Thread: 9.0 ?
Hello ! and happy new year to every one :-) i've had an idea coming over the last few weeks while i was reading the current commit fest web page ( http://wiki.postgresql.org/wiki/CommitFest_2008-11 ) To put it simple the idea is :"why not naming the upcoming version PostgreSQL 9.0 instead of PostgreSQL 8.4 ?" i guess this decision belongs to core-hackers. Furthermore i must admit that i've no idea what were the reasons to go from 7.4 to 8.0. I wasn't involved in the PostgreSQL community at the time, i haven't found any discussion about it on the mailing archives and the versioning page doesn't explain how the first digit evolves : http://www.postgresql.org/support/versioning By the way, when i see what's going to be inside the next major release i can't help telling me that this is gonna be a tremendous version and a "great leap forward" :) From my point of view this version will be an answer to many long-time users wishes : windowing function, embedded replication mechanism, improved FSM management, etc. i ain't nostradum but i've the feeling that this new version will amplify the numbers of migrations to PostgreSQL and we will see lots of new users in 2009. The 8.0 was the beginning of a new era, the forthcoming version may be another cornerstone. So it might be a good idea to increment the first digit. That would have a strong effect in public announcements and press releases. I think that will bring more spotlights to this new version and that will make the advocacy work a lot easier. i don't know if it's the right time and the right place to launch this debate but i'd like to know what you guys think about having PostgreSQL 9.0 released in 2009 :) -- damien clochard http://dalibo.org | http://dalibo.com
On Fri, 2009-01-02 at 11:49 +0100, damien@dalibo.info wrote:
>
>
> i don't know if it's the right time and the right place to launch this
> debate but i'd like to know what you guys think about having
> PostgreSQL
> 9.0 released in 2009 :)
Actually I was thinking the same, but I think we should use 9.0
*if/when* the hot standby and sync replication patches are committed.
Regards,
--
Devrim GÜNDÜZ, RHCE
devrim~gunduz.org, devrim~PostgreSQL.org, devrim.gunduz~linux.org.tr
http://www.gunduz.org
Attachment
On Friday 02 January 2009 05:49:57 damien@dalibo.info wrote: > Hello ! > > and happy new year to every one :-) > > i've had an idea coming over the last few weeks while i was reading the > current commit fest web page ( > http://wiki.postgresql.org/wiki/CommitFest_2008-11 ) > > To put it simple the idea is :"why not naming the upcoming version > PostgreSQL 9.0 instead of PostgreSQL 8.4 ?" > > i guess this decision belongs to core-hackers. Furthermore i must admit > that i've no idea what were the reasons to go from 7.4 to 8.0. I wasn't > involved in the PostgreSQL community at the time, i haven't found any > discussion about it on the mailing archives and the versioning page > doesn't explain how the first digit evolves : > IIRC, the primary reason 7.5 became 8.0 was due to the changes required to support win32, which touched a significant enough portion of the code base that even those people who did not plan to use the win32 support (or any of the new features) could have been effected by the changes in that release to underlying routines that had been reworked. > http://www.postgresql.org/support/versioning > > By the way, when i see what's going to be inside the next major release > i can't help telling me that this is gonna be a tremendous version and a > "great leap forward" :) > > From my point of view this version will be an answer to many long-time > users wishes : windowing function, embedded replication mechanism, > improved FSM management, etc. > while these changes are things that a lot of people want, you should be cautious in that not all of the items in commitfest page are guaranteed to get in (and it's probably unlikely they all will at this point) also, it's important to keep in mind that while some of these changes are significant, things like the dead space map and changes to FSM are really evolutionary changes, not revolutionary changes. (I think in-place upgrades are probably the biggest revolutionary change left out there, but that's just my opinion) > i ain't nostradum but i've the feeling that this new version will > amplify the numbers of migrations to PostgreSQL and we will see lots of > new users in 2009. The 8.0 was the beginning of a new era, the > forthcoming version may be another cornerstone. > > So it might be a good idea to increment the first digit. That would have > a strong effect in public announcements and press releases. I think that > will bring more spotlights to this new version and that will make the > advocacy work a lot easier. > > i don't know if it's the right time and the right place to launch this > debate but i'd like to know what you guys think about having PostgreSQL > 9.0 released in 2009 :) > Current policy is that we don't increment the version number for marketing purposes, and at this point it's probably premature to have the discussion until we get a complete picture of what items not yet committed will actually make it in. -- Robert Treat Conjecture: http://www.xzilla.net Consulting: http://www.omniti.com
> Current policy is that we don't increment the version number for marketing > purposes, and at this point it's probably premature to have the discussion > until we get a complete picture of what items not yet committed will actually > make it in. > Also, it's going to be painful for our redistributors when we switch over to 10.0, so we're setting a really high bar for that first digit. We took 10 years to go from 6.0 to 8.0. Linux is still on version 2, as is Java, and Perl has been version 5 for ~~ 12 years now. So, no rush. ;-) --Josh
On Monday 05 January 2009 19:57:22 Josh Berkus wrote: > Also, it's going to be painful for our redistributors when we switch > over to 10.0, Huh? Clearly, they have had enough time to practice on 8.2.9 -> 8.2.10 or whatever.
On Mon, 2009-01-05 at 21:04 +0200, Peter Eisentraut wrote:
> Clearly, they have had enough time to practice on 8.2.9 -> 8.2.10 or
> whatever.
Oh, I spent 2 weeks in order to get used to typing 2 digits instead of 1
digit :P
--
Devrim GÜNDÜZ, RHCE
devrim~gunduz.org, devrim~PostgreSQL.org, devrim.gunduz~linux.org.tr
http://www.gunduz.org
Attachment
Devrim GÜNDÜZ wrote: > On Mon, 2009-01-05 at 21:04 +0200, Peter Eisentraut wrote: > >> Clearly, they have had enough time to practice on 8.2.9 -> 8.2.10 or >> whatever. >> > > Oh, I spent 2 weeks in order to get used to typing 2 digits instead of 1 > digit :P > > Perhaps we should also adopt an ubuntu-like strategy of naming the releases. That'll give people the impression of major version changes instead of the number. For example, perhaps the next version could be code named "Cornucopious Core" or something ;-) Kind of like "Hardy Heron", or "Dapper Dan" . I think today people tend to refer to the name of the release rather than the version... I, of course, am not totally serious here... -- Chander Ganesan Open Technology Group, Inc. One Copley Parkway, Suite 210 Morrisville, NC 27560 919-463-0999/877-258-8987 http://www.otg-nc.com
All, > Perhaps we should also adopt an ubuntu-like strategy of naming the > releases. That'll give people the impression of major version changes > instead of the number. For example, perhaps the next version could be > code named "Cornucopious Core" or something ;-) Kind of like "Hardy > Heron", or "Dapper Dan" . I think today people tend to refer to the Gods forfend! Not that you were serious, but I actually rank the Ubuntu release naming scheme as "experimental failure" (kind of like "Postgres95"), and wish Ubuntu would go back to naming the releases after the date, or just use numbers like everyone else. I'm forever trying to remember whether the current release is "Dapper Dalmation" or "Stellar Sparrow" or "Woody Woodpecker" or "Moose & Squirrel". And don't get me started on Apple and their releases of OSX "Ocelot" and "Caracal". It's a release naming scheme which caters exclusively to insiders. Seriously, though, the real issue we'll run into with PostgreSQL 10 is that there's several Linux distributors (including, I think, Red Hat) which are using a package serial scheme which doesn't include a leading "0". So the upcoming version is 80400, not 080400, and will cause them to do some rejiggering when we do eventually release version 10. --Josh
On Tuesday 06 January 2009 20:59:46 Josh Berkus wrote: > Not that you were serious, but I actually rank the Ubuntu release naming > scheme as "experimental failure" (kind of like "Postgres95"), and wish > Ubuntu would go back to naming the releases after the date, or just use > numbers like everyone else. If you go to the Ubuntu web site, they offer you release 8.10 and 8.04 LTS for download. Seems perfectly normal. > I'm forever trying to remember whether the > current release is "Dapper Dalmation" or "Stellar Sparrow" or "Woody > Woodpecker" or "Moose & Squirrel". And don't get me started on Apple > and their releases of OSX "Ocelot" and "Caracal". It's a release naming > scheme which caters exclusively to insiders. Operating systems vendors are particularly prone to do that, apparently. (cf. also Microsoft, Sun) > Seriously, though, the real issue we'll run into with PostgreSQL 10 is > that there's several Linux distributors (including, I think, Red Hat) > which are using a package serial scheme which doesn't include a leading > "0". So the upcoming version is 80400, not 080400, and will cause them > to do some rejiggering when we do eventually release version 10. With some bemusement I notice your posts on this topic whenever "PostgreSQL 10" is mentioned anywhere. But it is quite frankly complete nonsense. Any packaging system worth anything can handle that without any problem. Not to mention that your hypothesized "package serial scheme" bears no similarity with reality. So, relax, we'll be fine.
On Tue, Jan 6, 2009 at 7:30 PM, Peter Eisentraut <peter_e@gmx.net> wrote: > > With some bemusement I notice your posts on this topic whenever "PostgreSQL > 10" is mentioned anywhere. But it is quite frankly complete nonsense. Any > packaging system worth anything can handle that without any problem. Even Windows will cope with it :-p -- Dave Page EnterpriseDB UK: http://www.enterprisedb.com
On Tue, 6 Jan 2009, Josh Berkus wrote:
> Seriously, though, the real issue we'll run into with PostgreSQL 10 is that
> there's several Linux distributors (including, I think, Red Hat) which are
> using a package serial scheme which doesn't include a leading "0". So the
> upcoming version is 80400, not 080400, and will cause them to do some
> rejiggering when we do eventually release version 10.
There isn't any such serial version scheme that I'm aware of for the RHEL
packages. Here's a RHEL4 install showing the expected x.y.z number:
$ rpm -qi postgresql-libs
Name : postgresql-libs Relocations: (not relocatable)
Version : 7.4.19 Vendor: Red Hat, Inc.
Release : 1.el4_6.1
As for where that comes from, this is what the latest Fedora .spec file
building recent packages looks like:
Summary: PostgreSQL client programs and libraries
Name: postgresql
Version: 8.3.5
Release: 2%{?dist}
No serial scheme to be found there. emacs has had version numbers >10 for
a long time; here's one of those packages that has a version number like
the PG packages will have eventually, works fine:
$ rpm -qi xemacs
Name : xemacs Relocations: (not relocatable)
Version : 21.4.15 Vendor: Red Hat, Inc.
Release : 11.EL4
I just took a quick glance at the Debian, Gentoo, and SuSE packaging as
well, and they all seemed OK too. Looks to me like if this problem
existed at some point, it's already been resolved in all the major Linux
distributions. I'd be surprised if there were really "several" left where
this is still a concern.
--
* Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD
Josh Berkus a écrit : > All, > >> Perhaps we should also adopt an ubuntu-like strategy of naming the >> releases. That'll give people the impression of major version changes >> instead of the number. For example, perhaps the next version could be >> code named "Cornucopious Core" or something ;-) Kind of like "Hardy >> Heron", or "Dapper Dan" . I think today people tend to refer to the > > Gods forfend! > > Not that you were serious, but I actually rank the Ubuntu release naming > scheme as "experimental failure" (kind of like "Postgres95"), and wish > Ubuntu would go back to naming the releases after the date, or just use > numbers like everyone else. I'm forever trying to remember whether the > current release is "Dapper Dalmation" or "Stellar Sparrow" or "Woody > Woodpecker" or "Moose & Squirrel". And don't get me started on Apple > and their releases of OSX "Ocelot" and "Caracal". It's a release naming > scheme which caters exclusively to insiders. > Every naming scheme is only understandable by insiders. After apart from debian fanboys, who can tell where "etch" and "sarge" names come from ? It's strictly the same thing with version number, except us, who can say what were the differences between the 8.0 and the 8.1 :-) The only advantage of release naming is that it's more fun than numbers. how about "Elegant Elephant" ? "Persistent Pachyderm" ? "Marvelous Mastodont" ? :o)
2009/1/9 damien clochard <damien@dalibo.info>: > Josh Berkus a écrit : >> All, >> >>> Perhaps we should also adopt an ubuntu-like strategy of naming the >>> releases. That'll give people the impression of major version changes >>> instead of the number. For example, perhaps the next version could be >>> code named "Cornucopious Core" or something ;-) Kind of like "Hardy >>> Heron", or "Dapper Dan" . I think today people tend to refer to the >> >> Gods forfend! >> >> Not that you were serious, but I actually rank the Ubuntu release naming >> scheme as "experimental failure" (kind of like "Postgres95"), and wish >> Ubuntu would go back to naming the releases after the date, or just use >> numbers like everyone else. I'm forever trying to remember whether the >> current release is "Dapper Dalmation" or "Stellar Sparrow" or "Woody >> Woodpecker" or "Moose & Squirrel". And don't get me started on Apple >> and their releases of OSX "Ocelot" and "Caracal". It's a release naming >> scheme which caters exclusively to insiders. >> > When we talk about 'labels' in the most cases we talk about SO's. If you see, comercial databases don't use 'names' AFAIK. Make the software more seriusly. In relation of more big number of versions, i think it is not really a good point. I think less digits shows you that the software is more 'stable' IMHO. I think too that the 9 version must support inside tools to full replication. > Every naming scheme is only understandable by insiders. After apart from > debian fanboys, who can tell where "etch" and "sarge" names come from ? > It's strictly the same thing with version number, except us, who can say > what were the differences between the 8.0 and the 8.1 :-) > > The only advantage of release naming is that it's more fun than numbers. > > how about "Elegant Elephant" ? "Persistent Pachyderm" ? "Marvelous > Mastodont" ? :o) > "Eternal Mamooth" ... :P Regards > > > -- > Sent via pgsql-advocacy mailing list (pgsql-advocacy@postgresql.org) > To make changes to your subscription: > http://www.postgresql.org/mailpref/pgsql-advocacy > -- Emanuel Calvo Franco ArPUG / AOSUG Member Postgresql Support & Admin
On Mon, Jan 5, 2009 at 6:57 PM, Josh Berkus <josh@agliodbs.com> wrote:
>
>> Current policy is that we don't increment the version number for marketing
>> purposes, and at this point it's probably premature to have the discussion
>> until we get a complete picture of what items not yet committed will
>> actually make it in.
>
> Also, it's going to be painful for our redistributors when we switch over to
> 10.0, so we're setting a really high bar for that first digit.
Assuming the sync replication and hot standby get committed and we
bump the version to 9.0, there will be a huge 'awesomeness' factor
needed to bump it to 10.
Frankly I cannot even imagine what new feature would mandate bumping
from 9 to 10. :-)
> We took 10 years to go from 6.0 to 8.0. Linux is still on version 2, as is
> Java, and Perl has been version 5 for ~~ 12 years now. So, no rush. ;-)
While I don't like the versions to be bumped up too quickly, I think there
is one pretty important reason.
Major version bump serves not only as a PR statement. It is also
'early warning' indicator -- "Hey, we've changed so much stuff / added
so many new features so you'd better be careful.". And I would think
we owe it to users. :-)
Best regards,
Dawid
--
.................. ``The essence of real creativity is a certain
: *Dawid Kuroczko* : playfulness, a flitting from idea to idea
: qnex42@gmail.com : without getting bogged down by fixated demands.''
`..................' Sherkaner Underhill, A Deepness in the Sky, V. Vinge
Dawid, > Major version bump serves not only as a PR statement. It is also > 'early warning' indicator -- "Hey, we've changed so much stuff / added > so many new features so you'd better be careful.". And I would think > we owe it to users. :-) Yes. Speaking entirely for myself, it's possible that we'll start using the first digit to indicate file format changes which make in-place upgrade impossible. --Josh