ContributorREVU

Please check the status of this specification in Launchpad before editing it. If it is Approved, contact the Assignee or another knowledgeable person before making changes.

Summary

Permit Contributors to review and comment on new packages in REVU in such a way that a Contributor review does not count as a rejection, and a Contributor cannot advocate.

Release Note

REVU Now supports Contributor comments. Members of ~revu-uploaders may suppport the reviewing process by commenting on new uploads.

Rationale

REVU should be used as an asynchronous packaging feedback tool.

  • Contributors will learn to review and be confronted with different packages and packaging styles
  • Contributors can cross-check their packages and share ideas and solutions
  • Permitting contributors to comment on new packages could speed up the review process

Contributors are able to help each others on other channels (bugs, answers, forums, IRC, wiki, etc.), they should also be able to post comments on REVU.

Use Cases

  1. Alice is a Contributor with experience in python packaging. Alice notices a new python package on REVU and leaves a comment to share her experiences with the packager.
  2. Bob is a Contributor very active in NEW packaging, and wants to reduce the review load for MOTUs. Bob reviews packages on REVU, and comments regularly on all aspects of the packaging.
  3. Chris is a MOTU with limited time. Chris reviews the REVU list, notices Bob's comments, and leaves a short comment asking for a cleanup of the issues Bob mentioned, pushing the package into the "Needs Work".

Design

  • A Contributor comment is treated like a Packager comment, (no allowed to advocate, no treated as rejection).
  • It is felt that there should be no additional distinction between comments from a contributor and those from MOTUs, just like for any other channels: IRC, LP, etc.
  • Contributors should be encouraged to review and comment other packages, but should also be warned about posting inappropriate or derogatory comments
  • Contributor comments should not push the package into the "Needs Work" list.

Implementation

Change REVU code to support contributor comments. Test on the REVU staging server. Deploy to REVU production server.

Code Changes

  1. Don't automatically exit commenting option when level is Contributor:

=== modified file 'details.py'
--- details.py  2007-11-10 20:45:38 +0000
+++ details.py  2007-12-07 22:29:35 +0000
@@ -115,7 +115,7 @@
        if login is None:
                return  'please login at main page'
        if (level == 'Contributor') and (_getuploader(c, upid) != login):
-               return 'Sorry, Commenting for contributors only on their own uploads'
+               print 'WARNING: please respect others people work, use constructive remarks'
        tmpl = cfg.getTemplatesPath('commentform.html')
        vars = { 'login' : login,
                 'upid': upid,
  1. Disable advocation checkbox when level is Contributor:

patch pending
  1. Modify sorting criteria to not push Contributor commented packages to the "Needs Work" queue

patch pending

StefanPotyra: After a glimpse at the big sql statement, imho no code change is needed here.

Outstanding Issues

Do we need comments moderation, scoring, threading, ... ?

Emmet Hikory - I think we don't need moderation: CoC applies. I don't think we need scoring, as most review comments are either actionable with a new upload or represent advocation (whether MOTU advocation or just Contributor satisfaction). I don't think we need threading, as focus should be on reviewing the package, rather than reviewing reviews.

[StefanPotyra] - I'd like to mark Contributor comments as such. Maybe MOTU Comments could display a MOTU icon next to the text marking these as MOTU Comments? Quick update: MOTU icons done in trunk, deployed to staging atm.

BoF agenda and discussion

proppy: what do you think about MOTU helpfull  revu others MOTU packages
proppy: ?
proppy: (sorry)
sistpoty: [TOPIC] having motu hopefuls review other MOTU packages
MootBot: New Topic:  having motu hopefuls review other MOTU packages
ScottK: proppy: As lolng as you make it clear you're not a MOTU, I think it's great.
imbrandon: we have a bad habbit about making things "required" though before they are put into use for a while, unlike debian policys where they are used for a long time before they are policy
persia: proppy: It's good.  It's encouraged, and no they can't have access to REVU to do it yet.
proppy: and MOTU reviewing you'r review
sistpoty: persia: they can, but only on a individual basis
ScottK: proppy: If you do the review interactively on #ubuntu-motu then that happens automatically.
Nafallo: imbrandon: agreed.
persia: sistpoty: Yes, but we can't distinguish Contributor comments from MOTU comments for sort order, so it breaks the MOTU review of Contributor comments part of the goal.
ScottK: sistpoty: That needs to be treated as a special case and not as a general capability right now.
persia: proppy: If you're worried about flooding the channel, use a pastebin to write the review, and post the link to #ubuntu-motu
sistpoty: persia: right. but I guess someone who does a good job for some time should get review rights...
proppy: persia: ok
proppy: it was just a suggestion 
sistpoty: ScottK: exactly, that's what I wanted to write
Nafallo: sistpoty: shouldn't that person be a MOTU as well then?
DktrKranz: I like this proposal. A non-MOTU can provide useful feedbacks, reviewer will choose to follow these advices or not.
ScottK: proppy: It's a good one.
persia: sistpoty: I disagree.  If someone is doing an excellent job of reviewing, they should be prodded to apply for MOTU.
proppy: maybe a threaded forum is more approriate for that than revfu
ScottK: persia: Think of it as a final exam.
sistpoty: Nafallo, persia: sure, but there are always delays, and I recall having someone grant review rights as a task of the application *g*
persia: ScottK: For some special cases, maybe.
sistpoty: I guess we all can agree that the real bug is in revu, which should get fixed, right?=
ScottK: And since we haven't mechanized a way to distinguish MOTU/non-MOTU comments, then a special case is all it can ever be until someone codes the changes in REVU.
persia: sistpoty: That might happen, but it's part of the application process, not a general case: Contributors should first prove their reviewing skills in #ubuntu-motu
sistpoty: persia: I guess we agree there :)
proppy: what about posting review in a mailing list then ?
sistpoty: proppy: that mailing list exists already: motureviewers@tauware.de ;)
proppy: then distingish between motu and non motu is just a matter of identify ?
persia: proppy: Be warned that not every uploader reads that list (nor every reviewer)
proppy: like an @ubuntu.com email ?
sistpoty: motu-reviewers even
ScottK: proppy: The trick is the data base schem needs changing.
persia: proppy: Essentially.  It's not email, it's membership in ~ubuntu-dev
ScottK: proppy: Not all MOTUs use their ubuntu.com address.
sistpoty: ScottK: we could abuse the current scheme a little bit... I guess there's some space left which would require only minimal changes (but those with care *g*)
proppy: ScottK: you mean not at all ?
proppy: ScottK: or not for reviewing business ?
Nafallo: @ubuntu.com is for members, not developers.
ScottK: sistpoty: Excellent.
ScottK: proppy: I meant not at all.
proppy: ScottK: ok ne
***ScottK for example.
***persia proposes the drafting of a spec for how non-MOTUs would comment, for general review
sistpoty: [IDEA] fix revu to allow non-motu contributors
MootBot: IDEA received:  fix revu to allow non-motu contributors
sistpoty: persia: that would be excellent, would you take care for this?
proppy: thanks for feedbacks
persia: sistpoty: I don't understand the problem well enough: I was happy using pastebin as a Contributor.
persia: proppy: Would you be up for drafting a spec for how Contributors would add REVU comments?
ScottK: proppy: How about you make a first draft?
proppy: persia: np
persia: proppy: Thanks.
imbrandon: pastebin requires IRC :(
proppy: persia: the difference betweend pastbin and revu
sistpoty: persia: it would be in place reviews... debian-mentors acts quite like this
persia: imbrandon: I don't consider that a downside.
proppy: persia: is that pastebin is live
proppy: persia: REVU is asynchronous
ScottK: proppy: It alse needs to talk about how it gets decided who can review.
sistpoty: (the mailing list, not the webpage=
imbrandon: persia: i do, i know many contributors i have worked with will NOT cone to IRC
imbrandon: come*
persia: proppy: True.  I'll probably have more useful input when more awake :)
sistpoty: hehe
proppy: should I use the spec template ?
imbrandon: plus irc limits you to those that are in your timezone
imbrandon: or sleep/schedule
ScottK: imbrandon: Sleep is for the weak.
imbrandon: hehe
persia: proppy: That would be best.
sistpoty: proppy: I guess just informative text is nice.. of course you can do an "official spec" if you want ;)
proppy: persia: url ?
proppy: persia: I mean of the wanted spec
proppy: persia: not the template
imbrandon: ok i got to run, i trust someone will post the minutes to the ML ?
sistpoty: imbrandon: sure, I'll do
ScottK: Maybe sistpoty would volunteer to mentor proppy on spec writing....
imbrandon: killer, thanks
***persia suggests https://wiki.ubuntu.com/Spec/ContributorREVU but thinks someone else might have a better suggestion.
proppy: https://wiki.ubuntu.com/MOTU/NonMOTUReviewProcess ?
sistpoty: ScottK: /me needs the spare time for rewriting revu :P
persia: sistpoty: You don7t get to rewrite until there's a spec, no? ;)
proppy: I suck when it comes to choosing CamelCasedNames
DktrKranz: sistpoty, is source code available somewhere?
imbrandon: sistpoty: i just started proding the codebase too and did you see the staging site yet ?
persia: DktrKranz: In the LP project
DktrKranz: ah, I'll have a look, then
proppy: [LINK] https://wiki.ubuntu.com/Spec/ContributorREVU
MootBot: LINK received:  https://wiki.ubuntu.com/Spec/ContributorREVU
Spec left the room ("what an unfortunate name :p").
sistpoty: DktrKranz: sure, https://code.launchpad.net/~revu-hackers/revu/trunk
sistpoty: imbrandon: no, not yet
sistpoty: ok, shall we move on?
***persia seconds


CategorySpec

Spec/ContributorREVU (last edited 2008-08-06 16:33:54 by localhost)