LWN daily pointed me to Jack Moffitt's critique of the BitKeeper license:
i.cantcode.com/writing/bitkeeper.html
It is worth a read. (Though it is a bit odd that he does not mention that the definition of the OSI basically explains the same differently what the FSF defined earlier. *wink* )
| With all of the restrictions that have been outlined, BitKeeper | seems almost free.
| If we amend the requirements to accomodate something that is almost | Free Software or almost Open Source, these boundaries will diminish | gradually and not only will our campaign be unsuccessful but we will | have weakened our efforts.
| BitKeeper under the terms of the BitKeeper License is neither Free | Software or Open Source software. It fails to meet the criteria and | therefore fails to meet the standards of freedom that these criteria | define and that the community have adopted. If the aim is to promote | Free Software and Open Source and preserve these rights and | freedoms, we should not be satisfied with almost free or almost open | source licenses like the BitKeeper license.
| 4. Conclusion | | Sometimes it is tempting to sacrifice our rights and freedoms for | convenience, but we should not do so. There are many problems with | CVS and other Free source management packages, and it would be nice | to move to a more robust and more well-designed tool. We are better | off to repair or fix the tools which are free, or if that is not | acceptable to create new free tools that preserve the the rights and | freedoms we enjoy.
| I also encourage Free Software hackers to avoid or cease using | BitKeeper in their own projects. It might not be as convenient to | use other tools, but in the long term we should be more concerned | with preserving those rights and freedoms we currently exercise and | enjoy daily. I personally have stopped using BitKeeper
| I encourage the entire community to support the efforts of Free and | Open Source projects in this area.
http://www.gnu.org/software/cvs/cvs.html http://subversion.tigris.org/ http://regexps.com/#arch http://prcs.sourceforge.net/ http://aegis.sourceforge.net/index.html
Bernhard Reiter bernhard@intevation.de wrote:
[...] http://subversion.tigris.org/ [...]
This seems to be where the smart money is right now, from people I know. What features does BitKeeper have that are needed to be added to subversion?
MJ Ray markj@cloaked.freeserve.co.uk wrote:
This seems to be where the smart money is right now, from people I know. What features does BitKeeper have that are needed to be added to subversion?
To answer my own question after digging Kernel Traffic [1] for a bit, it seems that a more peer-to-peer operation is the main difference. From reviewing the links, I think Aegis is a better fit for this type of development, but again I wonder why bk is the preferred choice. There are also active supporters of Aegis on lkml, but they don't seem to be as, erm, "aggressive" as bk's developers. He reminds me of someone, not just in the "lm" signature. ;_)
During the original discussion[2][3], the argument seems to have concentrated on personal attacks, ignoring substantive responses to nearly all questions put and the fact that bitkeeper was restricted software. BK then seems to have become the repository by sheer persistence.
2. http://kt.zork.net/kernel-traffic/kt19991011_38.html#5 3. http://www.uwsg.indiana.edu/hypermail/linux/kernel/9909.3/0511.html
Is there anything we can practically do to change the bk decision? It looks like bk has raised more than enough technical problems of its own. Maybe Aegis solves some of them? Aegis also have problems marketing themselves. PDF manuals, really, in this day and age! ;_)
Needing a non-free tool to be a first-class citizen on development of a Free Software project always seems a shame to me.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Is there anything we can practically do to change the bk decision? It looks like bk has raised more than enough technical problems of its own. Maybe Aegis solves some of them? Aegis also have problems marketing themselves. PDF manuals, really, in this day and age! ;_)
IMHO, its pointless persuing this - linus knows the BK guys quite well (from what it seems). And while I sympathise with people that say its isn't Free, its is free as in money, and the developers will implement any sensible ideas people suggest, and have stated several times that it is a very complex bit of software, far more complex than say the kernel. This requires a full time staff to develop it, and they are being paid to develop it by companies that require such a software - the BK people are trying to give the best of both worlds, and I personally wish them the best of luck.
But on the otherhand I realise the danger (whatif BK 'goes bad' when we are all dependant etc) but they have added clauses that it will go GPL if they go out of business, etc etc. and are trying to counter most of the usual bad points.
I think that there are a lot better fights to pick than this one - but then again I see the dangers - I'm not blind to history.
JohnFlux
On Thu, 2002-03-07 at 06:55, John Tapsell wrote:
But on the otherhand I realise the danger (whatif BK 'goes bad' when we are all dependant etc) but they have added clauses that it will go GPL if they go out of business, etc etc. and are trying to counter most of the usual bad points.
They already have gone bad - it's proprietary software. Does anyone know what format the BK changesets/archives/etc. are in? I imagine it's a format that is fairly easy to decipher, but who knows...
Relying on proprietary software in this manner is like Microsoft giving away free copies of Word to the dev team and them using it to do all the documentation in. That may seem a great deal at first sight, but they're locking their data into a proprietary application: if they ever need to get it out again, it might be difficult. Exporting a complete repository with changes, history, etc., might be well near impossible.
Unfortunately, I think the BK thing has been brewing for years - people have been pushing it for years, and it seems that much of it has been built in response to Linus' requests, so it's probably a very good technical fit for the kernel. Just not a good ethical fit :(
Cheers,
Alex.
Alex Hudson home@alexhudson.com wrote:
They already have gone bad - it's proprietary software. Does anyone know what format the BK changesets/archives/etc. are in? I imagine it's a format that is fairly easy to decipher, but who knows... [...]
Yes, one of the criticisms I have read is that these changesets are nothing like the unidiffs we are all used to and there is not (currently) a conversion tool.
Unfortunately, I think the BK thing has been brewing for years - people have been pushing it for years, and it seems that much of it has been built in response to Linus' requests, so it's probably a very good technical fit for the kernel. Just not a good ethical fit :(
Yes, but I always thought Linus had some irrational hatred of version control systems, so didn't ever consider it as a realistic threat. Larry's opinions of free software are known and rather condescending, but have proprietary marketers won? :-(
On Thu, Mar 07, 2002 at 06:55:15AM +0000, John Tapsell wrote:
Is there anything we can practically do to change the bk decision?
IMHO, its pointless persuing this
There is no other option than trying to convince people that giving up freedom gradually is a bad idea and freedom will erode.
And while I sympathise with people that say its isn't Free, its is free as in money,
Please read the critique of the licence again.
and have stated several times that it is a very complex bit of software, far more complex than say the kernel.
Do you believe this? How many codelines does BitKeeper have compared to the kernel?
On Thu, Mar 07, 2002 at 02:17:49AM +0000, MJ Ray wrote:
Is there anything we can practically do to change the bk decision?
You could do something alternative, you could use GNU/Hurd which is written by people who want to have freedom and promote free software.
Jeroen Dekkers
On Thu, Mar 07, 2002 at 02:17:49AM +0000, MJ Ray wrote:
Is there anything we can practically do to change the bk decision?
Create awareness, spread the word. LWN reports about a small pedition: http://lwn.net/2002/0307/a/bitkeeper-petition.php3 There is a summary on the kernel page: http://lwn.net/2002/0307/kernel.php3
Obvious: Test, use and Improve Free Software alternatives.
On Thu, Mar 07, 2002 at 08:41:01PM +0100, Bernhard Reiter wrote:
On Thu, Mar 07, 2002 at 02:17:49AM +0000, MJ Ray wrote:
Is there anything we can practically do to change the bk decision?
Create awareness, spread the word. LWN reports about a small pedition: http://lwn.net/2002/0307/a/bitkeeper-petition.php3 There is a summary on the kernel page: http://lwn.net/2002/0307/kernel.php3
Linus wrote an email about the subject. http://lwn.net/daily/lt-bitkeeper.php3
| In short: nobody requires BK of anybody else.
I tend to disagree strongly on the following point.
| And I personally refuse to use inferior tools because of ideology.
For me using Free Software is not about ideology, but long term pragmatism. Thus Linus weights freedom of software less important then I do.
The interesting part of the mail is that he gives pointers on what we could do to improve the situation if we want to hack on software:
| The most productive thing people could do might be to just do a | BK->CVS gateway, if you really feel like it.
This might be a good idea if it will help the transistion to CVS or other Free Software tools. It is not a good idea if it helps people to continue using BK.
He also writes: | I took a look at | subversion, and it doesn't even come close to what I wanted.
Implying that he did not look at Aegis yet. If Aegis is worth a look, we could check it regarding Linus' needs and make it easy for him to try it.
Bernhard
Bernhard Reiter bernhard@intevation.de wrote:
The interesting part of the mail is that he gives pointers=20 on what we could do to improve the situation if we want to hack on software:
This is probably the most useful thing we can do. Petitioning requests have limited impact, as it seems some people see petitions as "demands" rather than "popular requests".
| The most productive thing people could do might be to just do a | BK->CVS gateway, if you really feel like it. =20 This might be a good idea if it will help the transistion to CVS or other Free Software tools. It is not a good idea if it helps people to continue using BK.
I think it would be more beneficial to create a BK->Aegis converter, as Aegis appears (hint: I'm not using it extensively yet) to be able to use SCCS format at the back (via CSSC). I'm not keen on the idea of gatewaying, as it will allow the server to remain non-free and the data to be held there, a hostage to restricted software.
Implying that he did not look at Aegis yet. If Aegis is worth a look, we could check it regarding Linus' needs and make it easy for him to try it.
The more I read, the more this looks like the way forward. If someone more qualified/experienced than me could do a head to head comparison, it would help me immensely. Questions which are to my mind immediately: If Aegis is that old, why does subversion exist? Why does BK exist? Why aren't the Aegis folks screaming at the world in frustration yet? Why aren't we all using it?
On Fri, 8 Mar 2002, Bernhard Reiter wrote:
On Thu, Mar 07, 2002 at 08:41:01PM +0100, Bernhard Reiter wrote:
On Thu, Mar 07, 2002 at 02:17:49AM +0000, MJ Ray wrote:
Is there anything we can practically do to change the bk decision?
Create awareness, spread the word. LWN reports about a small pedition: http://lwn.net/2002/0307/a/bitkeeper-petition.php3 There is a summary on the kernel page: http://lwn.net/2002/0307/kernel.php3
Linus wrote an email about the subject. http://lwn.net/daily/lt-bitkeeper.php3
| In short: nobody requires BK of anybody else.
I tend to disagree strongly on the following point.
Yes. You are right, it's a long term investment to choose Free Software.
| BK->CVS gateway, if you really feel like it.
This might be a good idea if it will help the transistion to CVS or other Free Software tools. It is not a good idea if it helps people to continue using BK.
Another issue with BK, it's the source availability with current licensing scheme. The binary version is available with a specific licensing scheme. The source version is available with another specific license. There is some issue about auditability of the security of BK ;-) For me, I consider that, a big issue for software security. You can have 2 versions in // with multiple security level (or backdoor ?) ;-)
He also writes: | I took a look at | subversion, and it doesn't even come close to what I wanted.
Implying that he did not look at Aegis yet. If Aegis is worth a look, we could check it regarding Linus' needs and make it easy for him to try it.
I'm using CVS daily it is working but :
- rename of directory or files is a pain (you have to dig into CVSROOT) - linking is imposible and frequent in big project - sync of cvs is not so easy if you don't want a single point of failure - partial commit sometimes happens - I will not talk about the security on the cvs server (except maybe in a linux vserver ;-)
So we have look around but there is no miracle solution :
- aegis seems a really good software but the concept are quite different from CVS (maybe not so bad idea). Moving from a big CVS in aegis is not so easy (but I'm maybe wrong) license : GPL
- arch looks cool and really an innovation but software is quite young http://regexps.com/src/src/arch/=FAQS/features.html We will build a second repository with arch to move our currents projects. (some issue with protocol and FTP) license : GPL
- subversion seems not so bad. license : Apache-style
So for us, we will use CVS and arch in the same time and if arch is more stable (for our projects) than CVS, we will remove it.
arch could be an official GNU project ?
just my comments.
alx
Alexandre Dulaunoy adulau-conos@perycles.unix.be.EU.org wrote:
So we have look around but there is no miracle solution :
- aegis seems a really good software but the concept are quite different from CVS (maybe not so bad idea). Moving from a big CVS in aegis is not so easy (but I'm maybe wrong) license : GPL
BitKeeper is also not so like CVS. I'm likely to take the step in trying to move some of my CVS work into Aegis soon. A conversion guide is supplied with it. If there is interest, I report back on how it goes.
- arch looks cool and really an innovation but software is quite young
I have heard criticisms of the current arch's designs. It sounds very Tom Lord ;-) I don't know if there will be a version that the computer scientists can love too.
On Fri, Mar 08, 2002 at 01:21:18PM +0100, Bernhard Reiter wrote:
On Thu, Mar 07, 2002 at 08:41:01PM +0100, Bernhard Reiter wrote:
On Thu, Mar 07, 2002 at 02:17:49AM +0000, MJ Ray wrote:
Is there anything we can practically do to change the bk decision?
Create awareness, spread the word. LWN reports about a small pedition: http://lwn.net/2002/0307/a/bitkeeper-petition.php3 There is a summary on the kernel page: http://lwn.net/2002/0307/kernel.php3
Linus wrote an email about the subject. http://lwn.net/daily/lt-bitkeeper.php3
I don't see why I should even care about Linux and especially Linus at all. Personally, I just help finishing the GNU system instead of caring about the GNU/Linux variant. I don't see why people supporting free software want to get associated with some bunch of people promoting non-free software, use the term open source and try to get all the credits themself for writing an operating system. They don't give credits to the biggest part of the operating system, GNU, a part that is such important that without it Linux wasn't even possible.
Jeroen Dekkers
On Fri, Mar 08, 2002 at 07:26:25PM +0100, Jeroen Dekkers wrote:
On Fri, Mar 08, 2002 at 01:21:18PM +0100, Bernhard Reiter wrote:
Linus wrote an email about the subject. http://lwn.net/daily/lt-bitkeeper.php3
I don't see why I should even care about Linux and especially Linus at all.
Linus has a long time record to develop nice Free Software. It is good to keep in contact with people who do this. In general it is good to keep communication and discussions going.
Personally, I just help finishing the GNU system instead of caring about the GNU/Linux variant.
The goal of the GNU project is to give you freedom to completly run Free Software for your computing needs. The Linux kernel certainly helps this goal in a significant way. If Linux would have been available when Hurd started, the GNU project would probably have adopted it. This is standard GNU project behaviour. Because the Linux kernel is a significant contribution, a name like GNU/Linux gives credit to it and also explains that there might be other variants.
http://www.gnu.org/software/hurd/hurd-and-linux.html
| When we started developing the Hurd in 1990, the question facing us | was, ``How can we get a free kernel for the GNU system?'' There was | no free Unix-like kernel then, and we knew of no other plan to write | one. The only way we could expect to have a free kernel was to write | it ourselves. So we started.
| If we did face the question that people ask---if Linux were already | available, and we were considering whether to start writing another | kernel---we would not do it. Instead we would choose another | project, something to do a job that no existing free software can do. | | But we did start the Hurd, back then, and now we have made it work. | We hope its superior architecture will make free operating systems | more powerful.
I don't see why people supporting free software want to get associated with some bunch of people promoting non-free software, use the term open source and try to get all the credits themself for writing an operating system.
You have to keep the perspective here. Linus himself never tried to get all credits for an "Operating" system. Additionally different people have different ideas on how to promote and progress things. Linus is definatly doing good things for the Free Software-Community. We want him to keep doing this.
On Fri, Mar 08, 2002 at 07:51:44PM +0100, Bernhard Reiter wrote:
On Fri, Mar 08, 2002 at 07:26:25PM +0100, Jeroen Dekkers wrote:
On Fri, Mar 08, 2002 at 01:21:18PM +0100, Bernhard Reiter wrote:
Linus wrote an email about the subject. http://lwn.net/daily/lt-bitkeeper.php3
I don't see why I should even care about Linux and especially Linus at all.
Linus has a long time record to develop nice Free Software.
I haven't seen any piece of nice free software from Linus. No, I don't consider Linux nice and you don't want to get me started about it.
It is good to keep in contact with people who do this. In general it is good to keep communication and discussions going.
True, but I don't see why we should care about it. We could rather support the Hurd instead of putting time in convincing Linus not to use BK. Both for technical and social reasons.
Personally, I just help finishing the GNU system instead of caring about the GNU/Linux variant.
The goal of the GNU project is to give you freedom to completly run Free Software for your computing needs. The Linux kernel certainly helps this goal in a significant way.
I don't like the phrase "The Linux kernel". It's a pleonasm, it implies that Linux is more than a kernel and that you're talking about just the kernel of the bigger thing, which clearly isn't the case.
If Linux would have been available when Hurd started, the GNU project would probably have adopted it. This is standard GNU project behaviour.
From early messages talking about the Hurd Thomas (author of the Hurd) tells that they know FreeBSD will be finished sooner than the Hurd, but instead of helping FreeBSD they want to create a revolutionair system.
Also RMS contradicts itself by saying "Instead we would choose another project, something to do a job that no existing free software can do." The Hurd can clearly do a lot of things which no piece of existing (free or non-free) software can do.
You can also turn around what you say: If Linux wasn't created, the current people hacking Linux would probably be hacking the Hurd, combined with the technical better papers of the Hurd would have created a much better system. A system which can compete better with the current non-free systems. It would also promote GNU and especially the GNU philosophy a bit more.
Because the Linux kernel is a significant contribution, a name like GNU/Linux gives credit to it and also explains that there might be other variants.
True, but still a lot of people call it "Linux", including most developers of the kernel named Linux.
http://www.gnu.org/software/hurd/hurd-and-linux.html
| When we started developing the Hurd in 1990, the question facing us | was, ``How can we get a free kernel for the GNU system?'' There was | no free Unix-like kernel then, and we knew of no other plan to write | one. The only way we could expect to have a free kernel was to write | it ourselves. So we started.
| If we did face the question that people ask---if Linux were already | available, and we were considering whether to start writing another | kernel---we would not do it. Instead we would choose another | project, something to do a job that no existing free software can do. | | But we did start the Hurd, back then, and now we have made it work. | We hope its superior architecture will make free operating systems | more powerful.
I don't see why people supporting free software want to get associated with some bunch of people promoting non-free software, use the term open source and try to get all the credits themself for writing an operating system.
You have to keep the perspective here. Linus himself never tried to get all credits for an "Operating" system.
He does by calling the OS "Linux" instead of "GNU/Linux". He knows that Linux is just the kernel and not the complete OS, but still he doesn't call the OS by its right name.
Additionally different people have different ideas on how to promote and progress things. Linus is definatly doing good things for the Free Software-Community. We want him to keep doing this.
I don't see much good things. I see broken 'stable' versions of Linux, waste of manpower, promoting of non-free software, etc.
Jeroen Dekkers
On Fri, Mar 08, 2002 at 09:22:27PM +0100, Jeroen Dekkers wrote:
On Fri, Mar 08, 2002 at 07:51:44PM +0100, Bernhard Reiter wrote:
I haven't seen any piece of nice free software from Linus. No, I don't consider Linux nice and you don't want to get me started about it.
Okay. But note that we also can have _some_ competition and different opinions within the Free Software camp.
It is good to keep in contact with people who do this. In general it is good to keep communication and discussions going.
True, but I don't see why we should care about it. We could rather support the Hurd instead of putting time in convincing Linus not to use BK. Both for technical and social reasons.
Some people will consider linux and important piece of Free Software. Others will consider Hurd more important. We need more people for the hurd, though.
I don't like the phrase "The Linux kernel". It's a pleonasm, it implies that Linux is more than a kernel and that you're talking about just the kernel of the bigger thing, which clearly isn't the case.
I wanted to make sure that people recognise that "linux" is the kernel and not an operating system.
[ About the reasons to develop Hurd ]
There are good reasons to continue to develop the hurd and put it into a useful state. However it is not as urgent anymore as there are some free software kernels.
Thus there are areas, where we do not have Free Software at all. To conquer these might be more important than to develop another kernel.
You have to keep the perspective here. Linus himself never tried to get all credits for an "Operating" system.
He does by calling the OS "Linux" instead of "GNU/Linux". He knows that Linux is just the kernel and not the complete OS, but still he doesn't call the OS by its right name.
To call it "GNU/Linux" is a lot about the awareness. If everybody would knew about the importance of the GNU-Project we could give it any name. Linus knows about the importance. IIRC he thinks it is okay if people call the whole operating system GNU/Linux. He also does not care if they call it differently, though. His wish is to not deal with the political part of the question AFAIK.
The FSF Europe certainly still regards his work on linux as good.
Additionally different people have different ideas on how to promote and progress things. Linus is definatly doing good things for the Free Software-Community. We want him to keep doing this.
I don't see much good things. I see broken 'stable' versions of Linux, waste of manpower, promoting of non-free software, etc.
I'd wish that more people were interested in the Hurd or the GNUStep projects. But we do not need to downgrade the archievements of the linux developers for this to be a legitimate wish.
Bernhard
On Fri, Mar 08, 2002 at 09:41:53PM +0100, Bernhard Reiter wrote:
On Fri, Mar 08, 2002 at 09:22:27PM +0100, Jeroen Dekkers wrote:
On Fri, Mar 08, 2002 at 07:51:44PM +0100, Bernhard Reiter wrote:
I haven't seen any piece of nice free software from Linus. No, I don't consider Linux nice and you don't want to get me started about it.
Okay. But note that we also can have _some_ competition and different opinions within the Free Software camp.
I certainly agree, but still I want to tell my opinion.
Some people will consider linux and important piece of Free Software. Others will consider Hurd more important. We need more people for the hurd, though.
GNU is the most important thing and I think nobody can deny that.
I don't like the phrase "The Linux kernel". It's a pleonasm, it implies that Linux is more than a kernel and that you're talking about just the kernel of the bigger thing, which clearly isn't the case.
I wanted to make sure that people recognise that "linux" is the kernel and not an operating system.
I'm not a native English speaker, but I think "the kernel Linux" make people recognise it better.
[ About the reasons to develop Hurd ]
There are good reasons to continue to develop the hurd and put it into a useful state. However it is not as urgent anymore as there are some free software kernels.
The Hurd is technically better and it also promotes free software and GNU better. The claims of microsoft saying that free software isn't innovative would just be void, for example.
Thus there are areas, where we do not have Free Software at all. To conquer these might be more important than to develop another kernel.
The Hurd isn't just another kernel. It isn't even a kernel, it is a multiserver system running on a microkernel (gnumach). Also Linux isn't really going good IMHO (just look at the 2.4 versions). IMHO GNU/Hurd also has the potential to be the world leading operating system.
Additionally different people have different ideas on how to promote and progress things. Linus is definatly doing good things for the Free Software-Community. We want him to keep doing this.
I don't see much good things. I see broken 'stable' versions of Linux, waste of manpower, promoting of non-free software, etc.
I'd wish that more people were interested in the Hurd or the GNUStep projects.
That's basically the reason why I write mails promoting the Hurd. Especially because we lack manpower.
But we do not need to downgrade the archievements of the linux developers for this to be a legitimate wish.
I don't want to downgrade, but this is my opinion based on seeing how Linux development goes, what Linux developers are saying and reading Linux code itself. And, also an other important thing: knowing that it can be in a different (and IMO better) way. I just don't want to give too much credits to them.
Jeroen Dekkers
Jeroen Dekkers writes:
useful state. However it is not as urgent anymore as there are some free software kernels.
The Hurd is technically better and it also promotes free software and GNU better. The claims of microsoft saying that free software isn't innovative would just be void, for example.
Oh really? Make me laugh! Not that I doubt the technical superiority of the kernel Hurd, but it won't stop MS from claiming to have invented -- well -- everything (starting with Bits+Bytes, Math etc.).
Thus there are areas, where we do not have Free Software at all. To conquer these might be more important than to develop another kernel.
The Hurd isn't just another kernel. It isn't even a kernel, it is a multiserver system running on a microkernel (gnumach). Also Linux isn't really going good IMHO (just look at the 2.4 versions). IMHO GNU/Hurd also has the potential to be the world leading operating system.
Hi Jeroen,
I'd just wish, you'd cut that out. I -- in example -- am looking forward to the HURD, since I hope ist will solve some of the problems I have with the current 'state-of-the-practice' systems, like, I'd like to see a real process local, user controllable + virtual operating environment (like the fs in plan9). The Unix philosphy is a bit outdated here (just look at the difference between networking and file system, it makes me shudder), but wether Hurd can contribute to a solution here remains still to be evaluated.
In the meantime I have the wish to work on a stable + free (as freedom) system, so Linux is indispensable. I do not think, that another OS war in the free sw camp is helpful now.
That's basically the reason why I write mails promoting the Hurd. Especially because we lack manpower.
Promotion is not like the press gangs, when they where looking for sailors 200 years ago. I think you'll have to accept other peoples needs and preferences.
Regards -- Markus
On Sat, Mar 09, 2002 at 08:47:37PM +0100, M E Leypold @ labnet wrote:
Jeroen Dekkers writes:
useful state. However it is not as urgent anymore as there are some free software kernels.
The Hurd is technically better and it also promotes free software and GNU better. The claims of microsoft saying that free software isn't innovative would just be void, for example.
Oh really? Make me laugh! Not that I doubt the technical superiority of the kernel Hurd, but it won't stop MS from claiming to have invented -- well -- everything (starting with Bits+Bytes, Math etc.).
But with the Hurd we can say that we are innovative. GNU/Linux is just a clone of Unix, I don't see a lot of innovative things in it. Of course there are a lot of new and good things which weren't in Unix, but the overall principle is just the same.
Thus there are areas, where we do not have Free Software at all. To conquer these might be more important than to develop another kernel.
The Hurd isn't just another kernel. It isn't even a kernel, it is a multiserver system running on a microkernel (gnumach). Also Linux isn't really going good IMHO (just look at the 2.4 versions). IMHO GNU/Hurd also has the potential to be the world leading operating system.
I'd just wish, you'd cut that out. I -- in example -- am looking forward to the HURD, since I hope ist will solve some of the problems I have with the current 'state-of-the-practice' systems, like, I'd like to see a real process local, user controllable + virtual operating environment (like the fs in plan9). The Unix philosphy is a bit outdated here (just look at the difference between networking and file system, it makes me shudder), but wether Hurd can contribute to a solution here remains still to be evaluated.
Somebody already asked whether this was possible on our mailinglist. The Hurd could have most of the plan9 features if just somebody took the time to implement it.
In the meantime I have the wish to work on a stable + free (as freedom) system, so Linux is indispensable.
I don't deny that, I use GNU/Linux to do most of the things.
I do not think, that another OS war in the free sw camp is helpful now.
Why war? I only want to say that the Hurd also exists as an alternative to Linux.
Jeroen Dekkers
Jeroen Dekkers jeroen@dekkers.cx wrote:
I don't like the phrase "The Linux kernel". It's a pleonasm, it implies that Linux is more than a kernel and that you're talking about just the kernel of the bigger thing, which clearly isn't the case.
Please, do not get me started on this. Talking of "the Linux kernel" is no different to talking about "the Hurd kernel system" (I guess the extra word is required). After all, saying "Linux" or "Hurd" on its own is only meaningful on its own if you know that it is a kernel.
On Fri, Mar 08, 2002 at 10:49:28PM +0000, MJ Ray wrote:
Please, do not get me started on this. Talking of "the Linux kernel" is no different to talking about "the Hurd kernel system" (I guess the extra word is required). After all, saying "Linux" or "Hurd" on its own is only meaningful on its own if you know that it is a kernel.
The funny thing is that the Hurd isn't a kernel and lives completely in user-space. :)
Jeroen Dekkers
Jeroen Dekkers jeroen@dekkers.cx wrote:
The funny thing is that the Hurd isn't a kernel and lives completely in user-space. :)
No, I think it is a multiserver system running on a microkernel providing the services usually offered by an OS kernel, which is why I estimated the word "system" is required. However, since you just smile and say "it's not that" without saying what it *IS*, I remain uneducated.
On Sat, Mar 09, 2002 at 12:00:59AM +0000, MJ Ray wrote:
Jeroen Dekkers jeroen@dekkers.cx wrote:
The funny thing is that the Hurd isn't a kernel and lives completely in user-space. :)
No, I think it is a multiserver system running on a microkernel providing the services usually offered by an OS kernel, which is why I estimated the word "system" is required. However, since you just smile and say "it's not that" without saying what it *IS*, I remain uneducated.
It is what you just said. I don't know of any word with which you can refer both to a multiserver system and a monolithic kernel. We probably have to invent one. :)
Jeroen Dekkers
Jeroen Dekkers jeroen@dekkers.cx writes:
It is what you just said. I don't know of any word with which you can refer both to a multiserver system and a monolithic kernel. We probably have to invent one. :)
What about "operating system `core'"? The Hurd servers (combined with the microkernel) provide the core for the GNU operating system, just like Linux does.
moritz
On Sat, Mar 09, 2002 at 11:58:56AM +0100, Moritz Schulte wrote:
Jeroen Dekkers jeroen@dekkers.cx writes:
It is what you just said. I don't know of any word with which you can refer both to a multiserver system and a monolithic kernel. We probably have to invent one. :)
What about "operating system `core'"? The Hurd servers (combined with the microkernel) provide the core for the GNU operating system, just like Linux does.
I consider gnumach + the Hurd + glibc as the core of the GNU operating system.
Jeroen Dekkers
Jeroen Dekkers writes:
On Fri, Mar 08, 2002 at 10:49:28PM +0000, MJ Ray wrote:
Please, do not get me started on this. Talking of "the Linux kernel" is no different to talking about "the Hurd kernel system" (I guess the extra word is required). After all, saying "Linux" or "Hurd" on its own is only meaningful on its own if you know that it is a kernel.
The funny thing is that the Hurd isn't a kernel and lives completely in user-space. :)
Hardly. -- Markus
"M E Leypold @ labnet" leypold@informatik.uni-tuebingen.de writes:
The funny thing is that the Hurd isn't a kernel and lives completely in user-space. :)
Hardly.
The Hurd itself lives indeed entirely in userspace. "The Hurd" means only those servers running on _top_ of the (micro)-kernel. Sure, without a kernel, on which the Hurd could run on, it wouldn't be of much use. moritz
Jeroen Dekkers jeroen@dekkers.cx wrote:
I don't see why I should even care about Linux and especially Linus at all.
Well, this will be a high-profile snub for Free Software if it becomes established. I know Linus is not a real ideologue, but this isn't a nice message to see getting in the media. Fortunately, BK has had its share of problems in this episode, so they're not coming out totally good either. The lumps they're taking from the likes of the Ohio State Uni Open Source club may at least make people think about what they're doing, even if people don't know what a "petition" is.
Personally, I just help finishing the GNU system instead of caring about the GNU/Linux variant.
That variant is the largest user base today, I think, so it is worth caring for and trying to educate people from it.
I don't see why people supporting free software want to get associated with some bunch of people promoting non-free software, use the term open source and try to get all the credits themself for writing an operating system.
Did you enjoy the bit where emails said that BK was open source too? Nice demonstration of why that term isn't good.
They don't give credits to the biggest part of the operating system, GNU, a part that is such important that without it Linux wasn't even possible.
Well, to be fair, Linus did choose the GNU GPL and didn't name Linux himself anyway...
Let's keep the channels open, eh? Going all "high ground Hurd" isn't the solution. You've probably given me incentive to at least try it on the next new machine when it arrives.
On Fri, Mar 08, 2002 at 07:40:11PM +0000, MJ Ray wrote:
Jeroen Dekkers jeroen@dekkers.cx wrote:
I don't see why I should even care about Linux and especially Linus at all.
Well, this will be a high-profile snub for Free Software if it becomes established.
IMHO Linux is already a high-profile snub for Free Software and GNU.
I know Linus is not a real ideologue, but this isn't a nice message to see getting in the media.
True, I would rather see GNU and free software taking a much bigger role in the media instead of Linux and open source.
Fortunately, BK has had its share of problems in this episode, so they're not coming out totally good either. The lumps they're taking from the likes of the Ohio State Uni Open Source club may at least make people think about what they're doing, even if people don't know what a "petition" is.
The whole BK thing is just another reason to dislike Linux for me. (Most of the other reasons are technical)
Personally, I just help finishing the GNU system instead of caring about the GNU/Linux variant.
That variant is the largest user base today, I think, so it is worth caring for and trying to educate people from it.
IMHO GNU/Hurd is a better base to educate people from, all developers care about freedom etc. It's also directly linked with the GNU project. But I understand that GNU/Linux is the current variant which has the bigger userbase and really works better and you don't want to drop it. That why I said "personally". :)
I don't see why people supporting free software want to get associated with some bunch of people promoting non-free software, use the term open source and try to get all the credits themself for writing an operating system.
Did you enjoy the bit where emails said that BK was open source too? Nice demonstration of why that term isn't good.
I don't read the Linux mailinglist, but I can understand that it would cause a lot of problems.
They don't give credits to the biggest part of the operating system, GNU, a part that is such important that without it Linux wasn't even possible.
Well, to be fair, Linus did choose the GNU GPL and didn't name Linux himself anyway...
Let's keep the channels open, eh? Going all "high ground Hurd" isn't the solution. You've probably given me incentive to at least try it on the next new machine when it arrives.
I would also like to warn you that the current version of the Hurd isn't really useful in a production environment. There are going to happen big changes in the near future which would really improve the system.
Our biggest problem is manpower, we know the solutions to most problems the Hurd has and we know enough places to make it better, better than any piece of software available at the moment. But we need people to hack it. I just want to say there is an alternative to Linux, which is not really better physophical but also technical.
Jeroen Dekkers
Jeroen Dekkers jeroen@dekkers.cx wrote:
IMHO Linux is already a high-profile snub for Free Software and GNU.
Why? Without it, GNU softwares would not have as many users as they do.
The whole BK thing is just another reason to dislike Linux for me. (Most of the other reasons are technical)
If you're talking general about the microkernel vs hybrid-monolithic, isn't that debate still open? In *theory* micro should win, but in practice, most of the successful ones are hybrids at the moment.
If you're talking about general crap in some of the code, then I'm sure certain pieces of GNU software can be found to have the same...
["BK is open source" comment]
I don't read the Linux mailinglist, but I can understand that it would cause a lot of problems.
Me neither, but it was amusing ;-)
I would also like to warn you that the current version of the Hurd isn't really useful in a production environment. There are going to happen big changes in the near future which would really improve the system.
I guess this is why there aren't more users. Can you point me to a good starting point for finding out more, warts and all?
Our biggest problem is manpower [...]
Well, it is difficult to help without a system to help on. Does there exist something similar to Xnest or User Mode Linux to test on? Can any of the virtual machine solutions host it? Are there people in experience in running it like that?
On Fri, Mar 08, 2002 at 10:44:50PM +0000, MJ Ray wrote:
Jeroen Dekkers jeroen@dekkers.cx wrote:
IMHO Linux is already a high-profile snub for Free Software and GNU.
Why? Without it, GNU softwares would not have as many users as they do.
The whole BK thing is just another reason to dislike Linux for me. (Most of the other reasons are technical)
If you're talking general about the microkernel vs hybrid-monolithic, isn't that debate still open? In *theory* micro should win, but in practice, most of the successful ones are hybrids at the moment.
If you mean with "hybrid" a monolithic kernel with modules, that's still monolithic. Even a single server running on a microkernel is monolithic. The Hurd is a multi-server system, giving each server its own process. This gives you stability, security and makes it easier to debug it.
If you're talking about general crap in some of the code, then I'm sure certain pieces of GNU software can be found to have the same...
Oh yes, GNU tar for example. Also gnumach is very old and ugly code in a lot of places. But the Hurd itself is really clean, as is glibc.
I would also like to warn you that the current version of the Hurd isn't really useful in a production environment. There are going to happen big changes in the near future which would really improve the system.
I guess this is why there aren't more users. Can you point me to a good starting point for finding out more, warts and all?
It's a chicken-and-egg problem. We don't have enough manpower to really get the Hurd in a good state and people aren't going to use/develop it because of the problems.
I suggest you start with http://hurd.gnu.org and if you want to install the Hurd you should read http://www.debian.org/ports/hurd (note however that we are going to recompile the whole archive because of a libc ABI change, we haven't build packages for months). There is not very much documentation, but at least the code is well commented.
Our biggest problem is manpower [...]
Well, it is difficult to help without a system to help on. Does there exist something similar to Xnest or User Mode Linux to test on? Can any of the virtual machine solutions host it? Are there people in experience in running it like that?
Yes, we already had that from the beginning, it's part of the Hurd's design. Because the Hurd runs in user-space on a microkernel, you can just start a subhurd using the boot program. You can debug the Hurd servers with gdb just like any other program. You also don't need any special rights to start a Hurd server, you can just do it as a normal user.
The Hurd also runs in bochs. The last time I tried it didn't work on plex86, I will try to get it working as soon as I've time to do it.
Jeroen Dekkers
Hi!
MJ Ray markj@cloaked.freeserve.co.uk wrote:
Jeroen Dekkers jeroen@dekkers.cx wrote:
The whole BK thing is just another reason to dislike Linux for me. (Most of the other reasons are technical)
If you're talking general about the microkernel vs hybrid-monolithic, isn't that debate still open? In *theory* micro should win, but in practice, most of the successful ones are hybrids at the moment.
There are proprietary microkernel systems which are quite good (except for being non-free, of course) and even fast. QNX and BeOS are examples. But there's no production-quality free microkernel system yet.
If you're talking about general crap in some of the code, then I'm sure certain pieces of GNU software can be found to have the same...
But certainly not the Hurd. :-)
Cheers, GNU/Wolfgang
On Sat, Mar 09, 2002 at 12:39:37AM +0100, Wolfgang J?hrling wrote:
Hi!
MJ Ray markj@cloaked.freeserve.co.uk wrote:
Jeroen Dekkers jeroen@dekkers.cx wrote:
The whole BK thing is just another reason to dislike Linux for me. (Most of the other reasons are technical)
If you're talking general about the microkernel vs hybrid-monolithic, isn't that debate still open? In *theory* micro should win, but in practice, most of the successful ones are hybrids at the moment.
There are proprietary microkernel systems which are quite good (except for being non-free, of course) and even fast. QNX and BeOS are examples. But there's no production-quality free microkernel system yet.
I think you can say L4 is a production-quality microkernel, it's also free. They even hacked Linux to run on it, L4Linux. L4 also proves that microkernels *can* be fast. For the people interested, see http://www.l4ka.org/ and http://os.inf.tu-dresden.de/L4/.
JEroen Dekkers
On Fri, Mar 08, 2002 at 07:26:25PM +0100, Jeroen Dekkers wrote:
I don't see why I should even care about Linux and especially Linus at all. Personally, I just help finishing the GNU system instead of caring about the GNU/Linux variant. I don't see why people supporting free software want to get associated with some bunch of people promoting non-free software, use the term open source and try to get all the credits themself for writing an operating system.
Since when? Their product is a free kernel, they never claimed to produce a complete operating system. Don't confuse the errors of some people with the developers of the Linux Kernel.
They don't give credits to the biggest part of the operating system, GNU, a part that is such important that without it Linux wasn't even possible.
They don't because they don't produce full operating systems. They don't because the Linux Kernel doesn't require GNU in order to function. They don't because GNU and some ppl like you don't give credit to Linux.
I'm sick of this...
Regards, Luciano Rocha
On Fri, Mar 08, 2002 at 09:07:05PM +0000, Luciano Miguel Ferreira Rocha wrote:
On Fri, Mar 08, 2002 at 07:26:25PM +0100, Jeroen Dekkers wrote:
I don't see why I should even care about Linux and especially Linus at all. Personally, I just help finishing the GNU system instead of caring about the GNU/Linux variant. I don't see why people supporting free software want to get associated with some bunch of people promoting non-free software, use the term open source and try to get all the credits themself for writing an operating system.
Since when? Their product is a free kernel, they never claimed to produce a complete operating system. Don't confuse the errors of some people with the developers of the Linux Kernel.
Most, if not all, call it the OS "Linux", not giving credits to GNU. Also I've heard at least one of the main Linux hackers saying that naming it GNU/Linux is just stupid because GNU is just a little piece and you could better name it Xfree86/Perl/QT/GNU/Linux, which totally missed the point.
They don't give credits to the biggest part of the operating system, GNU, a part that is such important that without it Linux wasn't even possible.
They don't because they don't produce full operating systems.
But they refer to the full OS often enough.
They don't because the Linux Kernel doesn't require GNU in order to function.
They do, they need gcc for example. Also there doesn't exist any real-life used OS with Linux and without GNU.
They don't because GNU and some ppl like you don't give credit to Linux.
Can you name an example? I've never heard someone calling the "GNU/Linux" system "GNU".
Jeroen Dekkers
On Fri, Mar 08, 2002 at 10:16:46PM +0100, Jeroen Dekkers wrote:
Can you name an example? I've never heard someone calling the "GNU/Linux" system "GNU".
I don't think that saying Linux is badly done and technically a disaster, and ignoring everything it has done for the Free Software community, is giving credit to Linux.
On Fri, Mar 08, 2002 at 09:32:20PM +0000, Luciano Miguel Ferreira Rocha wrote:
On Fri, Mar 08, 2002 at 10:16:46PM +0100, Jeroen Dekkers wrote:
Can you name an example? I've never heard someone calling the "GNU/Linux" system "GNU".
I don't think that saying Linux is badly done and technically a disaster,
Should I lie to give Linux credit?
and ignoring everything it has done for the Free Software community, is giving credit to Linux.
I give credit to Linux for whatever they do. From my point of view they took away the people from the Hurd and caused that GNU isn't as known in the community as it should be.
Jeroen Dekkers
I don't think that saying Linux is badly done and technically a disaster,
Should I lie to give Linux credit?
Surely not. But then, Linux's success has proved you wrong.
I give credit to Linux for whatever they do. From my point of view they took away the people from the Hurd and caused that GNU isn't as known in the community as it should be.
Linux's use makes GNU to be known as well. Also, you claim that it took people from the Hurd, when they weren't on Hurd to begin with...
Linux brought people and developers to Free Software, people that didn't do that just for Hurd's sake.
Regards, Luciano Rocha
On Fri, Mar 08, 2002 at 10:06:49PM +0000, Luciano Miguel Ferreira Rocha wrote:
I don't think that saying Linux is badly done and technically a disaster,
Should I lie to give Linux credit?
Surely not. But then, Linux's success has proved you wrong.
Sure, but that doesn't make it technically good. But there is sure one thing Linus really should get credit: Open Development, i.e. "release often, release early" thing. That's one of the biggest mistakes of the Hurd, the development wasn't open in the beginning.
I give credit to Linux for whatever they do. From my point of view they took away the people from the Hurd and caused that GNU isn't as known in the community as it should be.
Linux's use makes GNU to be known as well. Also, you claim that it took people from the Hurd, when they weren't on Hurd to begin with...
You're right. But certainly everybody is hacking GNU/Linux and not GNU/Hurd. I think it's likely that those people would have been hacking GNU/Hurd now if Linux didn't exist.
Linux brought people and developers to Free Software, people that didn't do that just for Hurd's sake.
GNU/Linux did that, not Linux on its own. Sure, Linux did a big job, but not on its own and I don't believe Linux could have do it without GNU.
Jeroen Dekkers
On Fri, 2002-03-08 at 21:32, Luciano Miguel Ferreira Rocha wrote:
On Fri, Mar 08, 2002 at 10:16:46PM +0100, Jeroen Dekkers wrote:
Can you name an example? I've never heard someone calling the "GNU/Linux" system "GNU".
I don't think that saying Linux is badly done and technically a disaster, and ignoring everything it has done for the Free Software community, is giving credit to Linux.
However, you have to concede that the GNU system started YEARS before the Linux kernel.
I would not consider calling the system "Linux" ignoring everything the GNU system has done for the community if I thought Linus (and some distributors) was lacking intelligence.
As I think most of us would agree, Linus is a pretty bright guy, so he must be calling the system "Linux" instead of GNU/Linux intentionally.
Now, if it is to increase personal promotion or if because of a pet-pissoff, I don't give a damn.
I used to call the system "Linux", until I read why it should be called GNU/Linux. I thought it made sense and was only fair to share credit.
Hugs, rui
Also there doesn't exist any real-life used OS with Linux and without GNU.
I use and develop EtLinux. It's Linux, Tcl and Tcl scripts. And not necesseraly glibc (libc5 is supported, and soon uclibc). Yes, it's used in real life.
I suspect you'll hate me on the technical level. You are welcome, and you're not alone.
I've never heard someone calling the "GNU/Linux" system "GNU".
I do it pretty often, especially when I talk in public. Despite I run linux+tcl.
Moral: never say never, you got it wrong twice. And you're a little too strong in your posts.
/alessandro, who would better stay off this thread. No, he doesn't use BK.
Alessandro Rubini rubini@gnu.org wrote:
Also there doesn't exist any real-life used OS with Linux and without GNU.
I use and develop EtLinux. It's Linux, Tcl and Tcl scripts. And not necesseraly glibc (libc5 is supported, and soon uclibc). Yes, it's used in real life. I suspect you'll hate me on the technical level. You are welcome, and you're not alone.
Do it with a scheme (you can convert Tcl to Scheme mostly automatically, IIRC) and I'll not only stop hating you: I'll be impressed.
I've never heard someone calling the "GNU/Linux" system "GNU".
I do it pretty often, especially when I talk in public. Despite I run linux+tcl.
I'm not 100% without listening to the recording, but I think RMS has, but then he was talking about GNU in general, not just one variant.
Moral: never say never, you got it wrong twice. And you're a little too strong in your posts.
Indeed. I get slapped for absolutes too.
/alessandro, who would better stay off this thread. No, he doesn't use BK.
Well, thank $TARGET for that. At least there's one reason not to flame you. Have you tried Aegis?
On Fri, Mar 08, 2002 at 11:17:19PM +0100, Alessandro Rubini wrote:
Also there doesn't exist any real-life used OS with Linux and without GNU.
I use and develop EtLinux. It's Linux, Tcl and Tcl scripts. And not necesseraly glibc (libc5 is supported, and soon uclibc). Yes, it's used in real life.
Isn't libc5 dead? I looked at uclibc but I really prefer glibc and I don't think uclibc will support everything glibc supports.
I suspect you'll hate me on the technical level. You are welcome, and you're not alone.
I don't think "hate" is the right word, "disagree" might be.
I've never heard someone calling the "GNU/Linux" system "GNU".
I do it pretty often, especially when I talk in public. Despite I run linux+tcl.
Moral: never say never, you got it wrong twice. And you're a little too strong in your posts.
Well, you're probably right.
Jeroen Dekkers
Now I'm going to be a little bad, sorry for that:
On Sat, Mar 09, 2002 at 12:19:24AM +0100, Jeroen Dekkers wrote:
Isn't libc5 dead? I looked at uclibc but I really prefer glibc and I don't think uclibc will support everything glibc supports.
uclibc is a library for embedded systems. It's *supposed* to not support everything glibc does.
And speaking of prefering glibc over uclibc just for what it supports, couldn't I say the samething about Hurd Vs Linux?
Regards, Luciano Rocha
On Fri, Mar 08, 2002 at 11:31:27PM +0000, Luciano Miguel Ferreira Rocha wrote:
Now I'm going to be a little bad, sorry for that:
On Sat, Mar 09, 2002 at 12:19:24AM +0100, Jeroen Dekkers wrote:
Isn't libc5 dead? I looked at uclibc but I really prefer glibc and I don't think uclibc will support everything glibc supports.
uclibc is a library for embedded systems. It's *supposed* to not support everything glibc does.
Why is Felix bashing glibc then for being bloated? They don't even have the same purpose.
And speaking of prefering glibc over uclibc just for what it supports, couldn't I say the samething about Hurd Vs Linux?
I know a lot of things Linux doesn't support and the Hurd does. Running your own ext2 server as a normal user, ftp in the filesystem, etc. ;-))
Jeroen Dekkers
Jeroen Dekkers jeroen@dekkers.cx wrote:
I know a lot of things Linux doesn't support and the Hurd does. Running your own ext2 server as a normal user, ftp in the filesystem, etc. ;-))
If you mean using ftp as a filesystem, then there are multiple ways to accomplish that at Linux kernel module level. Please promote Hurd and not bash Linux, especially if you don't know Linux that well.
On Sat, Mar 09, 2002 at 12:03:56AM +0000, MJ Ray wrote:
Jeroen Dekkers jeroen@dekkers.cx wrote:
I know a lot of things Linux doesn't support and the Hurd does. Running your own ext2 server as a normal user, ftp in the filesystem, etc. ;-))
If you mean using ftp as a filesystem, then there are multiple ways to accomplish that at Linux kernel module level. Please promote Hurd and not bash Linux, especially if you don't know Linux that well.
I'm not really bashing linux, I'm more bashing all monolithic kernels, Linux being among them. Of course you can accomplish that at the kernel module level, however there are a lot of issues.
What happens when the ftp client code crashes? Your whole systems goes down, because it's in kernel-space. That's the reason why only an user with special permissions can use ftpfs.
In the Hurd, it lives in user-space, nothing happens when the ftpfs server crashes apart from disappearing. Any user can run ftpfs, you don't need special permission.
Jeroen Dekkers
Alessandro Rubini writes:
Also there doesn't exist any real-life used OS with Linux and without GNU.
I use and develop EtLinux. It's Linux, Tcl and Tcl scripts. And not necesseraly glibc (libc5 is supported, and soon uclibc). Yes, it's used in real life.
Hey, I like that. Cool. -- Markus
Jeroen Dekkers writes:
Most, if not all, call it the OS "Linux", not giving credits to GNU. Also I've heard at least one of the main Linux hackers saying that naming it GNU/Linux is just stupid because GNU is just a little piece and you could better name it Xfree86/Perl/QT/GNU/Linux, which totally missed the point.
I like the point. :-)
By the way: I don't call it GNU/Linux. To most of the mutants I say: Redhat, SuSe, etc. Since Debian keeps to the GPL as far as possible (or even strictly?), avoids contamination with proprietary (YAST) or undocumented (RPM) setup tools and also produces the a first distro with a Hurd kernel, I'd say, that Debian ist nearest to what could be GNU/Linux (I think they use GNU/Debian).
GNU/Linux is really a strange name, since there is no such a thing as THE STANDARD LINUX and there won't be 1 GNU system if it is successful. The time is over, when only one variant was the right one (and that is the idea of freedom).
They don't because GNU and some ppl like you don't give credit to Linux.
Can you name an example? I've never heard someone calling the "GNU/Linux" system "GNU".
Well. The name GNU sort of mutated to another meaning: (Real) Free software. This might include one ore more kernels to run the software at, but after all, a kernel is also a piece of software and who would like to take the freedom from people to put the pieces together as THEY want?
IMHO there will be never THE GNU system, only a whole continuum od GNU based systems. And that is good.
Regards -- Markus
"M E Leypold @ labnet" leypold@informatik.uni-tuebingen.de writes:
Since Debian keeps to the GPL as far as possible [...]
Debian has it's own definition of Free Software - the Debian Free Software Guidelines (DFSG). The GPL license is only one license, who fits that defintion.
(I think they use GNU/Debian).
It's 'Debian GNU/Linux', 'Debian GNU/Hurd', 'Debian GNU/BSD'..
moritz
On Sat, Mar 09, 2002 at 09:59:41PM +0100, Moritz Schulte wrote:
Debian has it's own definition of Free Software
Basically they explain the same thing as can be explained with the four freedoms in a different way. The OSI criteria which evolved from the Debian Free Software Guidelines are therefore another description of the same thing.
Both DFSG and OSI focus more on when a license fits Free Software. And then there are minor interpretation differences between all these descriptions.
For me the way to explain Free Software over the four freedoms is the best definition. Longer in existance, short and not too much focused on the license.
On Sat, Mar 09, 2002 at 09:10:40PM +0100, M E Leypold @ labnet wrote:
By the way: I don't call it GNU/Linux. To most of the mutants I say: Redhat, SuSe, etc. Since Debian keeps to the GPL as far as possible (or even strictly?), avoids contamination with proprietary (YAST) or undocumented (RPM) setup tools and also produces the a first distro with a Hurd kernel, I'd say, that Debian ist nearest to what could be GNU/Linux (I think they use GNU/Debian).
You call it Debian GNU/Linux if you are referring to that special variant/distribution of the GNU/Linux system. You call it GNU/Linux if you are referring to the OS, i.e. all variants/distributions.
I would not say Debian is the nearest of GNU/Linux. Debian is just a distribution of software, the best distribution of GNU software IMHO. But it distributes more software than only GNU/Linux. It distributes 7000 packages. It also distributes more systems, GNU/Hurd is the only other system at the moment but on GNU/*BSD is being worked on.
GNU/Linux is really a strange name, since there is no such a thing as THE STANDARD LINUX and there won't be 1 GNU system if it is successful. The time is over, when only one variant was the right one (and that is the idea of freedom).
I don't see anything wrong with it. It's the GNU system with the kernel Linux,
They don't because GNU and some ppl like you don't give credit to Linux.
Can you name an example? I've never heard someone calling the "GNU/Linux" system "GNU".
Well. The name GNU sort of mutated to another meaning: (Real) Free software. This might include one ore more kernels to run the software at, but after all, a kernel is also a piece of software and who would like to take the freedom from people to put the pieces together as THEY want?
But why not name those pieces put together by their right name?
And why is software not developed and distributed by GNU not real free software? Linux is real free software, but not GNU software. BSD is also real free software, but also not GNU software.
IMHO there will be never THE GNU system, only a whole continuum od GNU based systems. And that is good.
Yes, there is one GNU system. That's the system developed by GNU and released by GNU. AFAIK GNU already decided to use Debian for the distribution. The official GNU system is thus Debian GNU/Hurd. (We don't call this Debian GNU to avoid confusing with other GNU variants like Debian GNU/Linux and Debian GNU/NetBSD, but the Hurd is really part of GNU)
Jeroen Dekkers
On Sat, Mar 16, 2002 at 02:09:02PM +0100, Jeroen Dekkers wrote:
IMHO there will be never THE GNU system, only a whole continuum od GNU based systems. And that is good.
Yes, there is one GNU system. That's the system developed by GNU and released by GNU. AFAIK GNU already decided to use Debian for the distribution. The official GNU system is thus Debian GNU/Hurd.
Sorry, but that is not possible because the FSF can not let Debian decide what is free software and what is not. The definition of free software differs slightly between GNU and Debian.
Contrary to believe, there _will_ be THE GNU system, and if RMS gets what he wants, it will be available before this year ends. And it will not be called Debian GNU/Hurd, but it will be based on Debian GNU/Hurd, and all improvements will be done in Debian GNU/Hurd if possible (we will only do it seperately in GNU/Hurd if we can not do it in Debian for whatever reasons).
Thanks, Marcus
On Sat, Mar 16, 2002 at 10:18:27AM -0500, Marcus Brinkmann wrote:
On Sat, Mar 16, 2002 at 02:09:02PM +0100, Jeroen Dekkers wrote:
IMHO there will be never THE GNU system, only a whole continuum od GNU based systems. And that is good.
Yes, there is one GNU system. That's the system developed by GNU and released by GNU. AFAIK GNU already decided to use Debian for the distribution. The official GNU system is thus Debian GNU/Hurd.
Sorry, but that is not possible because the FSF can not let Debian decide what is free software and what is not. The definition of free software differs slightly between GNU and Debian.
Is it really that different? Maybe Debian and GNU should both agree on what's free software and what not. I don't think it would be that difficult.
Contrary to believe, there _will_ be THE GNU system, and if RMS gets what he wants, it will be available before this year ends. And it will not be called Debian GNU/Hurd, but it will be based on Debian GNU/Hurd, and all improvements will be done in Debian GNU/Hurd if possible (we will only do it seperately in GNU/Hurd if we can not do it in Debian for whatever reasons).
Debian and GNU should just cooperate IMHO.
Jeroen Dekkers
Jeroen Dekkers jeroen@dekkers.cx writes:
Sorry, but that is not possible because the FSF can not let Debian decide what is free software and what is not. The definition of free software differs slightly between GNU and Debian.
Is it really that different? Maybe Debian and GNU should both agree on what's free software and what not. I don't think it would be that difficult.
There are other issues, apart from the free software debate.
For example, RMS is very reluctant to accept crypto code, and Debian is currently in the process of finally embracing all sorts of cryptography.
On Thu, 28 Mar 2002 02:24:17 +0100, Florian Weimer said:
For example, RMS is very reluctant to accept crypto code, and Debian
Huh? I am involved in a lot of discussions with him on *which* crypto software should be used for the GNU project. So, I guess your information is a bit outdated.
Werner
Werner Koch wk@gnupg.org writes:
On Thu, 28 Mar 2002 02:24:17 +0100, Florian Weimer said:
For example, RMS is very reluctant to accept crypto code, and Debian
Huh? I am involved in a lot of discussions with him on *which* crypto software should be used for the GNU project. So, I guess your information is a bit outdated.
RMS does not want to have pervasive crypto in the GNU project because he fears that GNU distribution sites might be shut down if they offer crypto code or code with crypto interfaces. Crypto code should be concentrated in a few packages, even if this leads to technical problems.
Personally, I think this is the wrong strategy, but I can understand him.
On Thu, 28 Mar 2002 15:52:46 +0100, Florian Weimer said:
RMS does not want to have pervasive crypto in the GNU project because he fears that GNU distribution sites might be shut down if they offer crypto code or code with crypto interfaces. Crypto code should be
I agree with him. The American export rules are subject to change at any time and some time last year we had been very close to this. OTOH that is not only acrypto debate as the WIPO "agreements" are far more threatening.
Werner
On Thu, Mar 28, 2002 at 03:52:46PM +0100, Florian Weimer wrote:
Werner Koch wk@gnupg.org writes:
On Thu, 28 Mar 2002 02:24:17 +0100, Florian Weimer said:
For example, RMS is very reluctant to accept crypto code, and Debian
Huh? I am involved in a lot of discussions with him on *which* crypto software should be used for the GNU project. So, I guess your information is a bit outdated.
RMS does not want to have pervasive crypto in the GNU project because he fears that GNU distribution sites might be shut down if they offer crypto code or code with crypto interfaces. Crypto code should be concentrated in a few packages, even if this leads to technical problems.
I don't understand, if you just place the main machines in Europe or somewhere else we don't have those problems. IMHO crypto should be intergrated in the system (and I'm happy Debian finally does this).
Jeroen Dekkers
Jeroen Dekkers jeroen@dekkers.cx writes:
RMS does not want to have pervasive crypto in the GNU project because he fears that GNU distribution sites might be shut down if they offer crypto code or code with crypto interfaces. Crypto code should be concentrated in a few packages, even if this leads to technical problems.
I don't understand, if you just place the main machines in Europe or somewhere else we don't have those problems.
At least GCC, wget, Emacs, Ghostscript and the Hurd will need crypto some day, so your suggestion does work too well in practice.
On Fri, Mar 29, 2002 at 01:39:16PM +0100, Florian Weimer wrote:
Jeroen Dekkers jeroen@dekkers.cx writes:
RMS does not want to have pervasive crypto in the GNU project because he fears that GNU distribution sites might be shut down if they offer crypto code or code with crypto interfaces. Crypto code should be concentrated in a few packages, even if this leads to technical problems.
I don't understand, if you just place the main machines in Europe or somewhere else we don't have those problems.
At least GCC, wget, Emacs, Ghostscript and the Hurd will need crypto some day, so your suggestion does work too well in practice.
wget already has ssl support.
why would gcc need crypto ?
Tomasz Wegrzanowski taw@users.sourceforge.net writes:
why would gcc need crypto ?
The included Java run-time library might need it.
On Friday 29 March 2002 12:39 pm, Florian Weimer wrote:
Jeroen Dekkers jeroen@dekkers.cx writes:
RMS does not want to have pervasive crypto in the GNU project because he fears that GNU distribution sites might be shut down if they offer crypto code or code with crypto interfaces. Crypto code should be concentrated in a few packages, even if this leads to technical problems.
I don't understand, if you just place the main machines in Europe or somewhere else we don't have those problems.
At least GCC, wget, Emacs, Ghostscript and the Hurd will need crypto some day, so your suggestion does work too well in practice.
Why would gcc need crypto?
On Thu, Mar 07, 2002 at 12:40:28AM +0000, MJ Ray wrote:
Bernhard Reiter bernhard@intevation.de wrote:
[...] http://subversion.tigris.org/ [...]
This seems to be where the smart money is right now, from people I know. What features does BitKeeper have that are needed to be added to subversion?
Arch has a quick comparison chart started.
It has always stopped me (personally) from trying Aegis or cook in the past, that I could never easily find a comparison what really is different to cvs or make and a real reason to use it.
My emails suggesting to add an overview comparision at a place easy to find where never satisfactory answered. (Note: This was _years_ ago.)
Now I personally tend to try SCons over make next time I need this technology. Bernhard
Bernhard Reiter bernhard@intevation.de writes:
LWN daily pointed me to Jack Moffitt's critique of the BitKeeper license:
i.cantcode.com/writing/bitkeeper.html
The license is already outdated. Now, you have 72 hours to do the logging, and you have to upgrade your BitKeeper installation once a year (and possibly accept a new license), or you might violate one of the logging requirements.
On Thu, Mar 07, 2002 at 11:27:52PM +0100, Florian Weimer wrote:
Bernhard Reiter bernhard@intevation.de writes:
LWN daily pointed me to Jack Moffitt's critique of the BitKeeper license:
i.cantcode.com/writing/bitkeeper.html
The license is already outdated. Now, you have 72 hours to do the logging, and you have to upgrade your BitKeeper installation once a year (and possibly accept a new license), or you might violate one of the logging requirements.
Its is good that he quoted the license he wrote about then. Rapidly chaing licences are not a good thing...
On Wed, Mar 06, 2002 at 11:54:46PM +0100, Bernhard Reiter wrote: Free Software revision control software seems to be less known. Here is another one I've discovered today. It is also a software configuration management system. No idea how this compares to CVS.
Then to make the list more complete "cpcms" might be released in the future (as Alessandro pointed out). Maybe called "opencm" then. http://www.eros-os.org/pipermail/dcms-dev/2002-February/000764.html
http://www.gnu.org/software/cvs/cvs.html http://subversion.tigris.org/ http://regexps.com/#arch http://prcs.sourceforge.net/ http://aegis.sourceforge.net/index.html