[opensuse-project] Feedback for openSUSE 10.2 and preparing for 10.3
We've briefly discussed during the last IRC meeting about what went bad in 10.2 and how we can do this better for future products, especially 10.3. I'm mainly interested in process feedback - and not on feedback that package x is broken (unless that shows a process problem). So, let's discuss what we can do better for our next distro, Andreas -- Andreas Jaeger, aj@suse.de, http://www.suse.de/~aj/ SUSE Linux Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
Tirsdag 16 januar 2007 18:10 skrev Andreas Jaeger:
I'm mainly interested in process feedback - and not on feedback that package x is broken (unless that shows a process problem).
So, let's discuss what we can do better for our next distro,
Great idea ;-) Here are a few points from me (some of which I already mentioned on IRC) 1) Planning of new features: I don't know how much planning is done wrt. new features, but I think it would be great if there could be discussions with the community early in the process, about which new features should be high priority for the development cycle. Some of the new features on like zypper, I only found out about after it had been in factory/betas for a while. Same for the sudo yast module, if I'm not mistaken it has some bugs, and I guess noone noticed it or knew about, and hence it wasn't tested much. I know we have the commit-list, but that's rather demanding to follow. 2) Bleeding edge vs. stability: I realize that we have certain obligations wrt. to testing experimental stuff for SLE, but I guess since the box with installation support and all, is still on offer, openSUSE is supposed to reasonably solid also. I don't want openSUSE to become Debian Sarge, but I think we could move the balance a little bit towards stability and still fulfill our obligations wrt. to hardening stuff for SLE and without requiring very many resources either. I think we can be a (little) bit more conservative when selecting versions to include, especially of core stuff, like xorg, and gcc which are both unstable releases. Other than the lack of fglrx-driver for a couple of weeks I'm not aware of issues because of it, but it certainly does not inspire confidence in our beloved distro. And I think that just 2-3 weeks of additional testing after feature freeze could do a lot for the overall impression people get of using the distro. 3) Translation This may be mostly a personal issue, being a tier2 translator, but iirc the translations weren't actually included in the distro, til after translation freeze, so translators had little or no opportunity to "test" their translations. Of course this is not how things were planned.. 4) What to test I think the devs should have a pretty good idea what might need extra testing. Let us know about it. Maybe the individual devs working with bleeding edge stuff/packages could drop a quick mail to factory list: "Hi guys, I'm working on this, it's new/changed a lot, please pound on it". This does not only apply to "in-house-development", but also new cups and other upstream projects, when they do big updates. All in all, I was pretty happy with the 10.2 process. But of course there's room for some improvement and some adjustments. I'm especially pleased about stuff like kickoff, zypper and opensuse-updater, which demontstrates that not everything revolves around SLE and that openSUSE and the community still holds some weight after all. Martin --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Tirsdag 16 januar 2007 21:54 skrev Martin Schlander:
So, let's discuss what we can do better for our next distro,
Of course I just remembered something else... Any chance of Novell (sponsored) projects taking openSUSE roadmap/feature freeze into accord a little bit? If I'm not mistaken mono did a new release very shortly after our feature freeze. We shipped banshee which is broken wrt. support for the evil iPods, shortly thereafter a fixed version is released. Our Compiz is 0.2.2 - a couple of months later they're at 0.3.6, there are lots more plugins. I'm not a big fan of any of those projects, but I still find it surprising that there seems to be no consideration/coordination with openSUSE whatsoever. Martin --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
I can only comment on compiz: On Jan 16, 07 22:20:47 +0100, Martin Schlander wrote:
Our Compiz is 0.2.2 - a couple of months later they're at 0.3.6, there are lots more plugins.
That's it, a couple of months later. A couple of months earlier compiz didn't even have a version number. This project is moving fast now. 0.2.2 was the latest version we could stabilize and patch for os10.2 Matthias -- Matthias Hopf <mhopf@suse.de> __ __ __ Maxfeldstr. 5 / 90409 Nuernberg (_ | | (_ |__ mat@mshopf.de Phone +49-911-74053-715 __) |_| __) |__ R & D www.mshopf.de --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Great post! As a regular "advanced user" If I may, I agree with 95% of it. Mainly The 2nd part: On Ut, 2007-01-16 at 21:54 +0100, Martin Schlander wrote:
2) Bleeding edge vs. stability: I realize that we have certain obligations wrt. to testing experimental stuff for SLE, but I guess since the box with installation support and all, is still on offer, openSUSE is supposed to reasonably solid also. . . .
Anyway, to these days I had no oportunity to say that IMHO 10.2 release is great step forward in "releasing process" and that great work done on 10.2 is motivating for other users like me to get involved. Jozef Peterka AKA KaiSVK --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Martin Schlander schrieb:
I think we can be a (little) bit more conservative when selecting versions to include, especially of core stuff, like xorg, and gcc which are both unstable releases. Other than the lack of fglrx-driver for a couple of weeks I'm not aware of issues because of it, but it certainly does not inspire confidence in our beloved distro.
gcc in openSUSE 10.2 is not an unstable release. Don't be fooled by the package name. Novell has its own branches in the gcc svn (just like Red Hat, Apple and other vendors) and makes its releases from there. Consider gcc as distributed in openSUSE 10.2 the "stable Novell release". Calling it "4.1.2_20061115" or "gcc (GCC) 4.1.2 20061115 (prerelease) (SUSE Linux)" is just for orientation purposes so that people know what FSF gcc release it corresponds to, but strictly speaking, SUSE gcc is not FSF gcc. gcc has been handled this way for a long time and it's good because otherwise it's hard to identify regressions early enough to fix them. Overall I'm very satisfied with how the packages are updated or not, I think that users should rather start testing earlier (long _before_ feature freeze - starting with an RC is too late) instead of doing the feature freeze earlier. This needs some support from the community, users should be encouraged to test earlier. Andreas Hanke --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Wednesday 2007-01-17 at 10:19 +0100, Andreas Hanke wrote:
Overall I'm very satisfied with how the packages are updated or not, I think that users should rather start testing earlier (long _before_ feature freeze - starting with an RC is too late) instead of doing the feature freeze earlier. This needs some support from the community, users should be encouraged to test earlier.
Mmm. Some users may do that, but others may not. I don't even know how to install a pre-RC release. Are there isos? If there aren't, then you can not increase much the number of betatestesters to the masses. IMO, you should instead slow the RC phase so that normal users can test that phase longer. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFFrfSLtTMYHG2NR9URAuYRAJ98imo5NPS1yI+mynwTVBewsmzkvgCeLuE9 uL86VMKBeH0qrslRpyD+GBA= =22f2 -----END PGP SIGNATURE----- --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On 17-01-2007 at 12:03, "Carlos E. R." <robin.listas@telefonica.net> wrote: -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
The Wednesday 2007-01-17 at 10:19 +0100, Andreas Hanke wrote:
Overall I'm very satisfied with how the packages are updated or not, I think that users should rather start testing earlier (long _before_ feature freeze - starting with an RC is too late) instead of doing the feature freeze earlier. This needs some support from the community, users should be encouraged to test earlier.
Mmm. Some users may do that, but others may not. I don't even know how to install a pre-RC release. Are there isos? If there aren't, then you can not increase much the number of betatestesters to the masses.
IMO, you should instead slow the RC phase so that normal users can test that phase longer.
For 10.2, there were ISOs around from the beginning of Alpha, going through to Beta up to RCs. So there ARE indeed ISOs available. for the very brave, is alwys the availability of the Factory Tree (not yet suggested when using gnome and you want to use the comp, as the move from /opt/gnome to /usr gives some 'nice' problems as user) The RCs are mainly there to fix big errors but they should be as close to the final as possible. So it is unevitable that more users go with Alpha / Beta versions. My support is granted for this project and I stay with Factory and Snapshots. Dominique --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Carlos E. R. schrieb:
Mmm. Some users may do that, but others may not. I don't even know how to install a pre-RC release. Are there isos? If there aren't, then you can not increase much the number of betatestesters to the masses.
Of course there are ISOs. There are ISOs from the first Alpha on!
IMO, you should instead slow the RC phase so that normal users can test that phase longer.
I'm very much against this because it will result in something like what we had with the 10.1 release, where users complained and complained and complained because the distro was released with KDE 3.5.1 packages (+backported patches!) when KDE 3.5.3 was already available. I dare saying that most users really want to have recent packages, especially for highly "visible" software like KDE, GNOME, and the kernel. We should try to make more testers participate in the development rather than slowing down the development for everyone. That way we would just encourage users to start testing even later and all users get fewer features because there is less time to implement them. There is no law that testing has to start after feature freeze. Whenever users find an annoying bug in the release, they tend to say that it could have been prevented if the release had been delayed. But that's not systematic, as whenever such a bug is fixed in a newer upstream version, they tend to say that the newer upstream version should have made it into the release which isn't possible with an earlier feature freeze. Andreas Hanke --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Hi all, I know that his is a stupid idea, but why don't make the 10.3 artworks like the 9.x (9.1 especially) ones? The YaST theme was very beutifull, and the kwin decoration, the defaults icons and wallpaper, everythink was really and very special. This is my Idea. Vincenzo --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On Tuesday 16 January 2007 22:54, Martin Schlander wrote:
2) Bleeding edge vs. stability: I realize that we have certain obligations wrt. to testing experimental stuff for SLE, but I guess since the box with installation support and all, is still on offer, openSUSE is supposed to reasonably solid also.
I don't want openSUSE to become Debian Sarge, but I think we could move the balance a little bit towards stability and still fulfill our obligations wrt. to hardening stuff for SLE and without requiring very many resources either.
I think we can be a (little) bit more conservative when selecting versions to include, especially of core stuff, like xorg, and gcc which are both unstable releases.
I disagree on this. If anyone wants stability, there's SLED10. And it has fast update servers too, no problems with mirrors or speed. An openSUSE release will feel like an "old baby" if it has old components (kernel, X, KDE etc). Looks bad. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Silviu Marin-Caea wrote:
On Tuesday 16 January 2007 22:54, Martin Schlander wrote:
2) Bleeding edge vs. stability: [...] I don't want openSUSE to become Debian Sarge, but I think we could move the balance a little bit towards stability and still fulfill our obligations wrt. to hardening stuff for SLE and without requiring very many resources either. [...]
I disagree on this. If anyone wants stability, there's SLED10. And it has fast update servers too, no problems with mirrors or speed.
An openSUSE release will feel like an "old baby" if it has old components (kernel, X, KDE etc). Looks bad.
I disagree with you and fully agree with Martin's opinion quoted above. SLED is not an alternative to openSUSE. I think the release should focus a bit more on stability than cutting-edge software. With the build service in place and available software repositories for KDE, kernel, etc., you can always upgrade your stable release to newer components if you like. But forcing everybody right from the beginning (release) to use cutting-edge versions of software is not the way to go! You should realize that openSUSE is not a playground and test bed for freaks but used on many production systems, for instance in small and mid-sized companies, or even at home - those installations are actually used for work and have to function! --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Thomas Hertweck wrote:
Silviu Marin-Caea wrote:
On Tuesday 16 January 2007 22:54, Martin Schlander wrote:
2) Bleeding edge vs. stability: [...] I don't want openSUSE to become Debian Sarge, but I think we could move the balance a little bit towards stability and still fulfill our obligations wrt. to hardening stuff for SLE and without requiring very many resources either. [...] I disagree on this. If anyone wants stability, there's SLED10. And it has fast update servers too, no problems with mirrors or speed.
An openSUSE release will feel like an "old baby" if it has old components (kernel, X, KDE etc). Looks bad.
I disagree with you and fully agree with Martin's opinion quoted above. SLED is not an alternative to openSUSE. I think the release should focus a bit more on stability than cutting-edge software. With the build service in place and available software repositories for KDE, kernel, etc., you can always upgrade your stable release to newer components if you like. But forcing everybody right from the beginning (release) to use cutting-edge versions of software is not the way to go! You should realize that openSUSE is not a playground and test bed for freaks but used on many production systems, for instance in small and mid-sized companies, or even at home - those installations are actually used for work and have to function!
I agree with the latter, too. The Build Service repositories offer a broad choice of bleeding-edge versions of many packages and subsystems. That should be the place for the latest X.org, KDE, GNOME, Apache, PHP, etc... A sane balance is what is needed. Not too old stuff, but not too risky either. Also wrt the reputation of openSUSE (which is very important in order to get customers for SLED/SLES IMO). If openSUSE is too broken, there won't be as much confidence in SLED/SLES as if openSUSE was really solid (but still shipping recent versions of packages.. though maybe the the very latest bleeding edge release candidate). cheers - -- -o) Pascal Bleser http://linux01.gwdg.de/~pbleser/ /\\ <pascal.bleser@skynet.be> <guru@unixtech.be> _\_v The more things change, the more they stay insane. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux) iD8DBQFFtSh+r3NMWliFcXcRAqY1AJ9z3GQTDU9+vWBaUrfzhDH4jdfiHgCeOqLK 8gSIWzjFZQa0iqWr0Urdc4s= =r8dE -----END PGP SIGNATURE----- --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Pascal Bleser wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Thomas Hertweck wrote:
Silviu Marin-Caea wrote:
On Tuesday 16 January 2007 22:54, Martin Schlander wrote:
2) Bleeding edge vs. stability: [...] I don't want openSUSE to become Debian Sarge, but I think we could move the balance a little bit towards stability and still fulfill our obligations wrt. to hardening stuff for SLE and without requiring very many resources either. [...] I disagree on this. If anyone wants stability, there's SLED10. And it has fast update servers too, no problems with mirrors or speed.
An openSUSE release will feel like an "old baby" if it has old components (kernel, X, KDE etc). Looks bad. I disagree with you and fully agree with Martin's opinion quoted above. SLED is not an alternative to openSUSE. I think the release should focus a bit more on stability than cutting-edge software. With the build service in place and available software repositories for KDE, kernel, etc., you can always upgrade your stable release to newer components if you like. But forcing everybody right from the beginning (release) to use cutting-edge versions of software is not the way to go! You should realize that openSUSE is not a playground and test bed for freaks but used on many production systems, for instance in small and mid-sized companies, or even at home - those installations are actually used for work and have to function!
I agree with the latter, too. The Build Service repositories offer a broad choice of bleeding-edge versions of many packages and subsystems. That should be the place for the latest X.org, KDE, GNOME, Apache, PHP, etc...
A sane balance is what is needed. Not too old stuff, but not too risky either. Also wrt the reputation of openSUSE (which is very important in order to get customers for SLED/SLES IMO). If openSUSE is too broken, there won't be as much confidence in SLED/SLES as if openSUSE was really solid (but still shipping recent versions of packages.. though maybe the the very latest bleeding edge release candidate).
cheers - -- -o) Pascal Bleser http://linux01.gwdg.de/~pbleser/ /\\ <pascal.bleser@skynet.be> <guru@unixtech.be> _\_v The more things change, the more they stay insane. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux)
iD8DBQFFtSh+r3NMWliFcXcRAqY1AJ9z3GQTDU9+vWBaUrfzhDH4jdfiHgCeOqLK 8gSIWzjFZQa0iqWr0Urdc4s= =r8dE -----END PGP SIGNATURE----- --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
I vote for stability too... /J -- "In theory, there is no difference between theory and practice. But, in practice, there is." - Jan L.A. Van De Snepscheut --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On Mon, Jan 22, 2007 at 10:11:26PM +0100, Pascal Bleser wrote:
I agree with the latter, too. The Build Service repositories offer a broad choice of bleeding-edge versions of many packages and subsystems. That should be the place for the latest X.org, KDE, GNOME, Apache, PHP, etc...
Well, did you see specific problems resulting from eager updates of the mentioned components or is this yet another philisophical discussion on this list? Robert -- Robert Schiele Dipl.-Wirtsch.informatiker mailto:rschiele@gmail.com "Quidquid latine dictum sit, altum sonatur."
How about this.... what if instead of offering 10.3 it was offered as 10.2.4 , the four representing the fourth quarter of the year, then offering 10.3.1 as the first quarter or first release of this year? then the any bugs in 10.2.4 could be worked out before 10.3.1 and then 10.3.1 would be the stable for the year and carry the 18 months of support that is offered. that way more releases could occur in the factory update channel with less worry about ISO's. James --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On Mon, 22 Jan 2007, James Tremblay wrote:
that way more releases could occur in the factory update channel with less worry about ISO's.
Who would provide security and other high priority fixes for those additional releases? Gerald --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On Tue, 2007-01-23 at 07:36 +0100, Gerald Pfeifer wrote:
On Mon, 22 Jan 2007, James Tremblay wrote:
that way more releases could occur in the factory update channel with less worry about ISO's.
Who would provide security and other high priority fixes for those additional releases?
Gerald
There would be none, the additional or quarterly updates would truly be Beta's, only the the first quarter release would carry any support, that way if the "consumer" wants long term support they install only the ISO'd first quarter release, and if they want real support they buy SLED. I do. when I say first quarter, it is only and example it could just as well be the mid summer release. as long as the idea is that everyone knows that OpenSUSE is first stable and then innovative, while SLED is "Hardened and Secure". --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Am Tuesday 23 January 2007 02:43 schrieb James Tremblay:
How about this.... what if instead of offering 10.3 it was offered as 10.2.4 , the four representing the fourth quarter of the year, then offering 10.3.1 as the first quarter or first release of this year? then the any bugs in 10.2.4 could be worked out before 10.3.1 and then 10.3.1 would be the stable for the year and carry the 18 months of support that is offered. that way more releases could occur in the factory update channel with less worry about ISO's.
Well, if we do more releases, people will want also new features inside of it (and maybe just because the new version has always also a bug fixed, so "you have to update !") At the end we will have less stable releases, because we do not have the time to test them anymore. And there is also no time to work on larger changes... bye adrian -- Adrian Schroeter SUSE Linux Products GmbH, Maxfeldstr. 5, 90409 Nuernberg, Germany email: adrian@suse.de --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On Monday 22 January 2007 21:41, Thomas Hertweck wrote:
I disagree with you and fully agree with Martin's opinion quoted above. SLED is not an alternative to openSUSE.
No, SLED is not an alternative, it's a product on its own and one of its main goals is stability. It has a much more appropriate mojo to provide that compared to openSUSE. Yeah it would be so fabulous if openSUSE would be stable and polished, have super new software and nice glitz, but it's damned *HARD* to achieve, technically and financially. SLED has the stability and polish, and reasonably new software. openSUSE has to be the "playground and test bed for freaks". This advances the distro.
But forcing everybody
What forcing? Hello! Who forced you?
You should realize that openSUSE is not a playground and test bed for freaks but used on many production systems, for instance in small and mid-sized companies, or even at home - those installations are actually used for work and have to function!
Then, wrong product in the wrong place. If you want stability for production, SLED is the right thing to run. I bet you would hesitate recommending "Windows XP Home" instead of full XP to any small company, so why do you have such a problem with SLED? --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Am Tuesday 23 January 2007 09:29 schrieb Silviu Marin-Caea:
On Monday 22 January 2007 21:41, Thomas Hertweck wrote:
I disagree with you and fully agree with Martin's opinion quoted above. SLED is not an alternative to openSUSE.
No, SLED is not an alternative, it's a product on its own and one of its main goals is stability. It has a much more appropriate mojo to provide that compared to openSUSE. Yeah it would be so fabulous if openSUSE would be stable and polished, have super new software and nice glitz, but it's damned *HARD* to achieve, technically and financially. SLED has the stability and polish, and reasonably new software. openSUSE has to be the "playground and test bed for freaks". This advances the distro.
hey, no. We really try to keep openSUSE Factory already always usable and move every experimients to a OBS project first. Additionally we do Alpha and Beta releases to improve the stability. So, I consider openSUSE a stable and production ready to use product (if you do not think so, we might need to find out the reasons and to work on that). The difference to SLED is basically that it is not exclusivly focused on desktop usage and that it will not receive a 7 years maintainance (because this is a lot of work, esp. in the last years). bye adrian -- Adrian Schroeter SUSE Linux Products GmbH, Maxfeldstr. 5, 90409 Nuernberg, Germany email: adrian@suse.de --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Tirsdag 23 januar 2007 09:29 skrev Silviu Marin-Caea:
Yeah it would be so fabulous if openSUSE would be stable and polished, have super new software and nice glitz, but it's damned *HARD* to achieve, technically and financially.
I don't think anyone is expecting SLED-like stability, just a little better, so that we could actually recommend and install openSUSE with our mothers and other n00bs and they won't have to suffer all those little annoyances that though small, makes it impossible for Joe Sixpack to use it. However maybe the bugs don't come from using bleeding edge software, but from lack of testing. It'd prolly be easier for the devs to see what "kinds" of bugs we have and what it would take to solve the quality problems we have.
so why do you have such a problem with SLED?
I haven't used SLED, but as I understand it's not really geared towards home users, wrt. package availability etc., and if I'm not mistaken for example kdewebdev3 is not in it (!!). That means users have to mess with 10.1 repos, and be very careful which packages to install and update and such, not exactly an easy polished experience for Joe Sixpack. But I'm hoping Novell, or others, will soon target the home user market in a serious way, I really hope there'll be a "SUSE Linux Home Desktop 11". I personally love my openSUSE, but I don't think it can be considered a viable solution for non-geek home users currently, and I think it should be. For the sake of Linux, Novell and the world in general. However I don't think we're so far off, like I stated earlier, I really think just 2-3 weeks of additional testing and a little bit of conservatism could work wonders. Martin --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Martin Schlander <suse@linuxin.dk> writes:
However I don't think we're so far off, like I stated earlier, I really think just 2-3 weeks of additional testing and a little bit of conservatism could work wonders.
We'll add for 10.3 two more weeks of beta testing - but I would prefer if folks start testing with Alpha1 ;-) Andreas -- Andreas Jaeger, aj@suse.de, http://www.suse.de/~aj/ SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
Tirsdag 23 januar 2007 11:34 skrev Andreas Jaeger:
We'll add for 10.3 two more weeks of beta testing - but I would prefer if folks start testing with Alpha1 ;-)
Sounds great, in return I'll start testing earlier ;-) I did start with alpha4 on 10.2, though. Martin --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Andreas Jaeger wrote:
Martin Schlander <suse@linuxin.dk> writes:
However I don't think we're so far off, like I stated earlier, I really think just 2-3 weeks of additional testing and a little bit of conservatism could work wonders.
We'll add for 10.3 two more weeks of beta testing - but I would prefer if folks start testing with Alpha1 ;-)
so you should be less conservative on the doc :-) don't say all the time that the alpha release can't be used for work. very little tests can be done offwork. Real bugs come when one use the app as he does all the time. so to test one needs to make his data at risk. I know it's very difficult to do, but we should define alpha/beta states from the data point of view. I give an example: I use very often the nightly built of seamonkey. I can't wait for the improvements :-). only _once_ in ten years or so I had a lost data (and I have usually fresh backups). All the others problems (like right now) are just seamonkey crashing suddenly or the windows menu not opening the window like it should: nothing harming. in such state many people can start testing. when is this state for openSUSE? alpha? beta? factory? I have a test 10.2 install. May I make factory the update folder and go on? jdd -- http://www.dodin.net Votez pour nous, merci - vote for us, thanks :-) http://musique.sfrjeunestalents.fr/artiste/Magic-Alliance/ http://photo.sfrjeunestalents.fr/artiste/jddphoto/ --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On Tue, 23 Jan 2007, jdd sur free wrote:
All the others problems (like right now) are just seamonkey crashing suddenly or the windows menu not opening the window like it should: nothing harming.
in such state many people can start testing.
when is this state for openSUSE? alpha? beta? factory? I have a test 10.2 install. May I make factory the update folder and go on?
I have one system that is exactly what I did. As soon as the factory was open I started updating this machine. I test daily. I only report the really bad bugs with factory. But I think if more people tested factory earlier we would have a better product. -- Boyd Gerber <gerberb@zenez.com> ZENEZ 1042 East Fort Union #135, Midvale Utah 84047 --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Andreas Jaeger wrote:
We'll add for 10.3 two more weeks of beta testing - but I would prefer if folks start testing with Alpha1 ;-)
Andreas, Any idea when that will be? The roadmap at http://en.opensuse.org/Roadmap still refers to 10.2 as in the future... -- Cheers, Richard (MQ) --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
"richard (MQ)" <osl@16hd.freeserve.co.uk> writes:
Andreas Jaeger wrote:
We'll add for 10.3 two more weeks of beta testing - but I would prefer if folks start testing with Alpha1 ;-)
Andreas,
Any idea when that will be? The roadmap at http://en.opensuse.org/Roadmap still refers to 10.2 as in the future...
Alpha1 is currently planned for mid-february - but I'm still discussing the complete roadmap, Andreas -- Andreas Jaeger, aj@suse.de, http://www.suse.de/~aj/ SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
Silviu Marin-Caea wrote:
[...]
No, SLED is not an alternative, it's a product on its own and one of its main goals is stability. It has a much more appropriate mojo to provide that compared to openSUSE.
I think you have no idea what SLED is. It's tuned to desktop usage. However, many people need more than just an office desktop client.
Yeah it would be so fabulous if openSUSE would be stable and polished, have super new software and nice glitz, but it's damned *HARD* to achieve, technically and financially. SLED has the stability and polish, and reasonably new software. openSUSE has to be the "playground and test bed for freaks". This advances the distro.
I disagree completely. Again, you seem to have no idea of real-world situations. Factory might be a playground and test bed, but certainly not the release version of openSUSE, and it's good to know that Adrian and others agree with that. I use openSUSE at home and I certainly don't want to suffer from your proposal - my systems have to function, and if openSUSE can't provide the necessary balance between stability and being up-to-date, then it's gone and replaced by something that actually works. And I know a lot of people that have already gone down that route because of the software management problems in 10.1. Don't forget what Pascal wrote yesterday: "Also wrt the reputation of openSUSE (which is very important in order to get customers for SLED/SLES IMO)." This is exactly what happened in our office at work where everything is based on RedHat releases - we tried openSUSE (10.1) and because of many problems we decided not to go any step further. Sorry Novell, you have lost one possible customer there...
[...] What forcing? Hello! Who forced you?
If you want to have always the latest software versions in official openSUSE releases, then you effectively force all others to also use those possibly unstable and untested versions although they might prefer the stable versions in pratice. It's usually not easily possible to downgrade certain packages to get a stable release, so people are stuck. However, on the other hand it's usually much easier to upgrade software. As a consequence, an official openSUSE release should focus a bit more on stability and testing than bleeding edge software and those who want the latest and greatest can easily upgrade their systems via online repositories. Nobody said we should create another Debian stable release. It needs a good balance.
[...] Then, wrong product in the wrong place. If you want stability for production, SLED is the right thing to run.
Again, the only comment I can make on such statements is that you seem to have no idea what SLED is and that your point of view is very restricted - you seem to care only about the things that might suit your situation. If openSUSE wants to be somewhat professional (I heard claims that it should be considered "the best Linux distribution in the world"), then there is no way to use official releases as playground or test bed for possibly unstable and untested software. Cheers, Th. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Silviu Marin-Caea wrote:
[...] No, SLED is not an alternative, it's a product on its own and one of its main goals is stability. It has a much more appropriate mojo to provide that compared to openSUSE.
I think you have no idea what SLED is. It's tuned to desktop usage. However, many people need more than just an office desktop client.
+1
Yeah it would be so fabulous if openSUSE would be stable and polished, have super new software and nice glitz, but it's damned *HARD* to achieve, technically and financially. SLED has the stability and polish, and reasonably new software. openSUSE has to be the "playground and test bed for freaks". This advances the distro.
I disagree completely. Again, you seem to have no idea of real-world situations. Factory might be a playground and test bed, but certainly not the release version of openSUSE, and it's good to know that Adrian and others agree with that. I use openSUSE at home and I certainly don't want to suffer from your proposal - my systems have to function, and if openSUSE can't provide the necessary balance between stability and being up-to-date, then it's gone and replaced by something that actually works. And I know a lot of people that have already gone down that route because of the software management problems in 10.1. Don't forget what Pascal wrote yesterday: "Also wrt the reputation of openSUSE (which is very important in order to get customers for SLED/SLES IMO)." This is exactly what happened in our office at work where everything is based on RedHat releases - we tried openSUSE (10.1) and because of many problems we decided not to go any step further. Sorry Novell, you have lost one possible customer there...
Same thing happened with me. I had 20 clients/customers go th Red Hat because of SUSE 10.1. Had 10.1 been more like 10.2 they would have gone with SUSE. Saddly Novell lost a lot of customers because of the initial release of 10.1. I like the re-release. It did assist some clients to stay with SUSE, but that was because it was after the re-release that they tested SUSE. As I said on IRC I really thing 10.2 was a good release. I think it missed of few fixes that it should have had. I feel more testing would had fixed them. So I am really glad the testing has been extended. I think we need a better heads up to get bugs to a higher bug state so they get fixed before release. I think many do not understand/choose the right priority. I know I choose the lowest possible. I think some of the bugs that are marked urgent or high should be critical/show stopper. I think many do not want to add burden to the devs and thus do not attach the proper priority so they do not get fixed before release. Maybe there should be a task on the wiki for people to go through bugs to set the priority. This task needs to be done by someone's knowledgeable in what should really become a must fix. I guess that is why I am now doing more tests with factory to try and get things fixed sooner. Thanks, -- Boyd Gerber <gerberb@zenez.com> ZENEZ 1042 East Fort Union #135, Midvale Utah 84047 --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On Tuesday 23 January 2007 21:37, Thomas Hertweck wrote:
Silviu Marin-Caea wrote:
[...]
No, SLED is not an alternative, it's a product on its own and one of its main goals is stability. It has a much more appropriate mojo to provide that compared to openSUSE.
I think you have no idea what SLED is.
You don't know me, so please don't go wild with such opinions.
It's tuned to desktop usage. However, many people need more than just an office desktop client.
There's is SLES, SLE Real Time, Novell Linux POS etc. Are we still talking about production environment, or are we losing focus?
Again, you seem to have no idea of real-world situations.
Right...
Factory might be a playground and test bed, but certainly not the release version of openSUSE, and it's good to know that Adrian and others agree with that. I use openSUSE at home and I certainly don't want to suffer from your proposal - my systems have to function
You have started the exaggeration, I did not say that openSUSE releases need to be made all out of SVN checkouts. I was just saying that openSUSE cannot fulfill the role of enterprise products. Whoever wants long term super stability or does not want to be concerned at all with IT but do their business, should look there.
This is exactly what happened in our office at work where everything is based on RedHat releases - we tried openSUSE (10.1) and because of many problems we decided not to go any step further. Sorry Novell, you have lost one possible customer there...
And you haven't tried SLED/SLES because... ? Smart move, really, trying a different product to asses another. Pray tell that you have tried Fedora and decided to use Red Hat.
What forcing? Hello! Who forced you?
If you want to have always the latest software versions in official openSUSE releases.
Yes, I want the latest _releases_ (read stable) of software. And I'm getting it and you too. Even Factory does not contain SVN checkouts for most stuff, but maybe for some important pieces.
It needs a good balance.
It already has the perfect balance, I love it the way it is. Already the product cycle has been lengthened from 6 months to 8 months. It's just fine, should be left it as it is.
world"), then there is no way to use official releases as playground or test bed for possibly unstable and untested software.
We should stop the rant, seems to me that sufficient clarifications have been made. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Silviu Marin-Caea wrote:
[...] There's is SLES, SLE Real Time, Novell Linux POS etc. Are we still talking about production environment, or are we losing focus?
You started to talk about SLED ("If anyone wants stability, there's SLED10"), not me. If somebody lost focus, it might have been you. The rest of us was mainly talking about openSUSE.
[...] You have started the exaggeration, I did not say that openSUSE releases need to be made all out of SVN checkouts. I was just saying that openSUSE cannot fulfill the role of enterprise products.
Not true. Simple proof: Msg-ID: <200701221643.59632.silviu_marin-caea@fieldinsights.ro>
[...] And you haven't tried SLED/SLES because... ? Smart move, really, trying a different product to asses another. Pray tell that you have tried Fedora and decided to use Red Hat.
I have to repeat myself: you have no idea...
[...] Yes, I want the latest _releases_ (read stable) of software.
Not true. Your own emails are contradictory. Simple proof: Martin Schlander: "I think we can be a (little) bit more conservative when selecting versions to include, especially of core stuff, like xorg, and gcc which are both *unstable* releases." Silviu Marin-Caea: "I *disagree* on this. If anyone wants stability, there's SLED10." Now you want to sell me that you're only interested in stable software? You should make up your mind before answering emails.
[...] We should stop the rant, [...]
True. It's not worth spending time to discuss this topic with you. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On Monday 22 January 2007 13:41, Thomas Hertweck wrote:
With the build service in place and available software repositories for KDE, kernel, etc., you can always upgrade your stable release to newer components if you like.
+1, but please keep Tomcat more current :)
You should realize that openSUSE is not a playground and test bed for freaks but used on many production systems, for instance in small and mid-sized companies
+1, we use it at work for e-commerce servers with revenues in the millions of US$. -- Glenn Holmer (Q-Link: ShadowM) http://www.lyonlabs.org/commodore/c64.html --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On 23-01-2007 at 13:49, Glenn Holmer <gholmer@ameritech.net> wrote: +1, we use it at work for e-commerce servers with revenues in the millions of US$.
I think for such a case, a guaranteed contractual license like SLES would probably really be the right thing. It gives you more than only the required software, but having a whole team behind for support queries. And I guess the price would not be a problem if you have the business running as you say. Dom --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
On Tuesday 16 January 2007 17:10, Andreas Jaeger wrote:
We've briefly discussed during the last IRC meeting about what went bad in 10.2 and how we can do this better for future products, especially 10.3. I'm mainly interested in process feedback - and not on feedback that package x is broken (unless that shows a process problem).
So, let's discuss what we can do better for our next distro,
I mentioned some of these on IRC, but I'll mention them here too: I'd really like to see Zypper become a really mature and efficient (+speedy) app for the command line, and I know a lot of things for Zypper are in the works, so let me mention some of them: * Handling of local/remote RPMs (#230223) * A little more caching, i.e. with searching (#213762) * ...tied into, being speedy, which I think is really important. I acknowledge I'll get bad flames for this ;-), but APT (and even Smart, too, even) is super speedy, and it's a little off-putting that that an apt-get install takes, quite literally, less than 10 seconds, but a zypper in something (though it always delivers), takes over a minute with five sources because of the parsing metadata. And just wishes, such as: * Nice and accessible command line interface+output; I think Smart does really well in this regard. One example: show download rate (#227903) * And a very idealistic: solve build dependencies (#169757) * Full metalink support, especially for release time, and a metalink client shipped with 10.3. This would include handling our own metalinks, which we can make work with the torrent too. * Minimal X onto one CD. I think it only requires like 4 packages from CD2. This may be covered by Jaeger's "Making the basesystem smaller" thread on the opensuse-factory list, which would be nice to see. Kind thoughts, -- Francis Giannaros Web: http://francis.giannaros.org IRC: apokryphos (irc.freenode.net)
Hello, Am Dienstag, 16. Januar 2007 18:10 schrieb Andreas Jaeger:
We've briefly discussed during the last IRC meeting about what went bad in 10.2 and how we can do this better for future products, especially 10.3.
For those who didn't follow the whole thread, I added a short summary at https://bugzilla.novell.com/show_bug.cgi?id=233261 Regards, Christian Boltz PS: AJ, even if the AI is closed now, I'll remember what was proposed and remind you if needed ;-) -- I would say we first handle world-domination and then see what we can do for languages. [houghi in opensuse] --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Andreas Jaeger wrote:
We've briefly discussed during the last IRC meeting about what went bad in 10.2 and how we can do this better for future products, especially 10.3. I'm mainly interested in process feedback - and not on feedback that package x is broken (unless that shows a process problem).
What I miss is a more general statement about the future of openSUSE releases, in other words information about upcoming (major) changes and developments. This might not be of particular interest for the average home user but I am sure that professional users would really appreciate it since it might affect their work environment (in both ways, positive and negative and it's always helpful if you know what to expect) and it might give a first impression of features in and possible problems with future enterprise versions (provided that Novell keeps the approach that an openSUSE release forms the basis of an enterprise release). I know about http://en.opensuse.org/Factory/News which lists (some) changes in the current development version of openSUSE with respect to the previous stable version. This is a good starting point - however, it could be enhanced (and it should be distributed via e.g. "announce" mailing list). From my point of view it would be good to know more about where openSUSE is heading in general, say "the global picture". Maybe such things are discussed on the factory mailing list, I don't know, unfortunately it's almost impossible to follow all discussion groups... Cheers, Th. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
Thomas Hertweck <Thomas.Hertweck@web.de> writes:
Andreas Jaeger wrote:
We've briefly discussed during the last IRC meeting about what went bad in 10.2 and how we can do this better for future products, especially 10.3. I'm mainly interested in process feedback - and not on feedback that package x is broken (unless that shows a process problem).
What I miss is a more general statement about the future of openSUSE releases, in other words information about upcoming (major) changes and developments. This might not be of particular interest for the average home user but I am sure that professional users would really appreciate it since it might affect their work environment (in both ways, positive and negative and it's always helpful if you know what to expect) and it might give a first impression of features in and possible problems with future enterprise versions (provided that Novell keeps the approach that an openSUSE release forms the basis of an enterprise release).
That approach is kept.
I know about http://en.opensuse.org/Factory/News which lists (some) changes in the current development version of openSUSE with respect to the previous stable version. This is a good starting point - however, it could be enhanced (and it should be distributed via e.g. "announce" mailing list). From my point of view it would be good to know more about where openSUSE is heading in general, say "the global picture". Maybe such things are discussed on the factory mailing list, I don't know, unfortunately it's almost impossible to follow all discussion groups...
We're looking right now at the global picture and discuss some ideas that have been floating around for 10.3. Once that's done and put together, I'll share it - and then let's discuss it, Andreas -- Andreas Jaeger, aj@suse.de, http://www.suse.de/~aj/ SUSE Linux Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
Hi! On 1/16/07, Andreas Jaeger <aj@suse.de> wrote:
We've briefly discussed during the last IRC meeting about what went bad in 10.2 and how we can do this better for future products, especially 10.3. I'm mainly interested in process feedback - and not on feedback that package x is broken (unless that shows a process problem).
So, let's discuss what we can do better for our next distro,
Many have voted for stability. Me too. (I use openSUSE both at work and at home.) But I'd like bring another aspect into this.: stability of the security and other patches. I've had more problems always from updates than the original release (once after bad kernel, I could not even boot! Latest example was X.org update that, AFAIK, killed my Opera). I wish I could suggest how the testing of the patches could be improved. Specially as people expect the security patches always immediately. But I cant. I hope somebody else can suggest something practical. I hope I didn't offend anybody with this. Just my 1.5 cents. Maybe I'm the only one anyways with update problems. Keep up the good work! -- HG. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-project+help@opensuse.org
participants (23)
-
Adrian Schröter -
Andreas Hanke -
Andreas Jaeger -
Boyd Lynn Gerber -
Carlos E. R. -
Christian Boltz -
Dominique Leuenberger -
Francis Giannaros -
Gerald Pfeifer -
Glenn Holmer -
HG -
James Tremblay -
Jan Karjalainen -
jdd sur free -
Jozef Peterka -
Martin Schlander -
Matthias Hopf -
Pascal Bleser -
richard (MQ) -
Robert Schiele -
Silviu Marin-Caea -
Thomas Hertweck -
Vincenzo Barranco