Jamie's Random Musings
Various thoughts and adventures, including but not limited to Linux, Windows XP and Widows Vista, and assorted bits of hardware new and old.
Tuesday 13 October 2009, 5:52 AM
Mandriva One 2010 RC2 Available
I am downloading it now, and will install it on everything around here over the next couple of days. If it is even better than the RC1 release, it will be very good indeed.
jw 13/10/2009
PreviousComments on this post
How do you find time to test out all these distros!? I have enough on my hands with Ubuntu, always finding new ways to do things and new software.
It must give you a good overview though of where the distros are headed -- are they _that_ different given they're based on Gnome & KDE?
Hi Jake, Thanks for reading and commenting, as always.
My current motto is "it's easy to be flexible when you're unemployed". A corollary to that is that I am able to spend as much time as I like doing things that I find interesting. That is likely to change somewhat as of next week, when I start my new job.
Although many of the popular distributions are based primarily on Gnome and KDE, there are getting to be quite a few who use Xfce, either as their primary desktop or as an alternate that can be selected by the user. Even within the KDE distributions, there are some who are still using KDE 3.5, and there are several different KDE 4.x versions in various distributions at the moment.
It's probably fair to say that the distributions aren't that different, in so far as they are all based on some fairly recent Linux kernel, and they almost all use one of the three or four popular desktops. But there are significant differences in a few areas, some cosmetic and/or transient, and a few much more significant:
- Add-on tools and utilities. Mandriva is well known for this, with their various drak utilties for network setup, patch tracking and installation, X.org setup and so on. Linux Mint has a very nice set of utilities built on top of an Ubuntu distribution. Fedora is another good example.
- Linux kernel versions. We seem to be in a phase where there is a lot happening with the Linux kernel, as evidenced for example by Ubuntu jumping from 2.6.28 in Jaunty to 2.6.31 in Karmic. This is one of the few times that I can remember where so many different kernel versions are included in different distributions.
- Device driver support, both variety and speed of inclusion. When I first got my HP 2133 Mini-Note, I couldn't find a single distribution which supported everything in it - the VIA C7-M CPU and Chrome 9 graphic card, and the Broadcom WiFi adapter, for example. Over the following six months, all of the major distributions added support for it, but the trick was keeping an eye on which one was doing what and when.
- Frequency, quality and ease of updates. The days of having to track, download and install updates and patches manually are pretty much over, but the way this is done automatically varies significantly between distributions.
- Additional software and packages included. The obvious areas here are Office/Productivity software, where many distributions include OpenOffice.org, a few still include KDEOffice, and some don't have anything at all, and Multimedia software, with all sorts of different audio and video media players, digital camera and photo management packages, and so on.
I could go on, but you get the idea. Of course, I haven't even mentioned superficial differences such as visual themes and graphics, but these can make a huge difference in both "first impression" and "ease of use".
jw 14/10/2009
Wow, that's some response! Doesn't the existence of so much choice in distros actually _detract_ from the appeal of Linux to the new user? One of the reasons I use ubuntu (apart from the hardware support, ease of installing apps, online support and general tidiness of it all) is that it's popular. You could say a "de facto standard", if you wanted to stir things up ;)
The multitude of Linux distributions is a disadvantage in some regards, no doubt. One of the biggest is that it gives critics and detractors something easy to point at, and when their major objective is FUD, it is easy to create a lot of the "U" by holding up a list of anywhere from 10 to 50 different Linux distributions and saying "how do you choose?". In the same vein, but a much more serious problem in my opinion, is that it reduces the credibility of any one of them with OEMs, and that is where the real acceptance decisions are made. (Hmm. I'm not sure "credibility" is exactly the word I want there, but I hope you get my meaning.) Even when an OEM is sufficiently {smart|brave|foolish|whatever} to offer a product with Linux preloaded, they immediately start getting inundated with "why did you choose that distribution, and not my personal favorite?". I think the results of this are pretty clear from some of the choices a few OEMs have made so far. There have been several times when I have wondered where in the world they dug up this or that Linux distribution to offer preloaded.
However, having all those different distributions makes the Linux arena far more fertile, productive and resilient than a monolithic single-supplier operating system could ever be. Microsoft made a gigantic, stupid mistake with Vista, and practically the entire world suffered as a result, because there was no alternative. If one of the Linux distributions were to do something massively stupid, the others could easily pick up the slack. Likewise, different distributions are packaged in different ways, with different desktops, different sets of preloaded utilities, packages and programs, so even though they are all coming from essentially the same "core", they can end up being very different from each other. I think one of the best examples of this is the progression from Debian to Ubuntu and then Linux Mint.
It's sort of like drawing a diagram of a "family tree". When you do it for Windows, there's not much depth, honestly, you can count every release from DOS through Win7 on both hands, can't you? There's not much breadth either, even if you are generous and consider DOS, Windows [1..3], Windows [95..ME], Windows [2000..7] and Server to be different branches. Doing the same for Linux, both the depth and breadth are so large that it is extremely challenging to draw it all out. If you want to compare it to a living organism, and go back to my previous statement about Vista, when you're lacking both breadth and depth, a single mistake can have massive consequences, and it is only the gigantic, monolithic power of Microsoft that allowed them to plow on despite that mistake - and even then, I don't think the long term consequences of that mistake are known yet (and I'm still hoping they will be very severe). The same kind of mistake made in any of the Linux distributions would not much more than a hiccup. In fact, not only would the other distributions continue unaffected, and absorb most of the disturbance caused by that hiccup, but it is very likely that those "healthy branches" would feed back into the one with the problem, helping it to recover much faster than it could do alone.
In summary, while I think there are some definite disadvantages to having so many Linux distributions, I think there are even more advantages, and they far outweigh the disadvantages. The more I think about that "family tree" comparison, the more I like it - given the choice between a single line, with very few branches and essentially no "new growth", and a wide, deep, continually expanding network, full of new growth and new ideas, I'll choose the latter every time.
jw 17/10/2009
As for the "many distros" of Linux, and the family tree analogy, I think you have to look at the branches. and how they grow. A lot of the MS branches make lateral moves while the Linux branches are always pointing up.
On the new Mandriva beta, I have been anxiously awaiting this one. Having been a Mandrake/Mandriva user since 2001 I have found that they have always been on the bleeding edge of technological advances. I have tried virtually every distro out there but always come back to Mandriva, or a derivative, like PCLOS.
The way that the open source community advance and develop there primary operating systems without doubt rides the crest of technology advancement, which is good and i can never see that practice ever being surpassed by any other model out there.
They is as already mentioned one major woe as it stands atm with the open source community within this area and that is sheer amount to choose from, what would be good to see is all the different OS teams come together to agree upon a single core model.
If they could all continually contribute to and support a single agreed upon core architecture for distribution, then the rest of the world could crack on writing the 3rd party software for it.
As for things like the user interface who's say they has to be a single face for the core logic why not use a plugin module system for example, if I bought a new touch screen monitor I could then head for the download site and just download a new interface module for it, that would then take advantage of the new screen.
It doesn't haft to end there either think of it as a bit like a theme but much better that literally plugs in & out of the core logic, without impeding on the operation of the core Operating system or existing 3rd party softwares installed.
I think if the above could be done then 3rd party developers along with a lot of end users would switch to it, in an incredible short amount of time.
