Thread: Crediting reviewers & bug-reporters in the release notes
Bruce, As a fork of the earlier discussion, I think we should credit two groups of people in the release notes: (1) bug reporters / testers (2) patch reviewers My suggestions to keep things simple is, rather than trying to credit people on a per-feature basis, we simply have a listof names at the end of the release notes. Also, that we limit it to people whose contribution to development was significant(i.e. reviewed more than one patch, or spent a lot of time testing and analyzing a bug). -- Josh Berkus PostgreSQL Experts Inc. http://pgexperts.com San Francisco
On May 14, 2011, at 2:14 PM, Joshua Berkus wrote: > As a fork of the earlier discussion, I think we should credit two groups of people in the release notes: > > (1) bug reporters / testers > (2) patch reviewers > > My suggestions to keep things simple is, rather than trying to credit people on a per-feature basis, we simply have a listof names at the end of the release notes. Also, that we limit it to people whose contribution to development was significant(i.e. reviewed more than one patch, or spent a lot of time testing and analyzing a bug). +1 -- Jim C. Nasby, Database Architect jim@nasby.net 512.569.9461 (cell) http://jim.nasby.net
Joshua Berkus wrote: > Bruce, > > As a fork of the earlier discussion, I think we should credit two groups of people in the release notes: > > (1) bug reporters / testers > (2) patch reviewers > > My suggestions to keep things simple is, rather than trying to credit people on a per-feature basis, we simply have a listof names at the end of the release notes. Also, that we limit it to people whose contribution to development was significant(i.e. reviewed more than one patch, or spent a lot of time testing and analyzing a bug). [ Sorry for the long delay in replying.] Putting those names in the release notes sends us down the slippery slope of putting names in there that have no practical purpose. Names on the features tell us the people to contact for problems with the feature Reviewer names, without being assigned a specific features, don't really have any practical value to people reading the release notes, and make the names pure advertising for those people (which is OK in itself but hard to justify in the release notes). Can we do something on the web site or press release on this? One compromise would be to put the reviewers name on the features they reviewed most. That would have some practical value in using their name in the release notes. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + It's impossible for everything to be true. +
On Sat, Jun 11, 2011 at 10:08:07AM -0400, Bruce Momjian wrote: > Joshua Berkus wrote: > > Bruce, > > > > As a fork of the earlier discussion, I think we should credit two groups of people in the release notes: > > > > (1) bug reporters / testers > > (2) patch reviewers > > > > My suggestions to keep things simple is, rather than trying to credit people on a per-feature basis, we simply have alist of names at the end of the release notes. Also, that we limit it to people whose contribution to development was significant(i.e. reviewed more than one patch, or spent a lot of time testing and analyzing a bug). > > [ Sorry for the long delay in replying.] > > Putting those names in the release notes sends us down the slippery > slope of putting names in there that have no practical purpose. With utmost respect, I disagree. Reviewers perform a function essential to our release process, and should get the credit they deserve alongside the coders whose efforts they complement. If it turns out we have an extra screen-full or two of names, that's a small thing. If we slight someone who put in a bunch of effort, that's a much larger problem, as it sends a message about that kind of contribution that we wouldn't actually mean. Cheers, David. -- David Fetter <david@fetter.org> http://fetter.org/ Phone: +1 415 235 3778 AIM: dfetter666 Yahoo!: dfetter Skype: davidfetter XMPP: david.fetter@gmail.com iCal: webcal://www.tripit.com/feed/ical/people/david74/tripit.ics Remember to vote! Consider donating to Postgres: http://www.postgresql.org/about/donate
David Fetter wrote: > On Sat, Jun 11, 2011 at 10:08:07AM -0400, Bruce Momjian wrote: > > Joshua Berkus wrote: > > > Bruce, > > > > > > As a fork of the earlier discussion, I think we should credit two groups of people in the release notes: > > > > > > (1) bug reporters / testers > > > (2) patch reviewers > > > > > > My suggestions to keep things simple is, rather than trying to credit people on a per-feature basis, we simply havea list of names at the end of the release notes. Also, that we limit it to people whose contribution to developmentwas significant (i.e. reviewed more than one patch, or spent a lot of time testing and analyzing a bug). > > > > [ Sorry for the long delay in replying.] > > > > Putting those names in the release notes sends us down the slippery > > slope of putting names in there that have no practical purpose. > > With utmost respect, I disagree. > > Reviewers perform a function essential to our release process, and > should get the credit they deserve alongside the coders whose efforts > they complement. If it turns out we have an extra screen-full or two > of names, that's a small thing. If we slight someone who put in a > bunch of effort, that's a much larger problem, as it sends a message > about that kind of contribution that we wouldn't actually mean. I should back up and explain that the reason for having usernames on release feature items is not to give credit, but rather to assign blame later, in case the features cause problems. The submitter becomes the default go-to person for problems with that item. Having a list of reviewers does not serve that purpose, hence the slippery slope comment. If you put reviewers, then there is no logic that says company names should not be next to contributed features. I am not saying that is wrong, but it is the logical extension. I wonder if the submitter names should not be in the release notes at all. Do other projects put names in there? -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + It's impossible for everything to be true. +
Bruce, > I should back up and explain that the reason for having usernames on > release feature items is not to give credit, but rather to assign > blame > later, in case the features cause problems. The submitter becomes the > default go-to person for problems with that item. Having a list of > reviewers does not serve that purpose, hence the slippery slope > comment. Well, unfortunately, while you and I might understand that, most of our community does not. And to the completely uninitiated,it sure looks like credit. All of this is becoming less of a problem as we update the contributor list regularly. If I can clean up the sponsors listas well, I think we'll hear less complaints. -- Josh Berkus PostgreSQL Experts Inc. http://pgexperts.com San Francisco
Joshua Berkus wrote: > Bruce, > > > I should back up and explain that the reason for having usernames on > > release feature items is not to give credit, but rather to assign > > blame > > later, in case the features cause problems. The submitter becomes the > > default go-to person for problems with that item. Having a list of > > reviewers does not serve that purpose, hence the slippery slope > > comment. > > Well, unfortunately, while you and I might understand that, most of > our community does not. And to the completely uninitiated, it sure > looks like credit. Agreed. > All of this is becoming less of a problem as we update the contributor > list regularly. If I can clean up the sponsors list as well, I think > we'll hear less complaints. Yep. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + It's impossible for everything to be true. +
On tor, 2011-06-16 at 18:24 -0400, Bruce Momjian wrote: > I wonder if the submitter names should not be in the release notes at > all. Do other projects put names in there? In the release notes, I'd say usually not. The names are usually recorded in a changelog, but nowadays that is often generated from the source control system. If I were to look for "blame" in the PostgreSQL source, I'd go to Git. I wouldn't object to removing the names in the release notes.
> If I were to look for "blame" in the PostgreSQL source, I'd go to Git. > I wouldn't object to removing the names in the release notes. So, new policy: no names in the release notes, and we update the contributor list with each release? One thing we'd talked about offlist was having a "this release" list at the bottom of the contributor list, for people who had contributed *only* to the current release. I think it might be time to revisit that. -- Josh Berkus PostgreSQL Experts Inc. http://pgexperts.com
On 06/17/2011 02:15 PM, Josh Berkus wrote: > So, new policy: no names in the release notes, and we update the > contributor list with each release? > I still have a TODO here to prototype something on the wiki to help track sponsored features better. We'd have to do that better in a big way if the release notes are moving in this direction. It's already way too difficult to get people to sponsor features, both for the sponsor and for the people developing it, and this will push some distance toward making it harder. Right now developers can point to the release notes and say "there's a feature like that I helped develop" when trying to convince someone to fund a similar new feature. If the new state of things is that they are just one name on a long contributor list, and you have to hit the individual git commits to confirm what people actually did, it will be a significant step backwards in that situation. The companies who fund features don't have time for that sort of thing. -- Greg Smith 2ndQuadrant US greg@2ndQuadrant.com Baltimore, MD PostgreSQL Training, Services, and 24x7 Support www.2ndQuadrant.us
Excerpts from Bruce Momjian's message of jue jun 16 18:24:16 -0400 2011: > I should back up and explain that the reason for having usernames on > release feature items is not to give credit, but rather to assign blame > later, in case the features cause problems. I call BS on this. This PoV is perfect for justifying that sponsoring companies do not need to get credited, but it's not really the truth. Credit *is* given by having people listed in the release notes, whether you explicitely admit it or not. And pissing off contributors by taking it away is not something to be done lightly. (If assigning blame and being point of contact is really the truth, why is there no email address?) I understand that you don't want to credit sponsoring companies, but I feel that you can decree that as new policy without pissing off individual contributors. If we go the route of Greg Smith's suggestion whereby we assign credit to sponsoring companies in a separate page, that seems to please everyone without collateral damage. I am not saying we should credit reviewers next to each item; but perhaps we can come to some agreement that they are credited elsewhere (for example, maybe in the same page that credits sponsoring companies, or a neighbor page). OTOH I think bug reporters fall in a completely different group than patch reviewers. I mean, they are generally reporting bugs in existing releases; they are not participating in the development process. -- Álvaro Herrera <alvherre@commandprompt.com> The PostgreSQL Company - Command Prompt, Inc. PostgreSQL Replication, Consulting, Custom Development, 24x7 support
\Alvaro Herrera wrote: > Excerpts from Bruce Momjian's message of jue jun 16 18:24:16 -0400 2011: > > > I should back up and explain that the reason for having usernames on > > release feature items is not to give credit, but rather to assign blame > > later, in case the features cause problems. > > I call BS on this. This PoV is perfect for justifying that sponsoring > companies do not need to get credited, but it's not really the truth. > Credit *is* given by having people listed in the release notes, whether > you explicitly admit it or not. And pissing off contributors by taking > it away is not something to be done lightly. I can tell you why _I_ added names to release note items starting 15 years ago. By putting names on the release note items, if a bug was found in a release, I could easily know which developer to contact to get it fixed. I could have trolled the CVS logs, but it is often complex to find the right item. And why not put the developer names in the release notes? Who was going to read it except other developers? Well, a lot has changed in 15 years, and this name thing is only now being revisited, which is fine. I find it a happy coincidence that the names I used to help me are now seen as motivating Postgres contributions. Just a reality check, but I don't think the names in the release notes were originally seen as motivating developers because the assumption was that only a handful of people even cared about our release notes. > (If assigning blame and being point of contact is really the truth, why > is there no email address?) I already had everyone's email address and it was inefficient to type it on every item. I could easily look up their email addresses in my mail program. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + It's impossible for everything to be true. +
Bruce Momjian wrote: > \Alvaro Herrera wrote: > > Excerpts from Bruce Momjian's message of jue jun 16 18:24:16 -0400 2011: > > > > > I should back up and explain that the reason for having usernames on > > > release feature items is not to give credit, but rather to assign blame > > > later, in case the features cause problems. > > > > I call BS on this. This PoV is perfect for justifying that sponsoring > > companies do not need to get credited, but it's not really the truth. > > Credit *is* given by having people listed in the release notes, whether > > you explicitly admit it or not. And pissing off contributors by taking > > it away is not something to be done lightly. > > I can tell you why _I_ added names to release note items starting 15 > years ago. By putting names on the release note items, if a bug was > found in a release, I could easily know which developer to contact to > get it fixed. I could have trolled the CVS logs, but it is often > complex to find the right item. And why not put the developer names in > the release notes? Who was going to read it except other developers? > > Well, a lot has changed in 15 years, and this name thing is only now > being revisited, which is fine. I find it a happy coincidence that the > names I used to help me are now seen as motivating Postgres > contributions. > > Just a reality check, but I don't think the names in the release notes > were originally seen as motivating developers because the assumption was > that only a handful of people even cared about our release notes. > > > (If assigning blame and being point of contact is really the truth, why > > is there no email address?) > > I already had everyone's email address and it was inefficient to type it > on every item. I could easily look up their email addresses in my mail > program. If you want proof, we only started using full names, e.g. not "(Tom)", in 9.0. It didn't matter if users knew who Tom was --- we did. Look at the 6.1 release notes --- they are mostly only first names. Second, if you want to get rid of the names, we can still place them in SGML comments so developers can see who did a feature. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + It's impossible for everything to be true. +
Bruce Momjian wrote: > > > (If assigning blame and being point of contact is really the truth, why > > > is there no email address?) > > > > I already had everyone's email address and it was inefficient to type it > > on every item. I could easily look up their email addresses in my mail > > program. > > If you want proof, we only started using full names, e.g. not "(Tom)", > in 9.0. It didn't matter if users knew who Tom was --- we did. Look at > the 6.1 release notes --- they are mostly only first names. > > Second, if you want to get rid of the names, we can still place them in > SGML comments so developers can see who did a feature. FYI, the general logic was that I used first and last names only if two people had the same first name. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + It's impossible for everything to be true. +
Excerpts from Bruce Momjian's message of jue jun 23 00:13:34 -0400 2011: > \Alvaro Herrera wrote: > > Excerpts from Bruce Momjian's message of jue jun 16 18:24:16 -0400 2011: > > > > > I should back up and explain that the reason for having usernames on > > > release feature items is not to give credit, but rather to assign blame > > > later, in case the features cause problems. > > > > I call BS on this. This PoV is perfect for justifying that sponsoring > > companies do not need to get credited, but it's not really the truth. > > Credit *is* given by having people listed in the release notes, whether > > you explicitly admit it or not. And pissing off contributors by taking > > it away is not something to be done lightly. > > I can tell you why _I_ added names to release note items starting 15 > years ago. By putting names on the release note items, if a bug was > found in a release, I could easily know which developer to contact to > get it fixed. Well, I am not saying that that wasn't the reason you did it. I am only saying that it's not the only purpose that it ended up serving. -- Álvaro Herrera <alvherre@commandprompt.com> The PostgreSQL Company - Command Prompt, Inc. PostgreSQL Replication, Consulting, Custom Development, 24x7 support
Alvaro Herrera wrote: > Excerpts from Bruce Momjian's message of jue jun 23 00:13:34 -0400 2011: > > \Alvaro Herrera wrote: > > > Excerpts from Bruce Momjian's message of jue jun 16 18:24:16 -0400 2011: > > > > > > > I should back up and explain that the reason for having usernames on > > > > release feature items is not to give credit, but rather to assign blame > > > > later, in case the features cause problems. > > > > > > I call BS on this. This PoV is perfect for justifying that sponsoring > > > companies do not need to get credited, but it's not really the truth. > > > Credit *is* given by having people listed in the release notes, whether > > > you explicitly admit it or not. And pissing off contributors by taking > > > it away is not something to be done lightly. > > > > I can tell you why _I_ added names to release note items starting 15 > > years ago. By putting names on the release note items, if a bug was > > found in a release, I could easily know which developer to contact to > > get it fixed. > > Well, I am not saying that that wasn't the reason you did it. I am only > saying that it's not the only purpose that it ended up serving. Agreed. If we agree to make that credit goal overt, there is then little logic to avoid company names in the release notes, except it is more overt than it is now. It is then a question of the level of credit given --- there is no logical prohibition. My big point is that credit has grown out of a practical need to blame developers, and we have never really made the "give credit" decision --- it just happened. Rather than accept what we have now, I think we need to decide on the credit goal and then its use in the release notes. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + It's impossible for everything to be true. +
On tor, 2011-06-23 at 10:59 -0400, Bruce Momjian wrote: > If we agree to make that credit goal overt, there is then little logic > to avoid company names in the release notes, except it is more overt > than it is now. It is then a question of the level of credit given > --- there is no logical prohibition. I think we'd do both the readers of the actual release notes and those who want to be credited a service if we separated those lists anyway. We could just have a list of names after the change list, "The following contributed to this release" or something like that.
On 6/24/11 5:47 AM, Peter Eisentraut wrote: > I think we'd do both the readers of the actual release notes and those > who want to be credited a service if we separated those lists anyway. > We could just have a list of names after the change list, "The following > contributed to this release" or something like that. I'd be fine with that. Or, for that matter, putting it on the web site. Note that we do have a whole list of *code contributors* who aren't getting credited except in the release notes: people who contributed one small patch to one release, and then not again. These folks don't make it onto the Contributors page, so we need to credit them *somewhere*. -- Josh Berkus PostgreSQL Experts Inc. http://pgexperts.com