fooOne, it would also replace foobar with fooOnebar, which you probably didn’t want. To get around this, use the word boundary markers \< and \>:

:%s/\/fooOne/g

This restricts the replacement to occur only when ‘foo’ exists as a word on its own (with a word boundary character on each side of it). In Emacs:

M-x replace-regexp RET \ RET fooOne

Backreferences

Backreferences (as used in the previous tutorial) can also be very useful. For example, say you wanted to change all the date references in a file from US-style (09/22/09) to UK style, with long year and a dot instead of a slash (22.09.2009). This regexp would do the trick in Vim:

:%s#\< \(\d\+\)/\(\d\+\)/\(\d\{2\}\)\>#\2.\1.20\3#g

For Emacs, use:

M-x replace-regexp RET \< \([[:digit:]]+\)/\([[:digit:]]+\)/\([[:digit:]]\{2\}\)\> RET \2.\1.20\3

OK, that looks quite complicated! First of all, let’s note that in vim, we use # rather than /, giving us s###g rather than s///g. This makes it easier to read if you’re looking for / in the pattern, and also means that you don’t need to escape any / characters.

As discussed in the previous article, each pair of escaped brackets, \(PATT\), store a backreference to PATT. Here we have three backreferences, with a word boundary in front and afterwards (the \< and \>), and separated by a slash between each of the backreferences (as in 09/22/09).

The first pattern we’re looking for is \d\+: this means at least one digit character (\d). So this will match 9, 09, 12, etc. In Emacs, this is written [[:digit:]]+ (there is no need to escape the + in Emacs regexp syntax, as you must do in Vim). You can also use [[:digit:]] instead of \d in vim if you prefer.

The second backreference pattern is the same as the first one, to match the number of days. The third pattern, \d\{2\} matches exactly 2 digit characters (\{n\} matches exactly n of the previous character type), because years aren’t usually written as single digits.

The replace string is then straightforward: reorder the three backreferences so that the day digits come first, then the month, then the year with 20 in front of it, all separated by a period.

Warnings:

* This upgrade path is for the x64 Edition only, from a Mint 7 Gloria system to a Mint 8 Helena system.
* There is no guarantee that it will work for you. In fact this is quite a risky process. If you’re experienced and if you know how to troubleshoot and solve common Linux problems (in particular X11, kernel modules and APT problems) then you’re probably OK. If you’re a novice user we recommend you perform a fresh installation of Linux Mint 8 instead.
* You should make backups of all your data before upgrading.

Upgrading vs a fresh install:

The recommended way to install Linux Mint 8 is to download and burn the liveCD. Another alternative is to upgrade your existing system as detailed in this blog post. Note that we do not recommend you to do this and we insist on the fact that you know the pros and cons associated with it and the risks involved in doing so.

Pros:

* With an upgrade, you keep the software you previously installed, you don’t need to reinstall it again as you would after a fresh installation of Linux Mint.
* With an upgrade, you keep your settings and your documents, you don’t need to restore them from a backup. Note that you can also do this with a fresh installation by using a dedicated home partition and that no matter which way you’re upgrading or performing a fresh installation, you should ALWAYS make a backup of your personal data.

Cons:

* Upgrading is slow, because you need to download the new versions of all the packages. In comparison, the liveCD contains about 2.5GB of compressed data in a single download of less than 700MB and the fresh installation takes between 10 and 15 minutes. Consequently, it’s much faster to perform a fresh installation of Linux Mint than upgrading your existing one.
* Depending on what’s installed in your system, package upgrades can trigger complicated conflicts. If you’re experienced with APT you’ll probably know how to solve the problem. If you’re not, you could end up half-way between Mint 7 and Mint 8, and if you’re unlucky even with an unusable system.
* There’s a code-freeze before Ubuntu gets released, and then testing goes on. The same happens with the Linux Mint release, and so what you end up running when you perform a fresh installation is a system that is known to us for its statibility and which has been thouroughly tested. An upgrade applies all available package updates no matter how unsafe they are, from level 1 to level 5, so if there are regressions in these updates you’ll get your system impacted by them.
* With every new kernel, there are differences related to hardware support. To know if your hardware is fully recognized by the new Linux Mint release you should try out the liveCD and check that things work before making the decision to install or to update your system. There are have been numerous people saying that their wireless cards, their sound device or some other part of their system was recognized with the prior version of Linux Mint or Ubuntu and didn’t work as expected anymore. For some people, the latest release isn’t necessarily the right one. It’s better to be safe than sorry, so if you’re going to download the liveCD to check your hardware support anyway, you may as well go for a fresh install and avoid downloading things twice.

Comparison with Ubuntu:

* Upgrading Ubuntu is not safer than upgrading Linux Mint. It’s equally risky.
* Ubuntu doesn’t mention the risks involved in package and release upgrades. Their policy is to fix whatever gets broken and to assume that the regressions caused on your system will get solved by future upgrades.
* Linux Mint insists on these risks and recommends a prudent approach to upgrades. Our policy is to avoid possible regressions by being selective on the updates we recommend to you.
* Upgrading Ubuntu is easy, and easier than upgrading Linux Mint. It shouldn’t be though, and if there’s any risk involved in you braking your system, then the least we can do is to write a long boring post about it, to make you think twice about doing it, and to throw warning signs at you before you click on the shiny button

Upgrading graphically (easier):

* Open a terminal and type the following commands: “apt update” and “apt install mint-upgrade-tool-x64”
* Open mintMenu and run “Menu->Administration->Upgrade to Linux Mint 8″
* Follow the instructions.
* If asked, always choose “Replace”.
* If asked, always choose “I”.
* When finished, reboot the computer.
* Change your wallpaper to “Fresh” and your theme to “Shiki-Wise”
* Run the upgrade tool from the menu again and let the upgrade tool finish all the way to the end.

Upgrading from the command line (faster):

Open a terminal and type the following commands:

* gksu gedit /etc/apt/sources.list (Change all occurrences of “gloria” to “helena”, and all occurences of “jaunty” to “karmic”, remove the lines for the Community section of the Linux Mint repository, then save the file and close the editor)
* apt update
* apt install mint-info-x64 (choose “Y” or “I” to install the package maintainer’s version)
* apt install mint-meta-x64(choose “Y” or “I” to install the package maintainer’s version)

In the terminal, repeat the following commands until both upgrade commands show no upgrades available:

* apt upgrade
* apt dist-upgrade
* apt install mint-meta-x64