Category Archives: Linux

ArchLinux, not just for the elite

Posted on by
41

… and I’m the living proof of it!

I had several colleagues, friends and people asking me whether they should run Arch Linux on their desktops or laptops. I even read someone’s blog today on his impression on Arch Linux and Ubuntu. It’s time for me to jump in and clarify what you should expect with Arch Linux as a desktop on a daily basis.

Arch Linux is a rolling release system. What this means is that you do not get releases at specific intervals in time, like you do with Ubuntu, OpenSuse or Fedora. Instead there is a constant stream of updates that are uploaded on the distribution servers and that you can pull almost everyday. These updates are uploaded after a testing period by the Arch Linux  testing community (you can switch to the testing mirrors if you wish) and it is up to you to choose if you want to install them or not.

Such a rolling release process eliminates the need to accomplish major upgrade and makes you gain time, as you typically end up installing your Arch Linux system once, or twice if you really screwed up something. Also, Arch Linux does not come with very specific tools (aside the pacman package manager) and therefore you do not end up with Unity vs. Gnome Shell or YAST and PUP, or whatever control center. You get the latest KDE version, the latest Gnome 3 version, the latest Unity and the latest Xfce (these are examples). Pretty much everything is configurable as the distribution gets to make choices on core components versions (glibc, python, etc.) and exercises its value and role on testing and QA (what happens after each kernel upgrade, etc.)

Yet all this does not mean the distribution is hard to use. Not at all. The installation process may take a while (several hours.. or less) and I would be tempted to claim that what takes time is to transfer your own content and granular application settings to the new system, such as themes, pictures, etc.

But let’s focus a bit on the installation process: that’s where things tend to get rougher. Arch Linux uses a command line installer. It does not make things very difficult to understand – besides, you can always refer to some very good documentation – but it definitely makes the process more intimidating and any issue or inconvenience tends to be perceived as a bigger annoyance than what it really is. Of course, such a comment has to be put in context of other Linux flavors where you insert a DVD and don’t do much aside choosing your keyboard and entering your name. Not so long away you still had to be careful when partitioning your hard disk even with an user-friendly interface. In any case, the installation process is what will make you reach a working, fully graphical and modern system or a glowing command-line mess. There’s nothing specific to avoid here, only know that your patience and work will be rewarded and that in a sense, such an installation is not that hard to perform.

Once your system is up and running, everything tends to run smoothly and you end up with a nice, fully customizable desktop. You can use  “community contributed packages” from the Arch Linux User Repository to complement your software tools, themes, fonts, games and utilities.

As a conclusion, I would say that while Arch is not as easy to install as, say, Ubuntu, once you’ve gone past it, you will be surprized how easy it is to use it, almost as easy than Ubuntu or any other distribution. Arch Linux is a very fun and stable distribution that successfully blends the bleeding edge, stability and hackability of Linux. Don’t be fooled by the rumours saying it’s for the elite. It is made for you, if you can give it 3 hours maximum of your time to install it, and it is likely you will never switch back.

AppSet: a refreshingly nice package manager for Arch Linux in the times of app stores

Posted on by
1

In this post I will not talk about LibreOffice or open standards  but I thought this could be of interest to GNU/Linux users out there so feel free to comment and discuss.

I’m a rather outspoken user of Arch Linux after having used and tried many other distributions (MandrakeSoft/Mandriva, Suse, Debian, Fedora, Ubuntu, Mint, SLAX, Chakra and even a few others) and I think I got to like the rolling release concept quite a lot. The rolling release concept essentially takes away the notion of milestone release for a Linux distribution and replaces it by incremental and almost continuous updates. Which means that everyday I can update my system and it’s thus almost always running the most recent stable software versions. Note that the upgrade is my choice only, I could stop doing this for 3 weeks instance and that would be fine. Using Arch Linux does not only mean embracing the rolling release distribution model. It also  means being ready to install your system from the command line (granted, you only do that once in theory) which can be tedious but not reall difficult. Another “side effect” of using Arch Linux is that the distribution’s package management is done entirely through the command line and with the help of the excellent package manager pacman. Pacman is however not a graphical package manager, or rather, it does not come with a default, out of the box graphical front-end. Several of them do exist but it does not seem to be in the culture of Arch Linux to use one on a regular basis. Enter AppSet. AppSet is a very nice graphical package manager written in Qt; it even got me use KDE again on par with Gnome. AppSet does not only run on Arch Linux, it also supports Chakra (a very close fork of Arch Linux) and works in theory with any other packaging system.

In the times of ubiquitous App Stores, even the ones that actually values and promotes Free Software such as the Ubuntu App Store, it’s good to hear you still have innovation happening in this field, with tools being developed that let users be in total control of their system. I’m sure that the AppSet team would welcome more contributors and more distributions!

PPAs and LibreOffice

Posted on by
5

Waow, two posts in two days. I have to calm down. Anyway, I wanted to answer publicly to a question that has been asked on our mailing lists, on IRC, through private mails, and I’m sure it will be asked by the press at some point. I’m not an Ubuntu user, I’m an Archer. But my fiancée and my father use Ubuntu on a daily basis, and I also took the time to fetch some information about LibreOffice PPAs. What seems to be the problem? As you are aware, we have just released LibreOffice 3.4.1 and our scheduled, fixed release time does not always coincide with Ubuntu (or any other distribution) release. Therefore people -but it seems mostly Ubuntu users- ask for PPAs of LibreOffice, in other words, packages that are ready to install on their system with the most recent stable version of LibreOffice.

First I would like to thank everyone for their interest in LibreOffice! Second, I think it’s very important to understand that there’s a difference between PPAs and the default version shipped by Ubuntu. Just like any other distribution, Ubuntu releases a full GNU/Linux system that comes with a set of fully defined and qualified packages. Unless Ubuntu chooses to upgrade these packages themselves, they won’t move or change until the next version of the distribution is released.  PPAs are a community based and convenient way to use more up-to-date version of software packages, but do not expect the same quality or to have a fault-proof software running; it’s an upgrade for the users who wish to enjoy their system with more spice and not run a stable and fixed system with little or no change from the original release.

In the case of LibreOffice, it is very important to realize that THE  DOCUMENT FOUNDATION DOES NOT RELEASE ITS OWN PPAs. Please don’t ask us to release a PPA. Inside the LibreOffice, each GNU/Linux distribution is responsible for its own packaging. We do however release .debs and .rpms but these are our very own. The packages you will find in many distributions tend to be optimized or repackaged by the distribution developers themselves.  So if you look at Launchpad, you will see that the LibreOffice PPAs are there and are mostly done by the official Canonical LibreOffice team. They’re members of the Document Foundation themselves, but do not expect the Document Foundation to release PPAs, as it is this team’s sole choice (mostly based on their own remaining available time, I would imagine).

If you really want to help, I would suggest you join the Ubuntu team, or start packaging your own PPA. And you will always have the possibility to install the .debs by yourself by downloading them directly on the LibreOffice website.  Also please check our general release plan for an overview of our general release timing in 2011 and 2012.

Links for the end of April

Posted on by
Reply

I am having a very busy month of April, but I mean, a really busy one. I am alive and kicking, but I am swamped.

Here’s a couple of links before an even more active month of May:

  • Ars Aperta has contributed to a pretty interesting project, dubbed ODFgr and hosted by the OpenDoc Society. The goal of this website is to provide any developer with even a limited knowledge of ODF with resources and tools to manipulated ODF documents. We tried to design a pedagogical platform that the largest number will understand. Most of the examples are listed by languages (we mostly have Python and Perl) and you can study both the explanation and learn how to reproduce and implement it. We hope it will be the right spot for anyone willing to get started on OpenDocument hacking and development.
  • Events-wise the month of May will be busy. I will attend the OASIS Board of Directors’ meetingin Berlin and meet with the Bitkom. The week after that Ars Aperta will join a session on the political and legal issues pertaining to Free Software development during the Linux Solutions 2011 event in Paris. I will also give another talk during the same event as part of the Document Foundation and our experience with forks. Spoons shall come next year.

Radical Innovation is needed for GNU/Linux distributions

Posted on by
6

There’s a certain movement these days in the world of GNU/Linux distributions.  I think we are experiencing one of these moments that starts with a question that has been asked and heard many times -should distros differentiate themselves in order to survive? & aren’t there too many distros out there?- and ends with a much more serious question: Innovating in the world of GNU/Linux. Rest assured this is not going to be that sort of rant where we conclude that “Linux is the copycat of other OSes” just like we will not, in fact answer the question of the pretendly too many distributions or their differentiation. That is, I will not really answer these questions; and the reason I won’t is that I think these are all bad questions that either miss the point or show a certain lack of understanding of  FOSS and GNU/Linux in general.
I guess by now all of you have heard of Mageia, the Mandriva fork. But these news overshadowed something else that is a developing situation
elsewhere and matters perhaps even more: OpenSuse.

In a nutshell, OpenSuse has been breaking away very slowly from its main sponsor, Novell, for about 2 and a half years. The first visible sign of this -which really was a weak signal nonetheless- was the decision taken by the community to switch back to KDE as their preferred desktop instead of Gnome. Of course, just like Mandriva/Mandrakesoft, Suse had always been more KDE oriented than  Gnome. Yet Gnome is where the business, the stability, and theenterprise applications are supposed to be found, and on Gnome lied Ximian, the Groupwise integration etc. Then the OpenSuse folks started to open a brainstorming plan in order to define a new strategy for OpenSuse, apparently independent of what Novell was planning to do or sell with respect to that. This strategy brainstorming session ultimately reached its conclusion a few days ago:

https://lite.co-ment.com/text/lNPCgzeGHdV/history-version/RE3kSeg3LGI/

As you will see, what OpenSuse intends to be is a general-purpose, desktop oriented distribution; which means at the same time that nothing will change in its actual orientations and that it even departs from its usual enterprise polish it always had had. But what this also means is that we will not see OpenSuse or Suse on handhelds or tablets or any other new markets. This is a significant information, especially if you see that whoever will buy the Suse part of Novell in early 2011 might not be able to have its own way if  it does not take the time to engage with the community: The OpenSuse project seems to be very autonomous and not at all ready to fall into whatever new goals any future sponsor might want to achieve. And if it takes a fork to dot it, there’s the Mandriva case.  But always remember that OpenSuse has a very strong userbase and market share, although it’s been declining ever since 2009. What will be interesting nonetheless will be what the future owner of the Suse brand will want to do and how it plans to innovate. OpenSuse can be a general-purpose distribution; the user base is there, but the value might be hard to create if there’s no real business story to tell behind it.

Back to Mandriva / Mageia now. It’s perhaps to early to say anything about Mageia, except they seem to be made of some pretty skilled  people; and that’s usually not the kind of engineers you find easily on the market. They claim to continue what Mandriva as a distro was good at, only in a better way, and without the perceived historical failures of the past management teams.

Interestingly enough, I think Mageia is bad news for Mandriva, and it means that Mandriva should find an innovative business model and acquire/change to a new focus. Let me explain. Reading the Mageia website and going around the Internet, here’s what I understand:
- Mageia realizes the need to be a linux distro for other kinds of
terminals (tablets, handhelds, etc.)
- Mageia has crafted two strong bulletpoints in its storytelling that DOES hurt Mandriva starting today: Mageia “is” Mandriva, since it is
made of the engineers who have coded Mandriva ever since a few years; second, Mageia is “better” since they understood what “is wrong”: the management of Mandriva. (Nobody ever found anything to complain about Mandriva as a distro, it’s still one of the best on the market).
- Mageia is soon to “take over” the market: everyone on the forums  seem to dig Mageia; and in a sense, it’s what the Mandriva community and the French FOSS community was expecting.

If the last claim sounds bold, think again: what is the value of having a Mandriva desktop outside of a corporate support contract (same goes for a server) now that there’s Mageia? The way to create value for Mandriva is to depart from the traditional all-purposes distribution model (which still does not mean they would have to “cut” the actual distribution) and innovate first at the distribution level, and then, if possible, go up the ladder by growing a very skilled technical team able to innovate as an operating system, either by contributing upstream again, which it hardly does anymore these days, or innovating on the user experience just like Ubuntu does and is now clearly intensifying as a strategy.

In the case of Mandriva and Mageia, what might become interesting to watch is the potential race between the two twin-distributions; one is now almost an empty shell, deprived of its developers, and the other one has developers but no resources. In any case, it’s time these two get a real shot at innovating, for the sake of the entire Free and Open Source Software ecosystem.