Debian Etch moves ever closer, and Ian Murdock – the project’s founder – has been interviewed about Debian’s politics, its lack of strong leadership, and Ubuntu’s ever-growing fame. He feels that Debian is too enveloped in process and politics, making it impossible for anybody to make big decisions, thereby hindering the pace of development. In addition, on his weblog Murdock has announced he is joining Sun.
debian is dying.
Haha! your turn now. (a BSD user).
*BSD is is dying.
Haha! your turn now. (a Debian user).
Yes, this post was probably exactly as wothless as yours.
You’re both dying.
Haha! your turn now. (A SUSE user).
So much for my average comment score.
Your all Dying (Debian, Linspire, Mepis, Xandros, Solaris, Suse, Redhat — Ill just convert you all).
Haha, Your turn now. (An Ubuntu User).
=)
Yes, but the point was it was a joke and that you took it seriously …
Interesting comment about getting Solaris to catch up with Linux with regards useability – this is a good thing in my opinion since Linux has clearly benefited over the last few years with the advent of user-friendly distros. Note the impact of Ubuntu and OpenSuse in the more widespread acceptance of Linux as a viable platform for the average user (if such a being exists!)
Not so sure about the emphasis placed on backwards compatibility. Whilst some is obviously needed, I hope this does not come at the expense of design elegance/simplicity and innovation.
On a side note, I await my free Solaris disks with anticipation – I have not tried this OS since v2.6, at which point I really was not up to the challenge!
I speculate that what Ian has in mind for Solaris is standardizing on the “Linux” userland (including the GNU toolchain) and implementing the Linux system call interface. Essentially, make everything like a Linux distribution except for the kernel. A Linux-compatible distribution featuring a kernel with enterprise-class stability and functionality, including DTrace, Containers, and ZFS. I would not at all be surprised if Ubuntu winds up involved in some way, shape, or form.
If Sun wants to close the gap between Linux and Solaris, this the best way to leverage all of the development that happens on Linux. Phase 1: open the kernel. Phase 2: support the de-facto standard free software development environment. Phase 3: gain mindshare. And that’s the ballgame.
If they succeed in marrying the strengths of Solaris with the strengths of Linux, this could really be something. I’ve been a naysayer about OpenSolaris and it’s plan to compete head-to-head with Linux, but I’ve changed my mind… sort of. I think that the competition between Linux and Solaris will be fierce, and with Red Hat, Novell, and IBM duking it out with Sun, things should get really interesting no matter which free software kernel you prefer.
Well, I think its about time that Linux had some real competition. This can only be good for users as this might force the linux devs to do things they wouldn’t normally in-order to offer the same level of functionality. The Linux kernel is great, but there can always be improvements.
Debian’s democratic process, strict developer guidelines, and rigorous testing (and the resulting stability and quality of the OS) is what makes Debian, well, Debian.
I do think that a predictable release schedule would be a good idea–perhaps a new major version every year.
I think the Linux world needs someone who won’t release until it’s *done*. Not that I’m likely to use it myself, but I can see distinct benefits to that philosophy.
I mean, maybe they could have feature freezes more often and thus release more frequently, but there would come a point where too much would be added as soon as the release happened and it was OK to move things to a ‘to-be-stable’ tree…
The real funny thing is that Debian became too enveloped in process and politics, and grew the inability to make big decissions during Murdocks leadership.
“The real funny thing is that Debian became too enveloped in process and politics, and grew the inability to make big decissions during Murdocks leadership.”
This statement seems to be revising history at bit…I thought Ian left as the “leader guy” ten years ago? Maybe I’m missing something, I haven’t followed Debian that closely.
Not true. The problem with Debian is that it takes a really strong leader to drive the project, and Ian Murdock was the greatest leader the project ever had. Recent DPLs have been limp noodles. They don’t rise the occasion, they don’t see the big picture, and they don’t motivate the development community.
I’m not sure whether it was the democratic process that was flawed or there truly was a lack of strong leaders with an interest in leading the project. Whichever it was before, it’s now the latter. Debian can be rescued, but the kind of person that could make it happen is not likely to accept the challenge–or be accepted by the development community. Mark Shuttleworth put out feelers. He asked the Debian community whether they wanted a strong leader to whip the project back into shape. They told him they liked things the way they are. So he started Ubuntu.
I’d like to quote what Debian developer Martin Michlmayr wrote in his blog a year ago, just before the 2006 DPL elections:
‘There’s a reason we have had “weak” leaders since Bruce [Perens]. While now a large number of people think that Bruce was the best thing since sliced bread, lots of people were really pissed off back then with him commanding people around. And what was the result? A constitution that would ensure that no leader would ever have such power again. And that’s what we’re currently stuck with.’
http://www.cyrius.com/journal/2006/03/09
http://www.debian.org/devel/constitution#5
So maybe he was a wise guy leaving Debian behind. But instead of learning something, they do the same mistake again and again.
But to be true you cannot compare the beginning of Debian with the current state.
will solaris be using dpkg and apt now? That would be nice
will solaris be using dpkg and apt now? That would be nice
I doubt it. I think the package format Solaris uses is part of the UNIX standard.
However, someone could always add support for that format to dpkg
Edited 2007-03-19 22:36
Why would you need dpkg when you already have pkg-get from Blastwave?.
Why would you need dpkg when you already have pkg-get from Blastwave?
Since you ask, I’m guessing you have not used either apt or pkg-get for very long. Don’t get me wrong…I love and use blastwave.org every day. It’s made life as a Solaris administrator much better. But the difference in usability and functionality between Debian’s package management and the combination of Sun’s decrepit pkgadd + blastwave.org’s pkg-get is like night and day.
Even if pkgadd was updated to dpkg/rpm levels of functionality and pkg-get was a feature-complete clone of apt-get, you would still have the issue that blastwave.org does not fully integrate with Solaris fully since it installs its own versions of common libraries (gtk, gnome, ssl, etc.). It does this for a couple reasons. One reason is to maintain compatibility across the versions of Solaris that blastwave.org supports. Obviously, since Debian is a full OS it does not have that problem; it updates its software based on distribution release. The second reason is that Sun’s own shipped versions of standard libraries and supporting utilities are so painfully old. Blastwave.org cannot possibly hope to work with these old libraries and maintain newer versions of software.
Give both a real shot and you’ll see why, despite blastwave.org’s valuable contribution to Solaris, lots of people are very excited about today’s news and hopeful that it means an improvement in Sun’s package management and support of common open source software.
Edited 2007-03-20 00:07
between two evils I’d rather take RPM since many more packages are under RPM than dpkg.
between two evils I’d rather take RPM since many more packages are under RPM than dpkg.
Which doesn’t do much good, since they’re all pretty much Linux-centric.
…and support of common open source software.
SUN supports a lot of “common” open source software just fine. However, a lot of open source software doesn’t support Solaris. That isn’t Sun’s fault, but developers who have adopted Linuxisms instead of sticking to POSIX, UNIX, etc. standards.
SUN supports a lot of “common” open source software just fine. However, a lot of open source software doesn’t support Solaris. That isn’t Sun’s fault, but developers who have adopted Linuxisms instead of sticking to POSIX, UNIX, etc. standards.
You’ve got that right – although there are some really good bits of software, there is a lot of crap out there written by programmers who think the world revolves around Linux. It pisses *BSD ports maintainers off, Solaris package maintainers, it pretty much pisses everyone off – even Linux users who find that some of the Linuxisms used are not working as expected on all boxes.
Please for the love god, I don’t want to see Solaris compromised with crap being sprawled through from GNU/Linux land through what is right now, a pristine environment of beautiful code.
As for the original article; it talks about usability – nothing to do with GNU or anything, usability; that’ll probably include system management tools, GUI improvements, and the likes; if Solaris were deficient because of technical issues with the OS, then they would be in dire straights, but Solaris is the most technologically advanced operating system out there; the problems that exist with it are trivial that can be fixed with a little man power and community drive.
<< Please for the love god, I don’t want to see Solaris compromised with crap being sprawled through from GNU/Linux land through what is right now, a pristine environment of beautiful code. >>
Hey, I, for one, want GNU tar, GNU find, GNU ls, bash, GNU grep, GNU awk and GNU sed, GNU C Compiler in the default PATH (/bin ?).
They are so much more useable than their solaris counterparts (I use solaris everyday BTW).
And pleaaase Ian, give us dpkg and apt-get by default on solaris. That would just rock.
Edited 2007-03-20 11:38
You’re saying that as a so-called ‘Solaris user’ that going:
PATH=/usr/sfw/bin:$PATH
export PATH
In your .profile is too much?
You forgot :
alias awk=’gawk’ (not supported by SUN)
alias find=’gfind’ (not supported by SUN)
alias tar=’gtar’ (not fully supported by SUN)
alias vi=’vim’ (not supported by SUN)
…
No GNU ls.
That is too much…
Oh, and what about dpkg and apt-get ?
Hey, I, for one, want GNU tar, GNU find, GNU ls, bash, GNU grep, GNU awk and GNU sed, GNU C Compiler in the default PATH (/bin ?).
They are so much more useable than their solaris counterparts (I use solaris everyday BTW).
And pleaaase Ian, give us dpkg and apt-get by default on solaris. That would just rock.
The solaris userland is necessary due to backwards compatability requirements.
If you want Solaris with a GNU Userland, I suggest Nexenta: http://www.gnusolaris.org
Solaris isn’t really Solaris if everything is GNU…
Why can’t SUN provide the same level of convinience to solaris’ userland ?
I want << find . -mmin 3 -iname ‘toto*’ -maxdepth 2 >>, << tar zxf toto.tar.gz >>, << sed -r >>, GNU getopts, etc…
What’s backwards incompatible with that?
What’s backwards incompatible with the GNU tools?
Do you have an example?
In my opinion, GNU userland would be miles ahead of solaris’ userland if some of it’s drawbacks were fixed (the need fo info, a little lack of consistency).
But the kernel, the zones and ZFS just rock and have no GNU counterparts.
Combining a fixed GNU userland, apt and the solaris kernel would result in ‘the ultimate unix’, in my humble opinon.
I hope Ian can provide this. But the first thing to fix is the package management, enough with the patch dependancy hell!! Patching solaris IS A FU**ING MESS!!
Edit : Nexenta is a no go, I want the level of support SUN can provide. SUN’s support is very good, congrats, guys!
Edited 2007-03-21 08:00
They can, and if you’ll notice, several enhancements have been done to the Solaris userland programs over the years slowly working towards that.
And yes, some of the programs behave certain ways and even the ways the options are done are dictated by certain standards. In many cases, you are correct, they could add extra options to programs without a problem, and they have been slowly enhancing the utilities that are distributed. It just isn’t a priority in the grand scheme of things, and I don’t blame them.
ZFS, DTrace, and Zones are all far more important than “convenience functionality” which is what you are describing.
Patching Solaris is a mess? How?
I think it’s rather easy, I just click on the update manager, select all, and hit install updates. It does all the work for me. What’s so hard about that?
As far as the package management, every person has various views and opinions on package management. I will not deny that apt-get is easier, but people still complain about package management on RedHat systems where apt4rpm is available!
As for the original article; it talks about usability – nothing to do with GNU or anything, usability; that’ll probably include system management tools, GUI improvements, and the likes; if Solaris were deficient because of technical issues with the OS, then they would be in dire straights, but Solaris is the most technologically advanced operating system out there; the problems that exist with it are trivial that can be fixed with a little man power and community drive.
Yes, and I think most Solaris admins would agree that while Sun’s administration documentation is second to none, sometimes you want a tool to help you get the job done.
SUN supports a lot of “common” open source software just fine.
When I say support, I don’t simply mean providing an OS that will compile said software. I mean to actually provide an integrated sun.com software repository and a quality set of package management tools to keep the software up to date. For an example, see Red Hat’s supported package listing.
When I think of the types of projects I’ve been on these past couple years, they have all required a number of open source packages. In each case (whether or not we used Solaris or Linux), a Red Hat solution would have been both easier and quicker to deploy due to its package repository (or channel, if you will). In terms of future maintenance, the Red Hat servers will take less time since security updates will be provided through the same channel as the base OS updates…unlike Solaris where I will have to track each package’s security list and recompile and reinstall each myself.
Ian may have done wonders with Debian, but he has no experience with Solaris. He has no clue how to deal with ISVs and get them to support Solaris, he has never done any programming on Solaris.
Right now what Solaris needs is a leader who has ties to the industry – how to get Adobe to port Acrobat to Solaris x86, How to get ATI to support Solaris x86, How to get drivers ported to Solris x86. Ian is a GPL bigot – didn’t Debian want to rename Firefox because of a minor license issue?
What Solaris needs is a pragmatic guy with both Linux and Solaris knowledge…..lets see there are plenty of guys from the Solaris community who should be Chief Operating System Strategist.
at a time when debian is already dwindling due to politics, and the fact that ubuntu is still peeing in the pool. we get more bad news. well im still pulling for you Debian, you can do it.
Nah, Debian is all but dead. Too much politics, way too slow a release cycle, only guys using it on servers will stick with it. Others will migrate to more user friendly, up to date distributions. That’s my honest take. It’ll be sad to see Debian go, but in reality, they’re their own worst enemy in many ways. Dinosaurs were nice as well, but they grew extinct cos they couldn’t compete. The same is happening (sadly) to Debian now.
Dave
I’m glad WSU had NeXT, SUN, DEC, HP-UX and others for me to draw upon.
http://ianmurdock.com/2007/02/04/why-solaris-should-adopt-gplv2/
Typically, that means creating a shim so your kernel can talk to the Linux device driver layer and, of course, making sure the licenses are compatible.
Sun’s DDI is completely incompatible with Linux’s kernel API (or the lack there of). I wonder has Ian Murdock ever programmed anything – since his qualifications seem to be in Business.
Honestly, I’m somewhat shocked by the utter and complete disregard of the benefits the democratic grassroots-like process of Debian has brought upon the GNU/Linux and FOSS movement in general. Debian is the mother of the most popular GNU/Linux distribution (Ubuntu) today, and the mother to many others. Many of the very things that make Ubuntu & Co great are technical decisions done by Debian and most of these technical decision could only be made in an idealistic environment, an environment where the sentiment “let’s do what we think is right” overruled the sentiment “let’s do what works now”. Debian is to be commended for its stalwart adherence to its own goals, even if it looks quaint to differently oriented points of view.
I’m in awe of Ian Murdocks ignorance. Debian, the project, came a long way since its foundation. Despite being idealistic in many ways it still is one of the top three GNU/Linux distributions (IM’s words, not mine) with many derivates and other more “pragmatic” distributions in its waters. I can’t see how a program leading to an important and influential position over the last ten years is wrong so suddenly. Debian is where it is because it is like it is, not despite of it! I’m sad to read stuff like this from one of its founders.
IM said himself that the democratisation thing happened after he left. So your appraisal is a bit of an exaggeration.
Some other things Debian that ideally are very good, such as the “release when it’s ready” approach, don’t work in actual life.
Humans are like that.
A lot of people need deadlines to get going, and part of the fun of having a release schedule for a Linux distribution is expecting it, working your way towards rc1, testing, and encouraging yourself and the other lazy people around you to get those bugs fixed in time.
Sure, that’s not the ideal situation, because if you don’t make the deadline, you look like a fool, and have to postpone it, or deliver buggy software; people will make fun of it. And sometimes things happen that don’t fit the common schedules, so you may end up with a half-baked release if you don’t all suddenly work a lot harder.
But that’s part of the game, and with all its flaws, such a system works better than Debian’s. Not because it is better, but because it fits human nature a bit better.
So, maybe Debian needs a yearly.. D-Day.
h3rman: IM said himself that the democratisation thing happened after he left. So your appraisal is a bit of an exaggeration.[/i]
Yes, you are right. But I would like to point out that I never said Murdock himself started the “democratisation”. I just think that Debian is what it is because of its focus on democracy, ideals and due process and not because it was founded by Ian Murdock.
Some other things Debian that ideally are very good, such as the “release when it’s ready” approach, don’t work in actual life.
Humans are like that.
A lot of people need deadlines to get going, and part of the fun of having a release schedule for a Linux distribution is expecting it, working your way towards rc1, testing, and encouraging yourself and the other lazy people around you to get those bugs fixed in time.
Yes again, that’s my point. It may be that Debian’s program is not very workable under a point of view that works under different set of assumptions, but the Debian program still brought Debian to where it is today. And nobody can argue that Debian is an influential, perhaps even indispensable part of the FOSS movement. While other distributions may outperform Debian under certain benchmarks today, and maybe always and forever will, we can’t argue that Debian according to other benchmarks is very successful if not leader of the pack. This has made Debian to the foundation many other distributions rely on. Debian has its own way of doing things and I argue that this is exactly the reason why that’s the case. Debian is not fancy or fast, it is reliable, steady and free and I personally find this refreshingly different.
EDIT: fixed tags
Edited 2007-03-20 00:26
Faith is the death of open-source, try reasoning. Maybe Linux matured somewhat, but most of it’s true believers didn’t make the switch from juvenile to adult.
>I’m in awe of Ian Murdocks ignorance.
People who live in reality call this mature.
Guys like Dennis of BLastwave, John Grunvald, Rich Teer, Juergen, Schilly etc have been working hard on Solaris and Simon Phipps goes and hires a Linux fan boy to make decisions (technical ones) that basically will no doubt make Solaris a second class citizen again.
I really don’t get Sun. I thought they would take a stand against Linux and pour good resources into make Solaris work better than MacOS
Simon Phipps goes and hires a Linux fan boy to make decisions (technical ones) that basically will no doubt make Solaris a second class citizen again.
Sorry, I didn’t see where it said Simon hired him. I’m pretty certain lots of people or a different person than Simon made that decision. At this point, I think it’s rather silly to say it was a good or bad one. Let’s wait and judge him by his actions at Sun, shall we?
For those keeping up with Debian politics;
Who do you think has a good platform for the Debian Project Leader Election?
The announcement that Ian Murdock is joining Sun has attracted praise from Mark Shuttleworth, and in that Linux Format interview Murdock said that so far as he is concerned, Ubuntu “is” Debian. So it’s not exactly rocket science to figure that a lot of folks might like an Ubuntu user land merged with a Sun base. Sun gets a primo installation to compete with Red Hat or Novell; Ubuntu gets the commercial clout it would otherwise have to spend years building for itself. And Debian? Well I guess the Chinese put in a bid for the mineral rights, considering the number of (alleged) nails now in its coffin.
But then who knows what will happen. Change in all around I see, etc. IM has done a great deal for F/OSS and is now set to do a great deal more, if this new venture pans out. He deserves all best wishes, imho.
“So it’s not exactly rocket science to figure that a lot of folks might like an Ubuntu user land merged with a Sun base.”
hmm reminds me of nexenta: http://www.gnusolaris.org
aren’t they trying to do exactly the same?!
cheerZ
dean
Whether the Debian developer community accepts it or not, Ubuntu has become the rudder on their ship: Ubuntu develops a graphical installer, Debian decides to finish theirs finally; Ubuntu develops Upstart, Debian merges it in ….
Now, you see TONS of things go from Debian into Ubuntu, but there are certain things that would NEVER have gotten done (let alone even started) in Debian, that Ubuntu does instead, and then when it’s working great, Debian developers say to themselves, “Geez, that really is cool,” and then they merge it in.
That’s how open source works! Both projects benefit from each other, just as even Fedora benefits from openSUSE.
So Ubuntu may be like the rudder, but what’s a rudder without a ship??!! We need Debian because the arguing is a healthy side-effect of a type of project needed in OSS, the software democracy. But … that’s not to say a benevolent dictatorship can’t accomplish much, and then the democracy will take what’s good.
Edited 2007-03-20 18:18