Slashdot Mirror
Perens: Unite behind Debian, UserLinux
An anonymous readers writes "Infoworld is running a report on the Desktop Linux Conference, at which Bruce Perens suggested that in order to get Linux to the enterprise desktop, the Linux community should base their efforts on one single distribution... based on Debian.
Perens went on to say that enterprises will be willing to pay Linux companies to engineer versions of Linux to suit their needs, but that the base distro should remain free. He suggested that by 2006, 30% of enterprise desktops will run Linux." Here is a wired story with more information about his proposed UserLinux project.
What makes Linux so great is that there are so many distros, and I can choose the one I like. One distro can never compare to hundreds of them.
If they're running Debian, then that's great. But you need to put Linux into the hands of the masses if you want to take over the desktop and the best way to do that is to seed the planet with Linux Live CD's with the same fury that AOL soils the planet with their CD's.
No gcc, no including twelve different versions of AWK; just the kernel, KDE or Gnome (pick just one), OpenOffice, games, and all the rest of the shit that makes everything go.
Right now, when you say "Linux" to a layperson, they don't know what the fuck you're talking about. A Live CD is a painless way for them to find out.
We can rebuild him. We have the technology.
Is this truly the only Earth I can live on?
I was quite amused when at a recent conference someone described Open Source as Free Software with a politics-obotomy...
Steve Jobs unexpectly announced today he thinks you should use a Macintosh.
Bill Gates made an interesting proposal that everyone use windows.
Scott McNealy outlined a plan he has in which everyone uses Solaris.
Larry Ellison, in a widely-publicized press conference, stated that everyone should give him money.
More on these sudden and shocking developments as news unfolds.
I don't think that the community needs to collectively focus their attention on one single distro. I just think that one single distro needs to rise above the rest and earn market acceptance as a solid desktop. The strength of Linux is that I can use a different distro suited to a particular task. If I need a quick solution for IDS, but don't have some powerful hardware, I can quickly setup snort and Acid on a Debain box and get it going. If I need a quick packet filtering firewall with easy to manage tools (for the IT staff here that isn't very Linux knowledgeble) I can setup Redhat 9 in about an hour and a half.
Somewhere in the near future we need a desktop distro that is every bit as good as Windows is when it comes to the desktop. Then I can say "when I need a quick desktop for someone that just needs web access, eDirectory, and Lotus Notes out of the box, I can use insert distro here."
A step back from what? Right now most US companies running a supported Linux in the enterprise are running Red Hat Enterprise Linux, and it comes with (or with support for) all the products they need, Ie. Java, Oracle, PowerPath, etc. etc. etc.
This is the same "argument" that RMS uses, Ie. It's better to have nothing than something. Life doesn't work like that, people always go for the path of least resistance. Hell even debian wasn't stupid enough to not have "netscape" available when that proprietry and the only real browser. Saying "It's not free" doesn't solve the problem of "I need, now" (and "need" is relative, some people "need" to be able to play proprietry games, etc.).
ustr: Managed string API with ave. 44% overhead over strdup(), for 0-20B
I'm fed up with all this blather about Linux on the desktop. Is it ready yet? What needs to be improved? Why hasn't it happened yet? etc. etc.
There is one thing that is going to get Linux on the desktop, and one thing only. That is that the big PC manufacturers (principally Dell and HP) start to seriously promote and sell desktop PCs with Linux already installed.
If that doesn't happen, then Linux on the desktop will probably never happen to a significant extent.
Because few people want/need to build their own system. Debian has shown an incredible ability to package stable and consistent software which has already become the basis for many different desktop distributions (Corel/Xandros, Lindows, Knoppix). Also Debian supports more architectures than any other Linux Distribution (correct?) and hence all the work done by the various parties would help to ensure that the computer market is held in balance in terms of architectures (i.e. if every Linux distro used Debian as a base, and Linux gained 30% of the desktops, then the ability for "the market" to switch architectures in the event of gross arrogance (i.e. AMD and Intel push through DRM technologies which require annual licensing) would be vastly improved compared to if the Linux distros in use were all derived from RedHat). Of course ideally Gentoo would also collaborate in this enterprise and would become debian derived (i.e. you could do a debian base install and then do "apt-get install gentoo-stageN" to have it use debian as the toolchain to build gentoo, perhaps even building the system out of debian source packages (with gentoo patches)).
Never underestimate the dark side of the Source
I see all these people saying "what is so great about Linux is all these different distro's to try, and Debian is only one".
I don't think you have used Debian. I love Debian because I can put the bare minimum on my machines and then build up from there whether it be Gnome or KDE or a strict web server box with no GUI. To build it up all I have to do is grab the packages I want with apt. I can roll my own distro in a way.
Not to mention Stable, Testing and Unstable are really all different distributions anyway.
What happens when the corporate backers of UserLinux decide that bills can't be met and they have to concentrate on an enterprise version? Bills don't pay themselves and there are reasons why RedHat isn't doing the consumer version anymore.
Debian, as a UserLinux, will survive the corporate onslaught precisely because it is free. Much as the Linux kernel will survive in the absence of corporate backing. That is the power of open source software.
Red Hat isn't doing a consumer version because it cannot afford to, because it must answer to shareholders, because it is commercially driven to profit. Debian suffers none of these drawbacks.
However, if Red Hat Enterprise were based on Debian, Red Hat would have minimal overhead in procuring a similar consumer version, while retaining all the benefits of a consumer presence. There is an enormous amount of work being put into the Debian distribution, and commercial entities that recognize and take advantage of it have the potential for great benefit.
It seems like there has been allot of anti-Redhat FUD lately. While I have always been a Debian fan, and I agree that every distro maker should base their distro on Debian, all this crap about Redhat leaving a hole in the consumer market because they made Redhat Linux a community project that is still heavily guided and sponsered by Redhat... that just smacks of anti-Redhat FUD.
Truth is that Redhat Linux 10 was released several days ago, and for trademark reasons it is called Fedora Core 1. Anyone who has used Redhat 8.x or Redhat 9.x, will be able to tell that Fedora Core 1 is Redhat 10.
I would love to see one internet based community developed meta-distrution of Linux, with one comprehensive package repository. This would be the Linux standard. Then companies that want to make a newbie-friendly Linux could cherry-pick the best software packages, make custom themes, and tweak everything and also provide support.
In my opinion, the thing that Redhat 8 through Fedora Core 1 do really great is that they cherry-picked a nice set of software packages, made a nice theme for the desktop, and put everything together into one nice coherent package.
Note that the good things that Redhat does with its distro do not conflict with having a Debian-foundation, and the fact that Redhat has decided to fracture the internet community because it refuses to have Fedora Core 1 be a customized Debian is just plain silly!
Other distros have shown the power of using a Debian based core: Knoppix, Libranet, and Lindows, to name 3 distros, all accomplish something slightly different.
1. Knoppix is a live CD based Linux distro with completely automatic hardware detection. Knoppix is a great toy, a great way to advertise Linux, and it makes for an uber rescue disk.
2. Libranet aims at being a general purpose desktop/server distro, and it adds value by greatly simplifying the installation and maintenance of the OS.
3. Lindows is supposed to be a newbie friendly / user-friendly Linux distro that emulates the look-n-feel of Windows. It is aimed at a large target market of casual computer users that want to save a few bucks.
So please tell me why Redhat couldn't use a Debian foundation for Fedora Core? All they had to do was create a small community layered ontop of the Debian community. Their job would be to cherry-pick software packages from the comprehensive apt repository that Debian already has, and integrate it all into one coherent system by tweaking settings and theming applications.
In conclusion, lets drop this Redhat ditched desktop Linux crap, and focus on the fact that Redhat is duplicating effort by not basing their community developed distro on Debian. It is starting to remind me of Christianity with its many demoninations.
What happens when the corporate backers of UserLinux decide that bills can't be met and they have to concentrate on an enterprise version?
My guess is that the UserLinux corporate backers are large IT *users*, not developers like Red Hat. If that is the case they don't need to make any profit on it - they want to save money by using it themselves.
Get a few big companies with hundreds of thousands of PC seats and each company's share of the investment to develop this kind of desktop distribution starts to look small compared to what they spent just on handling the latest MS virus.
Stop worrying about the risks of nuclear power and start worrying about the risks of not using nuclear power.
I said it before, and I'll say it again:& cid=6329 689
http://slashdot.org/comments.pl?sid=69340
One of the main reasons why GNU/Debian is perfect for a reference system, is that stable doesn't change that often.
Debian Woody (3.0) was released July 2002, with an update December 2002. How many version of Mandrake, SuSe, Gentoo or RedHat has come out since then?
If you are a developer, you really don't want a moving target like the other distributions. You really want to have stable target over some period of time.
Note that, even if Debian becomes the reference system, it doesn't mean that RedHat or SuSe, Gentoo can't have never libraries or KDE, or GNOME on their system. It just means that at the very least, they need compatible libraries installed by default.
And no, LSB is not enough. That is just a voluntary paper, and with no reference system, you still would have to test the major distributions to make sure your program is working.
With a working reference system, like Debian, you would only need to test against one distribution.
Je ne parle pas francais.
You also have to make it painless to do things like install/remove software and install/remove drivers.
I have been patiently trying to build up and use my Suse 8.2 system.
My biggest complaints so far?
- I don't want to have to do black magic command line crap to install my NVIDIA drivers
- Although I definitely agree with the root/user separation, its a pain in the a$$ to keep getting assaulted with a root password prompt when I want to change a system setting (flame away)
- many of the programs don't seem polished; that is, they seem to crash at odd times or don't do what they said they would when I hit 'ok'. (??)
- the interface needs to be more polished for the average user who doesn't want to understand the technical aspects of what a link is or what HDA1 is...
I LOVE that Linux exists, and I am growing to love it more....BUT...I am not an "average" user. I am somewhere in the haze between advanced Windows weenie and low level Linux novice.
I don't care how many LiveCDs you ship to my father-in-law or my wife (as examples). If they can't install drivers and programs, configure their systems, and navigate their PCs _easily_ and through the GUI _only_ you won't have an ice-cube's chance in He11 of getting them to use Linux. Oh - and if they can't buy software (games) for it at Best Buy you're screwed too.
Average users want a tool that looks pretty, does neat things, and makes their lives easier/more entertained. They don't give a rat's behind about shell scripts, Xfree, Xserve, CUPS, gcc or whatever. It just confuses them and turns them OFF to the product.
Hope you find these comments contructive - they are not meant to assault.
Is the juice worth the sqeeze?
"Maybe so, but geeks (me included) will want choice."
From the story header: "at which Bruce Perens suggested that in order to get Linux to the enterprise desktop"
Thankfully, that excludes geeks like you (and me). Business has absolutely zero to do with such small quibbles on whether or not we geeks have choice and everything to do with what is a good business decision. There are points at which the two intersect but they are not either mutually inclusive or exclusive.
Human tendency especially among the informed/educated is to think to one's condition. "That isn't what I would want/do so it will fail." That line of thinking is usually not conducive to successful businesses.
"How will the business people know that UserLinux is the one true distro any more than they know that Debian is the one true distro now?"
Because if everyone in the Linux community unites behind one version for the defacto business standard, then they will know. Keep choice for those that want it. For those that don't give a flying four-door (insert your favorite four-letter) about whether it uses KDE or Gnome or OOo/Hancom/KOffice or for that matter what the hell it runs. They just want something that works.
In fact, while I am at it, this is what businesses want (I know because I run one):
Web Browser (again, flavor DOES NOT MATTER)
-Related browsing technology (Java, Flash, etc)
-No Pop ups
E-mail
-No Spam
-Easy attachments, etc.
-Free from virus concerns
Office Software
-Maximum compatibility desired (MS Office, Word Perfect, etc.)
Custom/Special Application compatibility
-This is the big showstopper
-Like it or not, there are businesses that will NOT consider Linux because simple apps like ACT 2000 will not work
-Wine/Codeweavers integration is a must
Central Authentication/Access Control
-It may already exist. I honestly don't know.
-Must work as simple or close to Windows Domains
-Must dictate what applications/security/settings are available to users that log in
-Must provide a central point to push new apps/bug fixes
-Should NOT require physical access to a machine EVER except for system problems
Windows Update-like mechanism (See Red Hat Up2Date)
-Again, this should be automatic to the end-user but updates pushed by admins (after proper testing, etc)
-Updates should be pushable by group (so that prioritization is possible)
Integrated Virus protection/Firewalling
-Just because a plethora of Viruses/Worms are not out there now doesn't mean they won't be later
-The Firewall settings should be set upon logging in (see the above)
Hardware Support
-No, we businesses don't care if we have the source for a scanner driver as long as it is free and works
-Again, do not confuse your condition to that of a business. Free first. OSS second.
ISV/Vendor Support
-Widely document the standard system
-Provide migration documents for migrating Windows apps to Linux
-Provide documents to train programmers that are Windows-centric to think on a Linux-level
There you go. My rendition of what it needs. I am sure there is more but those are the things I have encountered in Enterprise computing that without, Linux will NEVER unseat windows in a reasonable amount of time.
Feel free to tear my assertations apart.
Whoah, GREAT insight into the problem. First you drop all properly package managed distros like Gentoo and Debian, and then you bitch about having trouble to maintain the systems or teach things to users.
Mhh, I'd suggest you go in other direction. For example, why not install Debian or Gentoo on a friends computer, and then teach them to use apt-get or emerge. Or maybe, tell him to give you a call if he ever has a new need (game, whatever). Then you can tell him: type "emerge doom3" or "emerge whatever". Of course, you'd leave sshd installed and will cron update on already installed packages.
You don't even need to explain anything to the user. Really. Now if you only had mentioned Slackware or the likes, yes, that one I will only install on some servers and make sure they are lightweight, and vulnerabilities do not pop-up. These babies work GREAT, but must be babysitted (though I've found I really know what's going on and no side effects will be expected).
unfinished: (adj.)
My perfectly average eight year old daughter uses Debian , a "non user friendly" distro according to your post with no problem at all. She prefers it to windows xp because "it has better games", "cooler menus", and "no blue screens!"
The system is a vanilla install with some extra game packs, Evolution, Mozilla, and KDE. Nothing fancy.
If my kid can do it with next to no training aside from a basic understanding of how to use a GUI based operating system, I have a hard time believing adults who have been working with computers for several years or more would have trouble with it.
When it comes down to it, people like to complain about change, so they make excuses about why they aren't adopting linux. "But it's harder to use," "It doesn't have enough software," "It doesn't have any GUI tools" are excuses I hear all the time.
If they would simply try a current version of linux, maybe use it for a week or two, they would find that none of these excuses are true of modern linux distributions.
This signature has Super Cow Powers
Just to give you a little background, I've been running Linux at home, starting with Slackware, moving to Red Hat, and trying out Debian, for 7 years. I've managed a remote datacenter of 10 Linux boxes with nothing but ssh and the occasional use of a cyclades terminal server for when things get really fubar (like the nfs server got a little hosed and i couldn't log into anything remotley). So I'm no stranger to building my own kernels and configuring hardware by hand, I just find Debian makes it so much harder than it needs to be.
She prefers it to windows xp because "it has better games"
Better games? Better games?
Go out and buy her some commercial games, you old cheapskate. Expecting a seven year old to be happy with just Solitaire verges on the abusive.