CWG meeting 2012-03-26

From OpenStreetMap Foundation

Meeting of the Communication Working Group 26th March 2012


License change

IRC log:

21:03 RichardF: good evening CWG
21:03 harry-wood: hello
21:03 rweait1: Hello.
21:04 JonathanB: Yo, dawg...
21:04 harry-wood: hehe
21:04 : fake_mackerski left the room (quit: Read error: Connection reset by peer).
21:04 : fake_mackerski [~fake_mack@] entered the room.
21:05 harry-wood: This evening we could discuss communicating the license change
21:05 RichardF: yes, I think so
21:05 harry-wood: but unfortunately… as far as I can see… we're still stuck waiting for the plan to sure up a bit
21:05 JonathanB: Do we have any facts to work with so far?
21:06 rweait1: Not that I can see.
21:06 JonathanB: What's Matt said? Sorry, haven't been keeping up.
21:06 harry-wood: Fundamental choice between gradual change, or offline period… still doesn't seem to be mde
21:06 harry-wood: Matt's been very quiet about the whole thing
21:07 rweait1: Our problem or not?
21:07 RichardF: from Dermot on IRC:
21:07 RichardF: sorry, not IRC, I mean email:
21:07 RichardF: Placeholder reply: No, not tomorrow. I'm working with Matt to work out a revised plan based on latest information.
21:07 RichardF: "not tomorrow" as in "downtime doesn't start tomorrow"
21:07 JonathanB: We can be proactive in as much as we can tell other groups the information we'd need to put out a proper announcement
21:07 fake_mackerski: pings at the mention of his name
21:08 fake_mackerski: Perhaps I can be of some help
21:08 harry-wood: ah hello fake_mackerski
21:08 harry-wood: probably a good idea to have you join us
21:08 fake_mackerski: Let me give you latest state...
21:08 rweait1: welcome, danfai. Care to let us know who you are?
21:10 : fake_mackerski_ [~fake_mack@] entered the room.
21:10 : fake_mackerski left the room (quit: Read error: Connection reset by peer).
21:10 : fake_mackerski_ is now known as fake_mackerski
21:10 fake_mackerski: Waiting for word from him before posting it to rebuild
21:10 rweait1: we can continue, of course
21:10 fake_mackerski: Sorry, bad internets here
21:10 fake_mackerski: Short version - DB import is really slow
21:10 fake_mackerski: Coupled with an offline rebuild that is probably too much downtime
21:11 fake_mackerski: So the revised plan is likely to involve a live rebuild
21:11 harry-wood: ah right hmm
21:11 rweait1: I don't see a problem with that.
21:11 fake_mackerski: Recall that downtime is still needed, for the server switch
21:12 rweait1: does it matter when that is?
21:12 fake_mackerski: And I'm proposing starting that on the 1st subject to Matt's approval
21:12 rweait1: must it be before or after the rebuild?
21:12 fake_mackerski: Interesting question
21:12 fake_mackerski: Doing it before is tidier and Matt is drawn to doing it that way
21:12 fake_mackerski: But I have no strong feelings on it
21:12 JonathanB: OK, but they're not closely tied?
21:13 fake_mackerski: If the rebuild list debates can reduce to that, it would be really nice
21:13 fake_mackerski: There may be synergies...
21:13 fake_mackerski: I'm thinking of the revised codebase
21:13 fake_mackerski: API in particular
21:13 danfai: Hey sorry, I was not at keyboard a short time.
21:13 fake_mackerski: Matt can say more on that
21:14 rweait1: is curious but not challenging the tech plan.  :-)
21:14 rweait1: is there anything that you'd like CWG to do?
21:14 fake_mackerski: Show solidarity?
21:14 rweait1: +1
21:14 rweait1: "add code" too?
21:14 fake_mackerski: Well, that's a bit simplistic...
21:14 fake_mackerski: Foster useful discussion
21:14 JonathanB: People will want to know what effect on normal mapping, if any, a live rebuild will have.
21:14 fake_mackerski: And yeah, moar code!
21:15 fake_mackerski: JonathanB: Less than an offline one, and not trying to be flippant there
21:15 danfai: I am 'only' a person who was interested about OSM, and I early had ideas of programming tah for BOINC. I am now looking, if there is any interest in it and if so, then I would try to make a team fore this.
21:15 harry-wood: I guess a live rebuild by definition has no effect on normal mapping
21:15 RichardF: we need to say what's happening at the very first opportunity.
21:15 fake_mackerski: Matt has also mentioned that the new server should be better able to take the extra load
21:15 rweait1: Do you expect a consensus on online-offline, shortly?
21:15 fake_mackerski: RichardF: Yes, I want to get this communicated fast
21:15 fake_mackerski: Needs Matt, though
21:15 RichardF: danfai: hello :) this is the OSM Foundation's Communications Working Group meeting. probably not very interesting to you, but you might find stuff in #osm-dev
21:16 rweait1: danfai: nice to meet you. That discussion might be better suited to #osm, or talk@ mailing list.
21:16 fake_mackerski: rweait1: Between Matt and me the consensus is more or less there
21:16 rweait1: How about Frederik?
21:16 harry-wood: hehe. well that's a start
21:16 RichardF: fake_mackerski: good. when you're 95% there, we can publish something on OSMF blog.
21:16 fake_mackerski: He's been demanding the live approach at every opportunity ever since we proposed something different
21:16 JonathanB: Do you need a rubberstamp from a higher body?
21:16 rweait1: so that sounds positive.
21:17 fake_mackerski: JonathanB: Only the community
21:17 danfai: @RichardF @rweait1 I know, I am only listening at the moment
21:17 fake_mackerski: And the sysadmins
21:17 rweait1: :-)
21:17 RichardF: JonathanB: we inform management@ before we do it, with the emphasis on "inform"
21:17 JonathanB: Gotcha]
21:17 fake_mackerski: Once Matt OKs my version of what we discussed I'll be communicating it to MT
21:17 RichardF: danfai: cool. you are very welcome to listen :)
21:17 rweait1: have you had a discussion with sysadmins yet?
21:17 fake_mackerski: Matt holds that line of communication
21:18 fake_mackerski: And he has made sure to fill in TomH better
21:18 harry-wood: awesome
21:18 rweait1: So back to what CWG can do for you?
21:18 fake_mackerski: * Responsible communication as individuals
21:18 rweait1: if you drop us a note with the consensus plan, we can spread it through multiple channels
21:18 fake_mackerski: * Positive engagement in discussions
21:19 fake_mackerski: And yeah, once we have a revised plan that would benefit from commuication, that
21:19 fake_mackerski: Also, perhaps work with Simon Poole to communicate to data consumers
21:19 rweait1: especially if you have any suggestions / requrests ...
21:19 fake_mackerski: He has a doc that is pretty fit for purpose but seems to lack the confidence that it will match the plan
21:19 fake_mackerski: That is, he lacks the confidence
21:20 fake_mackerski: The document is inanimate
21:20 RichardF: best to be safe
21:20 RichardF: shoots document
21:20 rweait1: "messenger, RichardF. Messenger."
21:20 RichardF: heheh
21:21 rweait1: Any thoughts, fake_mackerski: on when we should expect an update?
21:21 rweait1: we can be standing by.
21:21 rweait1: and we have the benefit of multiple timezones.
21:21 fake_mackerski: 2 seconds after I successfully ping Matt
21:21 fake_mackerski: Had him earlier but he had EWG
21:21 fake_mackerski: Is that a pub gathering?
21:22 RichardF: no, IRC
21:22 fake_mackerski: Last contact was a good few hours ago
21:22 fake_mackerski: He's online but not responsive
21:22 rweait1: fake_mackerski: just let us know.
21:22 fake_mackerski: And he knows he has it in his mailbox
21:22 fake_mackerski: Sure will
21:22 fake_mackerski: And I appreciate the support
21:22 rweait1: even if you give us an hours notice, we can have somebody waiting for the note, and to redistribute it.
21:22 fake_mackerski: It's been… Interesting just lately
21:23 fake_mackerski: I will probably do a 2-phase thing
21:23 fake_mackerski: First MT and Frederik
21:23 fake_mackerski: I want to de-scary this for him
21:23 : toffehoff [~toffehoff@] entered the room.
21:23 fake_mackerski: Are the important folk among you already getting MT?
21:23 fake_mackerski: MT mail, that is
21:24 rweait1: How does one answer that with minimal ego?
21:24 toffehoff: Hi all, sorry I'm late.
21:24 harry-wood: not really. I'm getting it.. and richardF is I think
21:24 harry-wood: oh and toffehoff is too of course
21:24 fake_mackerski: rweait1: I'll CC you if you are not on it
21:24 JonathanB: isn't
21:24 rweait1: I'm not. JoanathanB?
21:25 rweait1: :-)
21:25 fake_mackerski: But consider it embargoed until I push it to rebuild@
21:25 fake_mackerski: I mean, if somebody is screaming mental about it you might reassure them based on what you know
21:25 fake_mackerski: But I'm just trying to be careful. Really careful
21:25 RichardF: understood.
21:25 rweait1: do we have an idea about the off-line details for the server switch?
21:25 harry-wood: how much of the above conversation is also embargoed?
21:25 RichardF: there are certainly people actively looking out for slipups.
21:26 fake_mackerski: harry-wood: I think it's probably OK
21:26 harry-wood: alright
21:26 fake_mackerski: It's nothing I wouldn't tell anybody who asked and seemed willing to apply logic
21:26 harry-wood: sometimes people read the minutes… weirdly
21:26 fake_mackerski: I think less mystery is the key to this one
21:27 rweait1: You shouldn't be walking on eggshells. You're working like a maniac.
21:27 fake_mackerski: You have no idea ;)
21:27 fake_mackerski: I'm also writing a paper and getting a job
21:27 fake_mackerski: The cure for cancer is coming along nicely too
21:27 rweait1: Good to know.
21:28 rweait1: If we're keeping you. Feel free to dip out and drop a note to communication with more requests later.
21:28 fake_mackerski: OK, that suits
21:28 fake_mackerski: Thanks guys!
21:28 rweait1: Thank YOU.
21:29 RichardF: +1
21:29 harry-wood: So that's good. Sounds like there's a plan approaching imminently
21:29 rweait1: anything else for mackerski before he goes?
21:29 rweait1: or should we move to other CWG business?
21:30 rweait1: Is there any other business?  :-)
21:30 harry-wood: well I was pondering what the main communications need to be about this whole thing. Don't need hold up mackerski with that though
21:30 rweait1: or am I just offline and don't know it?
21:30 fake_mackerski: harry-wood: Parting comment on that...
21:30 rweait1: I think it's much simpler while on-line ...
21:30 harry-wood: probably notifying diff consumers is priority number 1
21:30 fake_mackerski: Most mappers don't care at all
21:30 RichardF: our three main audiences are mappers, data consumers, press.
21:31 RichardF: mappers want to know how it affects their mapping.
21:31 fake_mackerski: Some know it's important but can't easily grok the issues and are easily worried
21:31 fake_mackerski: Those will be happy to be told one day that it has happened
21:31 fake_mackerski: Well, relieved anyway
21:31 RichardF: ODbL FAQ on the wiki should be helpful.
21:31 rweait1: Thanks for that RichardF
21:31 fake_mackerski: Those who really care and need to are on rebuild@ or should be
21:32 RichardF: but I think that, for mappers, the main thing is to get the message out asap.
21:32 rweait1: So, as a rough plan:
21:32 RichardF: it doesn't have to be particularly detailed, just reassuring that "it's going to plan. hell, there is even a plan now."
21:32 rweait1: 1) announce what we can, when we can
21:33 RichardF: yep.
21:33 rweait1: 2) do a license upgrade in progress on 01 April
21:33 harry-wood: I think mappers would be interested if there's a going to be a cut off
21:33 rweait1: then 3) it's done when complete,
21:33 harry-wood: but only as sort of passing interest. it's not an important matter of notification
21:33 rweait1: Well there is the DB downtime for switchover.
21:33 rweait1: we'll want to get that out when we can.
21:33 rweait1: but it can be a separate notice.
21:34 RichardF: so the first communication I guess needs to be "we are beginning switchover on 1st April and not before"?
21:34 rweait1: I kind of like starting on 31 March, just because.
21:34 rweait1: I don't much care when it finishes.  :-)
21:34 RichardF: timezones ftw ;)
21:34 fake_mackerski: Like I said, hold onto that one until Matt can see if that fits
21:34 RichardF: oh yeah of course
21:35 rweait1: Not my call. Just a dream.  :-)
21:35 fake_mackerski: But that's how it's playing out in my mind
21:35 RichardF: nothing's going out yet, we're just discussing what looks likely
21:35 RichardF: personally I hope it takes a week so I can have a holiday!
21:35 harry-wood: when will the diff stream move to new URLs? before or after gradual "live" rebuilding?
21:35 rweait1: is it wise to suggest multiple short API outages for the duration?
21:36 rweait1: has to be after, harry-wood.
21:36 rweait1: Concurrent with the license switch.
21:36 fake_mackerski: One proposal is to stop the first diff feed before we start rebuilding
21:36 rweait1: all the cleaning is prior to the switch.
21:36 rweait1: I have no argument with that.
21:37 fake_mackerski: Avoids huge diffs and makes it harder for unplanned consumers to wait and see how much stuff has been deleted
21:37 rweait1: (speaking as a data consumer)
21:37 fake_mackerski: Sorry, makes it easier
21:37 toffehoff: Will the diffs during the rebuild also contain the clean-up edits?
21:37 rweait1: Can I help by working on the timeline document or something?
21:37 toffehoff: If so, it might be good to mention that....
21:38 RichardF: fake_mackerski: yes, and I'd also like data consumers to have to make a decision about when they switch back on
21:38 RichardF: they may decide to wait a week after the change, for example.
21:38 rweait1: toffehoff: I think he just said diffs might be off for the duration.
21:38 toffehoff: ok
21:38 harry-wood: somebody leaving a server to do it's own thing will they be alarmed to see lots of data disappearing suddenly. or just alarmed that their diffs broke one day?
21:38 toffehoff: Missed that I guess.
21:39 rweait1: harry-wood: If you are operating a server you shouldn't alarm easily. :-)
21:39 RichardF: breaking is better.
21:39 rweait1: i think diff-stop, is less surprising.
21:39 RichardF: we really do not want to have to deal with questions "how do I reverse these diffs?"
21:39 rweait1: I haven't seen a diff stop recently, but they did happen...
21:39 toffehoff: ... and easier to communicate (a diff-stop)
21:40 rweait1: fake_mackerski: again. Let me know if I can help by transcrbing something. I can hang out while you and the others talk on IRC or whatever.
21:40 rweait1: and I'll use a spel chekker.
21:42 harry-wood: perahps there should be two diff URL moves :-/ one to stop servers loosing data during the rebuild, and one to be clear that it's "ODBL from now onwards"
21:42 harry-wood: but let's leave this planning for dermot & matt
21:43 fake_mackerski: rweait1: Sure, thanks
21:43 fake_mackerski: RichardF: Yes, "we" feel the same about the active decision
21:44 harry-wood: Communicating what will happen with the diffs is a priority. Do we anticipate putting out that kind of information tomorrow then?
21:44 fake_mackerski: harry-wood: I suggest talking to Simon on that
21:45 fake_mackerski: He has the exact document
21:45 fake_mackerski: He just didn't know what to say about diff halting
21:45 rweait1: but he doesn't have the decision or the time.
21:45 rweait1: so tell us when and what, and we can edit his draft.
21:45 fake_mackerski: At LWG and elsewhere, halting was plainly identified as the Right Thing
21:45 rweait1: Simon's draft is good. Thanks to him, of course.
21:46 fake_mackerski: When isn't IMO important for the doc...
21:46 fake_mackerski: It's the order that's important
21:46 rweait1: That's decided then?
21:46 fake_mackerski: "Soon" can work, I think
21:46 fake_mackerski: Or perhaps talk about the 1st April plan and indicate that "the change" will cause xyz
21:46 rweait1: diffs off; start cleaning; new address ; license change; planet; diffs on?
21:47 fake_mackerski: Yes, that's what seems sane to me
21:47 fake_mackerski: doesn't admin the diffs, though
21:47 rweait1: does Matt?  :-)
21:47 fake_mackerski: But the idea of a change has been indicated to be sane and possible
21:49 rweait1: Should we plan on a progress report every two - three days during the cleanup?
21:49 rweait1: in other words, when will we know if the clean will take two weeks or two months?
21:49 rweait1: I know that we don't know this yet...
21:50 harry-wood: sounds good to me. But getting ahead of ourselves
21:51 harry-wood: It's the april 1st and what needs communicating before that, that I'm worried about
21:51 rweait1: Dermot, want to get back to it and let us know, ewhen you can?
21:51 RichardF: well, let's start three Google docs or whatnot, and draft our comms there.
21:52 rweait1: anything else for CWG?
21:52 rweait1: I sent a note about surveys recently.
21:52 harry-wood: oh yeah? on talk?
21:53 rweait1: on communication@, iirc
21:53 harry-wood: oh i see yeah.
21:53 rweait1: but there might have been a diffeent one on talk. :-)
21:53 harry-wood: I didn't see the original emails. People posting annoying surveys?
21:53 rweait1: I say yes, many others disagree.
21:54 rweait1: rather, too many surveys.
21:54 rweait1: We want to know more about our users. We should take charge of how they are surveyed.
21:54 rweait1: But. That's something $SOMEBODY would have to do, even if we decided that we should...
21:55 rweait1: So maybe it;s a dead idea.
21:55 : fake_mackerski left the room (quit: Ping timeout: 480 seconds).
21:55 rweait1: From the rousing support, I presume we can move on to other topics.
21:56 harry-wood: could have a policy stating that nobody should post surveys on the mailing lists… (which those who don't read the policy will not know about) …and that nobody should fill them in if they do
21:56 harry-wood: *suggesting* that nobody should fill them in. We can't stop people
21:56 RichardF: I personally don't have a huge problem with it. we are usually too busy shooting ourselves in the foot to worry about Google doing it for us.
21:56 JonathanB: The point is valid: it's (too) easy to post to talk asking people to talk part, regardless of the value of the survey
21:57 toffehoff: Well, I do see some advantages in trying to streamline surveys and also having access to the results.
21:57 rweait1: There are things we could do with a survey, that outsiders could not.
21:57 rweait1: We could present them to random new signups over a period of time.
21:57 RichardF: anyone who posts a survey request to talk will have their sample skewed by large proportion of unhinged nutters, anyway ;)
21:58 toffehoff: however, I also feel there is no real urgency right now. We could leave this to when the ODbL change is on it's way smoothly.
21:58 rweait1: we can distinguish users who update their answers from new answers
21:58 rweait1: Absolutely.
21:58 rweait1: no rush on this.
21:58 harry-wood: ok. Is there a rush on any other topics?
21:59 toffehoff: Nothing come to mind ....
21:59 RichardF: not AFAIK.
21:59 rweait1: So are we done for the day?
22:00 harry-wood: seems there'll be more to talk about tomorrow
22:00 rweait1: :-)
22:00 toffehoff: :-)
22:00 rweait1: I'll be on #osm
22:00 rweait1: everybody else as well?
22:00 JonathanB: Sometimes
22:00 rweait1: maybe we can use this channel as well if needed?
22:00 harry-wood: yeah I was just wondering that
22:01 rweait1: danfai, did you enjoy the meeting? Are you anxious to join a working group now? :-)
22:01 harry-wood: hehe
22:02 toffehoff: Hi danfai, didn't notice you on the list. Welcome.
22:02 danfai: hey
22:02 rweait1: are we adjourned then?
22:02 JonathanB: I'm good
22:03 toffehoff: Looks like it.
22:03 rweait1: ttfn, all
22:03 harry-wood: So maybe we can agree to all keep our eyes on the emails tomorrow. Get read for announcements… but maybe also have a meeting tomorrow if needed
22:03 toffehoff: ok
22:03 rweait1: I'll be on the road for a few hours. I'll watch for emails into the wee hours. Somebody plan to be up early (UK) to cover me?
22:03 danfai: yes I enjoyed, but I will watch for the first moments and perhaps later I can join a group
22:04 harry-wood: I'm normally doing portuguese classes on Tuesdays, but they finished for the term
22:05 toffehoff: No problem danfai. Don't forget to shout when things are going to fast or you don't understand things.
22:06 : rweait1 left the room.
22:06 toffehoff: Going to catch some sleep now....
22:06 toffehoff: Bye!
22:06 danfai: hehe, I wont shout so fast, if this is topic I'm not really fit in
22:06 danfai: bye
22:07 harry-wood: bye all
22:07 : toffehoff left the room (quit: Remote host closed the connection).
22:08 harry-wood: Oh by the way… exciting communications innovation —>