lwn qui fait des prédictions pour l'année suivante, les compares avec ce qui c'est réellement passé :
Morceaux choisis :
- We also predicted a growing backlash against enterprise Linux and their supporting business models, and the possible emergence of free alternatives. Certainly, resentment toward the enterprise distributors continues to exist in some parts of our community, and some of those people are doing something about it. But many of the projects which aim to undercut the enterprise Linux business model - CaOS, Whitebox Linux, UserLinux, etc. - appear to have made little progress over the last year.
Perhaps the largest surprise in this area is the emergence of Ubuntu Linux, which is an attempt to provide the best of a 100% free Linux distribution with longer-term support options. Ubuntu has succeeded in making a big initial splash; whether that will turn into a successful business remains to be seen.
- Kernel. The prediction that the 2.7 development series would start seemed obvious, but it was wrong. We did sense that the development process was changing, however, and predicted that the next development series would differ from 2.5. The pressures which might lead to a new development series still seem to be mostly absent - mostly because the 2.6 development model tends to prevent those pressures from building up.
- What we missed: LWN would like to apply a small patch to its 2004 predictions to fix a few bugs. So we now predict that, in 2004:
* Despite all appearances, software patents will not be enacted in the European Union. Yet.
* Mandrakesoft will emerge from bankruptcy, shake off much of its debt, and start to function as a profitable company.
* New FUD attacks against Linux will target total cost of ownership and intellectual property concerns; none will have much success.
* The Debian "sarge" release will not happen, and, in fact, will appear to be no closer at the end of 2004. Increasingly, Debian offshoot distributions will handle the task of creating release-ready versions of that distribution.
* Some large companies will publicly promise not to use their patents against Linux users, or, even better, to use their patent portfolios to defend (at least some) Linux users against patent attacks.
Linux timeline de 2004 pour revenir sur les temps forts de 2004 :
Avec une photo de M. Linus Torvalds qui fait caca :
Les prédictions de lwn pour 2005 :
Morceaux choisis :
- Free databases will see some high-profile deployments. The adoption of free database management systems is still in an early stage. Things will progress in 2005, to the point that some proprietary database vendors will see the need to start competing directly against the free alternatives. Perhaps 2005 is when we'll see some real free database FUD.
- There will be no 2.7 kernel in 2005, despite the requests for such a release from some quarters. The 2.6 process will continue to merge changes at a staggering rate, and nothing will come along which is so disruptive that it forces the creation of a new development series. The steady series of complaints about the quality of the 2.6 mainline releases will force some changes to the process - we may see more frequent releases or true "release candidates" for wider testing. But the simple fact is that the kernel developers - and the distributors who have the job of delivering stable kernels to their customers - are happy with things as they are, and will not be in a hurry to go back to the older way of doing things.
- Debian will get a new stable release out[*], one way or another. Much of the user base for stable Debian releases will, however, have moved on to offshoot distributions like Ubuntu. There will be a new round of soul-searching within the Debian Project over the value of its stable distribution and what that distribution should be.
Très risqué comme prédiction...
- In conclusion...
More people will notice that Linux users don't have spyware and adware problems, which will be getting steadily worse on other platforms. This issue, alone, will cause more people to look at free software. Many will get their feet wet with Firefox and stop there, but others will take the full plunge. As proprietary systems are turned into zombies which spam and spy on their alleged owners, pure exasperation will push a new round of Linux adoption.
Dans lwn de la semaine dernière (disponible gratuitement), "Four-level page tables" de Linux (en cours d'introduction dans 2.6.11) est couvert.
- Before the patch was merged, the x86-64 architecture could not effectively use the fourth level and was limited to a 512GB virtual address space. Now x86-64 users can have a virtual address space covering 128TB of memory, which really should last them for a little while.
Avec une petite note sur le modèle de développement de Linux pour la série 2.6 :
- The merging of this patch demonstrates a few things about the current kernel development model. Prior to 2.6, such a fundamental change could never be applied during a "stable" kernel series; anybody needing the four-level feature would have had to wait a couple more years for 2.8. The new way of kernel development, for better or for worse, does bring new features to users far more quickly than the old method did - and without the need for distributor backports. This patch is also a clear product of the peer review process. Andi's initial version worked fine, and could certainly have been merged into the mainline. The uninvited participation of another developer, however, helped to rework the patch into a less intrusive form which brought minimal changes to code outside the VM core. The end result is an improved kernel which can take full advantage of the hardware on which it runs.
Dans la série des articles sur les distributions AMD64, Mandrake 10.1 se fait tailler un costard :
- Is Mandrakelinux 10.1 (X86-64) worth ¤119? As we did not test the commercial edition of the product, we cannot really answer the question, but the FTP edition has given us enough warning signs to put any recommendations on hold. Frankly, it is hard to see how Mandrakelinux will compete with other 64-bit distributions on the market, especially with the likes of Fedora, but also Debian or Gentoo, which are free of cost and available for download immediately after release (or continuously updated). Additionally, all three of them have more up-to-date packages (Mandrakelinux 10.1 ships with GNOME 2.6 and KDE 3.2.3), fewer bugs (especially when compared to Fedora Core 3), and more responsive mailing lists and user forums, actively monitored by the distributions' developers. Mandrakelinux 10.1 X86-64 is not a bad product, but it is marred by lack of polish and some unnecessary commercial tricks.