Linux Vendors to Standardize on Single Distribution
Jon James writes "eWeek is reporting that a number of Linux vendors will announce on Thursday that they have agreed to standardize on a single Linux distribution to try and take on Red Hat's dominance in the industry. " The vendors in question are SuSe, Caldera, Conectiva, and Turbolinux. However, as the article also points out - Red Hat has a very well established lead in the corporate market - and Sun's decision to create Yet Another Linux Distribution (Sun Linux! Now With McNealy Vision!) will make the waters even more muddy.
Actually not, they are just adapting to a new form of existance. Targeting geeks it was simple, every geek has it's own needs and would like it's own distribution.
Targeting masses actualy defines being more organized and more uniform. This way linux development actualy speeds up, what's one of the main things of this merging.
Setting one standard and deploying jobs across few companys that had to do all the work untill now. Speed is increasing, uniforming gets better and most importantly. There is a higher organisation level
Signature Pro version 1.13.2-3 release 83.5 beta3try7 after-breakfast edition
The one reason I have not yet downloaded this is recompilation. I mean, compiling is pretty much a deterministic activity. Given similar compiler settings you'd expect the result to be the same each time. Apart from being deterministic it is also time consuming. Just compiling a pretty bare bones installation with gnome, kde, open office, mozilla would likely take me weeks by which time most of the packages would need recompilation because of updates!!! I'm all in favour of optionally compiling a few key things but I'm even more in favour of using pre-packaged binaries. Most of us probably would go for the 686 type code, so create binaries for all popular variants of X86 and distribute those (and maybe also other processors).
An alternative, admittedly far fetched, idea would (imagination going beserk here) be p2p compilation. Compilation can be distributed over computers and there likely is a small subset of all possible compiler settings that is most frequently used. Simply cache the results for such compilations and given a match in source code version processor architecture and compiler settings, reuse the result (and offer the replicated binary for download). If there is no match, compile yourself and offer the result. This should quickly eliminate redundant compilations and offer most of the advantages of compiling everything yourself.
Jilles
The one reason I have not yet downloaded this is recompilation. I mean, compiling is pretty much a deterministic activity. Given similar compiler settings you'd expect the result to be the same each time.
... incomopatabilities should mean major revision number changes, not minor), but it does happen. When borrowing packages and binaries from other distros this becomes more acute.
... again, this is very acute when moving from GCC 2.9.x to 3.x, and again, compiling everything yourself fixes that problem.
What isn't deterministic is what packages (and what versions of those packages, and what compile-time options for those packages you've selected) you've chosen to install. If you're using somelib.so.1.0.1 and someone else is using somelib.so.1.0.2, there is a small (but real) chance that a minor incompatability will result in a binary compiled against one displaying some occasional flackeyness when run against the other. This isn't terribly common (and it represents a mistake on the library maintainer's part when it does happen
Compiling on your own machine eliminates this.
There is also the problem of binaries compiled with different versions of the GCC compiler behaving is subtly different ways
If you have a decent processor, compiling isn't really that burdensome (the initial installation excepted of course). Most people start their daily or weekly upgrades in the evening before going to bed, making the burden effectively zero. In any event, the advantages are well worth the trouble, and the speed improvements are dramatic.
Your P2P idea is interesting (sort of a shared cpu cycle approach a la Seti@home). Again, the problem with having others compile for you (rather than sharing cpu cycles you use yourself) is that they will likely have slightly different libraries than you do, for some things at least, possibly compiled with different optimizations, so you cannot be 100% certain that what you are getting is exactly what you want. With Gentoo and Source Mage's approach you can be 100% certain that you are getting precisely what you want, and that it is compiled against precisely what is on your system.
The Future of Human Evolution: Autonomy
Anyone who has messed with Caldera, SuSE and Turbolinux knows that they do NOT produce a 100% redistributable version of their commercial offerings. In addition to now allow redistribution of their CDs, most either omit major packages, or limit usage to "personal." As of 8.0, SuSE has gotten even more restrictive no longer offering free downloads of many components. This alone has turned off this user from considering their software.
Conectiva, on the otherhand, has gained a lot of notariety in their efforts. The two biggest being the use of Apt for RPM, and one of their lead developers managing a Linux kernel branch alongside Alan Cox and only one other. I have not used their distro, and DistroWatch.COM does not differentiate between "free download" and "100% redistributable" so I cannot tell if they maintain the same GPL-anal approach as RedHat. For now, I'll assume so (please let me know if otherwise?).
So, for this strategy to work, assuming the rumor is true, I make the following 2 recommendations to the resulting conglomerate:
These vendors don't have to stop value-adding to their distros. In fact, this approach could still allow them to do so. But they really need to build some mindshare with those of us who like RedHat and Debian because of their 100% GPL-focus. Release a 100% Redistributable CD set which they all agree on. This has kept me from using Caldera, SuSE and Turbolinux over the years.
Then each can include their own CD #1 binary, "alternate," non-redistributable boot CD in their commercial, boxed sets so the value-added stuff can be installed (in addition to other, non-redistributable CDs). The idea is that the install packages should be the same for both the freely redistributable and commercial non-redistributable versions, even if the default/base freely redistributable ones are replaced by those in the commercial, non-redistributable CD(s). Simple, no?
This will get the masses to join them. If the new conglomerate can build a new, 3rd party software repository for Apt like Debian has for Deb, this would get me to use this new distro. And they would quickly find that a number of 3rd party free software / open source projects would make sure their packages are built for and distributed in this new RPM-Apt repository. God knows I'd be sold in a heartbeat, assuming the distro quality is as good as RedHat. With SuSE in the mix, I don't see this being an issue, since I have used their kernels before (and trust them as much as RedHat).
Right now I mix a custom distro (usually installed via NFS so I don't have to build CDs that are outdated quickly) use RedHat with Ximian and FreshRPMS added. Ximian is Ximian, and I don't forsee not using their Gnome set (this new "standard" distro will make it easy for them to support). FreshRPMS is RedHat-focused and uses RPM-Apt, but it is far from "comprehensive" with only about 50 packages or so. This is a far cry from Debian's 10,000+ Going to RPMfind or the older contribs is just not viable, and I don't bother much anymore. But I don't have nearly the package selection as Debian with RedHat and this frustrates me since I will not use Debian for other reasons (I'm not going to expand on them here, just note I said *I* will not use Debian -- not that Debian is "bad," not at all).
-- Bryan "TheBS" Smith
Independent Author, Consultant and Trainer
Here's a slightly different perspective: I am still relatively new to Linux (>2 years), and I'm just a rank beginner when it comes to programming. I sit in front of a Windows box all day at work because I have to, and I have a Windows box at home, too, mostly for (in)convenience.
Why do I use Red Hat? Why do I use Linux at all? Well, frankly, the more I use Windows, the more I like Linux. It's stable, powerful, non-stupid, (don't even get me started about Stupid Automagical Windoze Tricks) and it does exactly what I need in a way that works well for me. Also, I think the interfaces are fascinating, so I'm writing a paper about them (for the arts/social sciences community) now.
On the other hand, I neither have the skills nor the inclination (yet) to spend hours tweaking and reprogramming config files so that I can get something up and running. I like that it works. I like that I can do what I want with it, and I don't have to tinker with it incessantly.
Sorry if that sounds kind of anti-hackerish (it's not meant so), but I'm still trying to master the basics, and I wouldn't try to drive a Formula 1 racer while on my learner's permit, either.
I'm not a geek, I'm just a clever script.
I use Red Hat Linux for my servers. I'd not dream of anything else at the moment. Why?
1) Excellent support - whatever software I want to install, I can be quite sure that there's a RH version - often in RPM form. This reduces the cost of maintanence dramatically.
2) The RED HAT NETWORK is fantastic! I simply type "up2date -u" and 10 minutes later, I have all the relevant security patches installed! Just $5 per month, and their download servers are FAST. (I routinely see 15-20 Mbit connections - 10x-15x FASTER than an unfettered T1!)
3) Reliability. My Red Hat systems are stable. They work today, tomorrow and next year.
4) Stability of the distro. Red Hat has been around. They are profitable, or at least not burning capital very fast. I can feel good knowing that I'm investing my considerable time, money, and energy into a platform that will be there in the future, too.
With the above, I can fulfill my support contracts easily and cheaply, and focus on the delvery of service rather than simple maintanence.
Is Red Hat perfect? No. But it satisifies the above, and they are what I need to found my business upon.
I have no problem with your religion until you decide it's reason to deprive others of the truth.