pgFoundry

PgFoundry has been brought up quite a bit lately. How we want
it to succeed, how we want it to be the hub of PostgreSQL development.
So my question is... Why isn't it?

Speaking only for myself, I volunteered to have my project moved over
first as a test case. This was agreed, the original plan being to start
moving the project "within two weeks." It stretched into months, many of
them, and at some point I figured I'd asking and just wait to be notified.
That must have been over a year ago by now.

A lot of stuff happened to me personally over that waiting period (like
living in three noncontiguous countries) so I may easily have missed, or
even forgotten, an announcement on one of the mailing lists.

I'm not unique, however. The same may be true of others. So if you want
to encourage people to migrate to the new site, tell them about it and
repeat, repeat, repeat! Post an announcement for each project that makes
the jump. Bring it to the fore. Convey the fact--assuming it is one, of
course--that this is where the Postgres-related projects are headed. Keep
people updated. Did I mention "repeat" as well?

Jeroen

On Fri, 6 May 2005 01:32 pm, Joshua D. Drake wrote:

Hello,

PgFoundry has been brought up quite a bit lately. How we want
it to succeed, how we want it to be the hub of PostgreSQL development.
So my question is... Why isn't it?

Because it's not the hub of PostgreSQL development. I think it will be difficult to
build a culture of "This" is the place to be unless we actually kill gborg totally.
Currently there are still projects there, I'm personally never sure where to look
for a particular project. Even some of the more high profile projects like Slony-I
aren't even on PgFoundy. How can we expect people to take it seriously.

Issue two, which I know is being resolved, is that people find it painfully slow to
navigate. Who wants to search a sight that is painfully slow. But until the site is
running at a good speed, and can support a reasonably large number of projects
at that speed, are people going to be encouraged to move over? I don't think so.

I know there are issues with the CVS statistics. If I'm looking
for a project to forfill a function, looking at the statistics is a good way to determine
if the project is going anywhere or not. As well as releases. Currently every project
say "This project has no CVS history." and lists 0 commits and 0 adds. I don't think
this generally looks good for us. If there was no information, it would be better than
the false information.

Also a little more prominence on the PostgreSQL main page would be helpful.
I know there is a link, but to the unknowning user, what is pgFoundry about?
Maybe some advertisting about the fact that is you want something that runs with
your PostgreSQL server, head on over to pgFoundry to find it.

We should encourage any OSS projects that are for PostgreSQL to use
pgFoundry instead of any other hosting source. One very basic example is the
nss library I have been working on. I recently found that in February, another fork
of the nss library had popped up on debian's Gforge site. I had no idea it existed,
and they had no idea I existed, and they use PostgreSQL fairly exclusively. Where
were they looking for an nss library when then needed one? Well, it obviously wasn't
at pgFoundry.

Why is PostgreSQL not hosted at pgFoundry?
It seems that it has everything we would need?

This is possibly true, it gives the advantage of trackers and many functions that
the lists are used for. Maybe it's less likely we would lose patches and/or bugs.

I don't really have a lot of knowledge about the benefits disadvantages, so I'll
leave the people who actually use CVS and things like that to make a comment.

To keep this on track, consider my question as if it were 2 months from
now and pgFoundry was all up on the new servers etc...

Personally, this is a problem. It's another 2 months away. In that time, I think we
also need to focus on getting rid of gborg and redirecting people to pgFoundry.
But can the current setup handle the load, and can we get the move from gborg done?
Also is the upgrade path for moving servers easy, if it is it's probably more reason to
push for the full closure of gborg.

Now, despite the apparent large number of complaints. I think pgFoundry is a very good
idea, and will work well in the long run. I just think we need to get some things going
well to get people on the site more. Once that happens, people will instinctively look there.

Sincerely,

Russell Smith

Personally, this is a problem. It's another 2 months away. In that time, I think we
also need to focus on getting rid of gborg and redirecting people to pgFoundry.
But can the current setup handle the load, and can we get the move from gborg done?
Also is the upgrade path for moving servers easy, if it is it's probably more reason to
push for the full closure of gborg.

Gborg is considered deprecated. The projects that are there may or may
not move. Although the goal it to get them to.

Also at this point Gborg has nothing to do with the initial question. I
am not asking about Gborg. I am asking why we are not placed PostgreSQL
at the core of what is supposed to be the PostgreSQL Projects website,
pgFoundry.

Sincerely,

Joshua D. Drake
Command Prompt, Inc.

Now, despite the apparent large number of complaints. I think pgFoundry is a very good
idea, and will work well in the long run. I just think we need to get some things going
well to get people on the site more. Once that happens, people will instinctively look there.

Sincerely,

Russell Smith

--
Your PostgreSQL solutions company - Command Prompt, Inc. 1.800.492.2240
PostgreSQL Replication, Consulting, Custom Programming, 24x7 support
Managed Services, Shared and Dedication Hosting
Co-Authors: plPHP, plPerlNG - http://www.commandprompt.com/

On Fri, 6 May 2005, Joshua D. Drake wrote:

Personally, this is a problem. It's another 2 months away. In that time,
I think we
also need to focus on getting rid of gborg and redirecting people to
pgFoundry.
But can the current setup handle the load, and can we get the move from
gborg done?
Also is the upgrade path for moving servers easy, if it is it's probably
more reason to
push for the full closure of gborg.

Gborg is considered deprecated. The projects that are there may or may not
move. Although the goal it to get them to.

Also at this point Gborg has nothing to do with the initial question. I am
not asking about Gborg. I am asking why we are not placed PostgreSQL at the
core of what is supposed to be the PostgreSQL Projects website, pgFoundry.

This has been discussed several times before ... pgFoundry offers us
nothing that we don't already have, and takes away several things ...
also, with pgFoundry moving 'State side', it has one more check against
moving the core project over to it ...

With PostgreSQL *not* US based, we are not subject to the whim's of the US
government, nor its export restrictions, nor its lawyers ...

This is one of those "check the archives, its been discussed before" kinda
threads ;(

----
Marc G. Fournier Hub.Org Networking Services (http://www.hub.org)
Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664

mr-russ@pws.com.au (Russell Smith) writes:

Because it's not the hub of PostgreSQL development. I think it will
be difficult to build a culture of "This" is the place to be unless
we actually kill gborg totally. Currently there are still projects
there, I'm personally never sure where to look for a particular
project. Even some of the more high profile projects like Slony-I
aren't even on PgFoundy. How can we expect people to take it
seriously.

Various people are keen on moving Slony-I over to PGFoundry, but not
until the perpetual "two months" expires, and it gets to the necessary
"more stabler" point.
--
(format nil "~S@~S" "cbbrowne" "acm.org")
http://www.ntlug.org/~cbbrowne/sap.html
Rules of the Evil Overlord #78. "I will not tell my Legions of Terror
"And he must be taken alive!" The command will be: ``And try to take
him alive if it is reasonably practical.''"
<http://www.eviloverlord.com/&gt;

This has been discussed several times before ... pgFoundry offers us
nothing that we don't already have,

Actually it does:

1. Public presentation of the project development
2. Public presentation of the project bugs
3. Public presentation of the project tracker and docs

Perception is everything :)

The new people, the people who are coming to PostgreSQL they are looking
for the Web, they want an interface that they can peruse, easily search,
see priority etc...

That is not mailing lists :)

and takes away several things ...

Curious to know what those are?

With PostgreSQL *not* US based, we are not subject to the whim's of the
US government, nor its export restrictions, nor its lawyers ...

I am not even going to touch this one.

Sincerely,

Joshua D. Drake

--
Your PostgreSQL solutions company - Command Prompt, Inc. 1.800.492.2240
PostgreSQL Replication, Consulting, Custom Programming, 24x7 support
Managed Services, Shared and Dedication Hosting
Co-Authors: plPHP, plPerlNG - http://www.commandprompt.com/

On Fri, 6 May 2005, Joshua D. Drake wrote:

This has been discussed several times before ... pgFoundry offers us
nothing that we don't already have,

Actually it does:

1. Public presentation of the project development

Sounds like what http://www.postgresql.org is either doing, or should be
extended to do ...

2. Public presentation of the project bugs

Same as 1, but maybe a bit more forcefully by ppl like Tom ... god, Josh
Berkus even spent the time going through all the various 'bug tracking
tools' out there to find something that would satisfy the requirement by
ppl like Tom to *not* have a web interface, while providing one, and none
of them (and, if I recall correctly, *especially* not pgFoundry) came
close ... and Josh put alot of work/effort into the research, if I recall
correctly, including looking at some potential "commercial" offerings that
were willing to work with us ...

Again, this was heavily discussed at least a couple of times ...

3. Public presentation of the project tracker and docs

Isn't that what http://www.postgresql.org is for?

The new people, the people who are coming to PostgreSQL they are looking
for the Web, they want an interface that they can peruse, easily search,
see priority etc...

Sounds like what http://www.postgresql.org is either doing, or should be
extended to do ...

----
Marc G. Fournier Hub.Org Networking Services (http://www.hub.org)
Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664

Marc G. Fournier wrote:

On Fri, 6 May 2005, Joshua D. Drake wrote:

This has been discussed several times before ... pgFoundry offers us
nothing that we don't already have,

Actually it does:

1. Public presentation of the project development

Sounds like what http://www.postgresql.org is either doing, or should be
extended to do ...

No that is public presentation of the project not project development.

2. Public presentation of the project bugs

Same as 1, but maybe a bit more forcefully by ppl like Tom ... god, Josh
Berkus even spent the time going through all the various 'bug tracking
tools' out there to find something that would satisfy the requirement by
ppl like Tom to *not* have a web interface, while providing one, and
none of them (and, if I recall correctly, *especially* not pgFoundry)

The new version of PgFoundry provides a SOAP interface so there is no
reason why we could not create a simple interface for command line
access (which I would use as well)

3. Public presentation of the project tracker and docs

Isn't that what http://www.postgresql.org is for?

Well not in my mind although it could be extended to. Case in point
where do I go on the website to see what the current todo and progress
based on the todos?

I click on Developers->RoadMap->TODO

That is great (seriously) but from a developer (at least in my mind) it
doesn't tell me what I want to know.

Now look at:

http://gforge.org/tracker/?atid=105&amp;group_id=1&amp;func=browse

http://gforge.org/forum/?group_id=1

http://gforge.org/pm/task.php?group_project_id=54&amp;group_id=1&amp;func=browse

The color scheme is bad but I think it makes my point.

We don't have anything like this, and I think it would be beneficial.

The new people, the people who are coming to PostgreSQL they are
looking for the Web, they want an interface that they can peruse,
easily search, see priority etc...

Sounds like what http://www.postgresql.org is either doing, or should be
extended to do ...

I am talking about new developers. Developers want good, solid
information that is easy to find. That is not always the case as we have
recently been speaking about at length in several different threads.

Sincerely,

Joshua D. Drake

----
Marc G. Fournier Hub.Org Networking Services (http://www.hub.org)
Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664

--
Your PostgreSQL solutions company - Command Prompt, Inc. 1.800.492.2240
PostgreSQL Replication, Consulting, Custom Programming, 24x7 support
Managed Services, Shared and Dedication Hosting
Co-Authors: plPHP, plPerlNG - http://www.commandprompt.com/

On Fri, May 06, 2005 at 10:01:45AM -0700, Joshua D. Drake wrote:

With PostgreSQL *not* US based, we are not subject to the whim's of the
US government, nor its export restrictions, nor its lawyers ...

I am not even going to touch this one.

Why? While I wouldn't put the terms exactly the way Marc did, I
_would_ say that the US approach to "intellectual property" in
software is sufficiently reaching as to qualify in the "over"
category. Canada's rules are bad enough, but having worked under
both regimes, I'm moderately convinced that having a non-US "home
base" for CVS is a good idea.

A

--
Andrew Sullivan | ajs@crankycanuck.ca
When my information changes, I alter my conclusions. What do you do sir?
--attr. John Maynard Keynes

Andrew Sullivan wrote:

On Fri, May 06, 2005 at 10:01:45AM -0700, Joshua D. Drake wrote:

With PostgreSQL *not* US based, we are not subject to the whim's of the
US government, nor its export restrictions, nor its lawyers ...

I am not even going to touch this one.

Why?

Because I think if you think that having the CVS anywhere will matter to
the US, I think your dead wrong.

If the US wants to come after the project they are going to one way or
the other. They will either:

A. Prosecute US developers working on the project
B. Prosecute Non-US developers with countries they have extradition
treaties with.
C. Send in the army to overthrow your dictator and hunt you.

It doesn't matter either way.

FYI, a US business can rather successfully sue a Canadian one.

I am not saying I agree or disagree with the above 3. Frankly it is none
of anybody's business what I think about it. However I am no fool in
thinking that another country provides any veil if the US actually wants
something you have.

Sincerely,

Joshua D. Drake
Command Prompt, Inc.

While I wouldn't put the terms exactly the way Marc did, I

_would_ say that the US approach to "intellectual property" in
software is sufficiently reaching as to qualify in the "over"
category. Canada's rules are bad enough, but having worked under
both regimes, I'm moderately convinced that having a non-US "home
base" for CVS is a good idea.

A

--
Your PostgreSQL solutions company - Command Prompt, Inc. 1.800.492.2240
PostgreSQL Replication, Consulting, Custom Programming, 24x7 support
Managed Services, Shared and Dedication Hosting
Co-Authors: plPHP, plPerlNG - http://www.commandprompt.com/

On Fri, May 06, 2005 at 11:35:19AM -0700, Joshua D. Drake wrote:

FYI, a US business can rather successfully sue a Canadian one.

Yes, but in Canada, only for actual violations of Canadian law. Most
of the time, the practical effect of this is nothing, because the
Canadian businesses have US assets that the USian business can go
after in USian courts. (This tactic is the one that's been
attempted, for instance, to perform the enforcement of the
extraterritorial claims of the US Cuba embargo.) But since the
project isn't a legal entity and has no assets, we don't have that
problem. Individual contributors might, of course, but we can't do
anything about that anyway.

I am not saying I agree or disagree with the above 3. Frankly it is none
of anybody's business what I think about it. However I am no fool in
thinking that another country provides any veil if the US actually wants
something you have.

Well, sure. But the code isn't really something they can "have" any
more than they can "have" the idea of public key cryptography. The
more relevant question is a cost-benefit one: the US government spent
(IMHO) too much time harassing US security researchers over PGP, for
instance, but didn't do very much attempting to make life difficult
for non-US researchers. I submit that was mostly because the
diplomatic pain that it was likely to cost wasn't worth the pointless
benefit of trying to put the cat back in the bag. I think that there
is a potential nonzero advantage in having the project hosted outside
of the strict legal reach of the US Congress. The Parliament of
Canada isn't a whole lot better, but its tendency to feature debates
about who will be best at bribing some part of the country to keep
quiet about the divorce means that it is less likely to spend as much
time attempting to tell people what to think. It's a very modest
benefit, to be sure, but not one to give up without thinking. (In
the Canada case, of course, it comes at the potential cost that in
any year, the project could find itself in a new, and previously
non-existent, country.)

A

--
Andrew Sullivan | ajs@crankycanuck.ca
A certain description of men are for getting out of debt, yet are
against all taxes for raising money to pay it off.
--Alexander Hamilton

No that is public presentation of the project not project development.

I don't see that people are going to be able to participate in development
if they don't use the mailing lists.

I am not arguing that but public mailing lists are no place to track
status of sub projects or tasks. They are for discussion.

Sincerely,

Joshua D. Drake

--
Your PostgreSQL solutions company - Command Prompt, Inc. 1.800.492.2240
PostgreSQL Replication, Consulting, Custom Programming, 24x7 support
Managed Services, Shared and Dedication Hosting
Co-Authors: plPHP, plPerlNG - http://www.commandprompt.com/

On Fri, 6 May 2005, Joshua D. Drake wrote:

No that is public presentation of the project not project development.

I don't see that people are going to be able to participate in development
if they don't use the mailing lists.

I am not arguing that but public mailing lists are no place to track status
of sub projects or tasks. They are for discussion.

first off, there is really no reason why a 'PostgreSQL Server' project
can't be created on pgFoundry for the express purpose of this sort of
thing, with pointers to where the real work is happening ...

but ... who is going to maintain it? if developers aren't willing to deal
with a formal bug tracking system, I can't see how a 'web based project
tracking system' is ever going to get updated, unless somehow maintaining
of that ties in with the Weekly News that David/Elein/etc have been
producing ... ?

But, if that is the case, then why not just archive the Weekly News on the
main web site and provide a link/sub-page expressly for that?

----
Marc G. Fournier Hub.Org Networking Services (http://www.hub.org)
Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664

On Fri, May 06, 2005 at 11:09:36 -0700,
"Joshua D. Drake" <jd@commandprompt.com> wrote:

Marc G. Fournier wrote:

On Fri, 6 May 2005, Joshua D. Drake wrote:

1. Public presentation of the project development

Sounds like what http://www.postgresql.org is either doing, or should be
extended to do ...

No that is public presentation of the project not project development.

I don't see that people are going to be able to participate in development
if they don't use the mailing lists.

If they want an overview of what is happening, they should read the weekly
news summaries. The TODO list will let them know what new features have
been completed for the next release.

Joshua D. Drake wrote:

No that is public presentation of the project not project development.

I don't see that people are going to be able to participate in
development
if they don't use the mailing lists.

I am not arguing that but public mailing lists are no place to track
status of sub projects or tasks. They are for discussion.

Joshua,

I think you formulated the question wrong. It shouldn't be "why aren't
we using pgFoundry?" but "what can we do to improve our processes?" Only
after that question is answered should toolsets be considered. For
example, the TODO list is not really a task list at all - its status is
very nebulous. Some things will probably never be done, and many many
things will be done which aren't on it. So before we ask about task
tracking, we need to ask where the list of tracked tasks comes from. In
the past I have been told "there isn't and can't be such a list". In
that case, use of task tracking software is just a waste of bandwidth.

I'd like to see a bit more in the way of formal process, but to be done
right that also needs some resources (i.e. someone's time) which is not
something we are rich in.

cheers

andrew

"Joshua D. Drake" <jd@commandprompt.com> writes:

No that is public presentation of the project not project development.

I don't see that people are going to be able to participate in development
if they don't use the mailing lists.

I am not arguing that but public mailing lists are no place to track status of
sub projects or tasks. They are for discussion.

What does it mean to "track" the status of something? How would the status
change except by discussion? What would be the point of announcing the status
of something without allowing people to comment?

I think you have a severely flawed idea of how free software development
proceeds. What you're describing sounds like something a manager of a
commercial project would want. Perhaps it's something the managers of the
people working on Postgres on behalf of some corporate sponsors might want but
in those cases I doubt they would want the information to be public anyways.

In the free software world there's no top-down management of the project with
managers issuing direction and expecting feedback reports. People only want
tools that make their lives easier. Not tools that make other people's lives
easier at the expense of their own convenience. The programmers are not
beholden to any corporate interests (other than their own sponsors, who
presumably are getting all the feedback they're looking for privately).

I'm rather surprised Postgres doesn't have a good bug tracking system. That's
something most projects find pretty essential. Strangely enough the reason
seems to be that Postgres really doesn't have many bugs... Unlike web browsers
or GUIs or most of the other free software projects out there, databases don't
tolerate bugs well. Any serious bug is cause for an immediate point release.
The only use for a bug tracking system would really be for tracking all those
pesky "IWBNI" bugs that never rise to urgent status.

But "tracking the status" of sub-projects is just not the kind of thing free
software people do. They send emails when they have something to say.

--
greg

On Sat, 7 May 2005, Greg Stark wrote:

But "tracking the status" of sub-projects is just not the kind of thing
free software people do. They send emails when they have something to
say.

in defence of Joshua's idea, there are some "large projects" within our
development that would be nice to see some sort of 'time lines' for, but
mainly those are the sorts of things that have set milestones to look for
... I believe that Simon's work on the PITR stuff might fall under that,
where he had various 'stages' he was working towards ...

But, I think that the # of "sub projects" that this would apply to is so
few that setting anything up more formally then Simon sending out an
updated patch would be more work then the derived benefit ...

----
Marc G. Fournier Hub.Org Networking Services (http://www.hub.org)
Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664

Greg,

I'm rather surprised Postgres doesn't have a good bug tracking system.
That's something most projects find pretty essential. Strangely enough the
reason seems to be that Postgres really doesn't have many bugs... Unlike
web browsers or GUIs or most of the other free software projects out there,
databases don't tolerate bugs well. Any serious bug is cause for an
immediate point release. The only use for a bug tracking system would
really be for tracking all those pesky "IWBNI" bugs that never rise to
urgent status.

Actually, a bug tracker would be useful for two purposes:

1) Tracking bugs that are actually feature requests, in an effort (possibly
futile) to cut down on the number of "when will PostgreSQL be able to use an
index on MAX()?" requests that we get. A certain amount of this is in our
FAQ, but the FAQ has the flaws of both not being easily searchable, and
having a very manual update process so that it frequently gets out of date.

2) Tracking bugs that were fixed (and features that were added) in particular
releases so that users know when they need to upgrade. For example, if a
user had an index corruption problem with 7.4.1, it would be useful for them
to know that an upgrade to 7.4.5 (as I recall) would fix it. Currently,
they'd have to read all of the release notes from 7.4.2 through 7.4.5 and
decipher insider terminonolgy to figure it out -- not always easy to do, and
even harder to convince your boss.

The problem is that a bug tracker would not be useful to the Postgresql
*developers*; our current system works pretty well for us now. Except for
one possibility:

IF we had a formal bugtracker, are there people who are not currently
contributing to PostgreSQL who would be willing/able to read, test, and
analyse bug reports? With the addition of several companies to our
community, it's a possibility, and would make the trouble of using a bug
tracker worthwhile, I think.

Comments?

--
Josh Berkus
Aglio Database Solutions
San Francisco

What does it mean to "track" the status of something? How would the status
change except by discussion? What would be the point of announcing the status
of something without allowing people to comment?

No one said anything about not letting people comment or discuss. What I
am suggesting is a better public presentation of what the heck is going
on with PostgreSQL development.

I think you have a severely flawed idea of how free software development
proceeds.

Then you obviously aren't paying attention. Look at other major OSS
projects. They have these things in place. Even the Linux kernel has a
bugzilla (although I am not advocating bugzilla). Not to mention KDE,
Gnome, Debian..

These projects also have reasonably defined milestones for particular
releases and show status of those milestones during the release.

What you're describing sounds like something a manager of a

commercial project would want. Perhaps it's something the managers of the
people working on Postgres on behalf of some corporate sponsors might want but
in those cases I doubt they would want the information to be public anyways.

What I am describing is what other large OSS projects already do.

In the free software world there's no top-down management of the project with
managers issuing direction and expecting feedback reports.

No but there are people in charge of particular tasks. There are people
only working on certain things. Like the work that the people did on PITR.

People only want

tools that make their lives easier. Not tools that make other people's lives
easier at the expense of their own convenience. The programmers are not
beholden to any corporate interests (other than their own sponsors, who
presumably are getting all the feedback they're looking for privately).

I am not suggesting that anybody be beholden to anybody accept maybe the
community itself.

Sincerely,

Joshua D. Drake