Working Group Minutes/DWG 2013-02-07

2013-02-07 DWG meeting minutes


Minutes by: Paul Norman


IRC nick Real name
pnorman Paul Norman
sly Sylvain Letuffe
rweait Richard Weait
woodpeck Frederik Ramm


  • Matt Amos
  • Eugene Sandulenko
  • Henning Scholland
  • Grant Slater


  • Meetings
    • Frequency
    • Meetings
    • Non-meeting minutes
  • 2013 workplan
  • OTRS alternatives
    • report from Sly on french use of Trac
  • Need for policy/guidance work
  • DMCA
    • bulk edits
    • imports
    • mechanical edits
  • "bully" editors leaving project after being told off by DWG - need different approach?


  1. Meetings
    • Most work done by email
    • Consider merits of meetings vs only email/irc after this meeting
    • Should produce either reports or minutes, depending on above
    • Reports would be based on emails and work done by the DWG
  2. 2013 Workplan
    • Amend existing draft to add
      • specific mention of DMCA takedown notices, in conjunction with LWG where necessary
      • Add mention of reports or minutes as discussed above
      • PDF, review and post
  3. OTRS alternatives
    • French having some success with trac
      • Not perfect tool
      • Not linked to OSM
      • Public instance
    • Lack of interface between OSM messages and OTRS an issue
    • Many OTRS features not useful when all messages have to go through
    • email <-> OSM gateway or ability to CC users would help with OTRS
    • A list of desired features in a tracking system needs to be created. Follow-up to be done on dwg list
  4. DMCA
    • Paul to draft text to send to board+lwg on DMCA process and guidance required
  5. Bulk edits (imports and mechanical edits)
    • structured discussion is necessary
    • No process decided on for implementation, but drafts need to be discussed with community
    • Value in keeping mechanical edits and imports as seperate policies, even if there is duplication between them
  6. Bully editors
    • Discussion deferred due to time constrains

Action items


  1. pnorman to PDF workplan and send to list
  2. pnorman to draft text on DMCA
  3. Frederik to start discussion on bully editors on DWG list


  1. List of tracking system features to be developed


Timestamps are PST.

10:48 -!- sly has joined #osm-dwg
10:49 <sly> hi every one
10:49 <pnorman> Hey
10:50 <pnorman> i'll reply to that topic when logins come back up
10:55 -!- woodpeck has joined #osm-dwg
10:56 <pnorman> hey
10:57 <rweait> Hey.
10:57 <pnorman> since I arranged the meeting I guess I'll chair
10:58 <woodpeck> ok
10:58 <rweait> Are we quorate?  
10:59 <woodpeck> does somene have a log file?
10:59 <pnorman> henning is on vacation, matt and grant didn't respond
10:59 <woodpeck> s/have/record/
10:59 <pnorman> Ya, I log everything by default
10:59 <woodpeck> assuming we'll publish?
11:00 <rweait> I think we should, except any parts we need to redact for privacy.
11:00 <pnorman> we may want to turn it into minutes - it depends if we get into cases
11:00 <woodpeck> agreed
11:00 <sly> +1
11:00 <rweait> Agenda?
11:00 <woodpeck> i can upload to osmf wiki
11:00 <woodpeck> agenda link is in the channel message
11:01 <pnorman> see topic - i'll stick the text into the log
11:01 <pnorman> first of all, meeting frequency. we should really have meetings at least quarterly if not more often. thoughts?
11:02 <woodpeck> if they prove useful - yes
11:02 <woodpeck> if they prove a waste of time and are just there to have a meeeting - no ;
11:02 <woodpeck> ;)
11:02 <woodpeck> it does look better on the outside if we have regular meetings
11:03 <rweait> We do discuss our work pretty frequently by email and irc.
11:03 <woodpeck> in lieu of meetings we could also create a quarterly "report" that we publish
11:03 <woodpeck> explaining what we discussed and decided
11:03 <rweait> Perhapswe can plan on another meeting in 3 months and evaluate at that time.
11:03 <woodpeck> +1
11:04 <pnorman> perhaps rediscuss at the end of this meeting?
11:04 <rweait> We're set for one hour?
11:04 <woodpeck> more would be torture
11:04 <rweait> :-)
11:04 <rweait> not with this crew.  We're the reasonable ones.  Remember?
11:05 <pnorman> I like the idea of some kind of regular report since most of what we do is done by email
11:05 <rweait> I like the idea of that being automated, if it is done at all.
11:06 <rweait> sly, anything to offer?
11:06 <sly> nope
11:06 <woodpeck> btw _sev are you here or just idle?
11:06 <sly> as long as I'm not writing the report
11:06 <woodpeck> sly: :-)
11:07 <pnorman> i suspect it'll have to be manual, turning a month or quarters worth of emails into a short report isn't the type of thing you can do automatically
11:07 <woodpeck> let's say we want something to show when people ask what the dwg does - either reports or minutes. and discuss details later. ok?
11:07 <rweait> +1
11:07 <pnorman> +1
11:08 <rweait> next item pnorman?
11:08 <rweait> 2013 workplan?
11:08 <sly> about reports (since I don't have much to say about it) why not wait for someone to ask us one ?
11:08 <pnorman> 2013 workplan. sly had expressed the concern that one of the items was too perscriptive
11:09 <rweait> sly: then reporting has to be done as an "emergency". doing something on a schedule can be planned.
11:09 <pnorman> I think we have had requests
11:10 <sly> well, my answer to such request would probably be "here you go, that's a dump of 3 month emails please go ahead", but I guess I'm not really suited for reports and diplomatie ;-)
11:11 <pnorman> the problem is that there is likely going to be private stuff in that dump of emails
11:12 <sly> My comment would then be +0 = no opinion
11:12 <woodpeck> let's discuss the workplan
11:12 <sly> yes
11:12 <woodpeck> let's discuss the workplan
11:12 <sly>  sly had expressed the concern that one of the items was too perscriptive > what item are you refering to ?
11:13 <pnorman> the "action" stuff
11:13 <woodpeck> can you post a link to your latest version?
11:13 <woodpeck> i'm without my email archive a.t.m.
11:14 <woodpeck> dropbox or something
11:14 <pnorman> [temporary link omitted]
11:16 <sly> pnorman: about action, since that was internal draft, it is only a small matter of words, I don't feel we really to change it for the use we have
11:16 <sly> *need
11:16 <pnorman> PDF it and stick it on the wiki, replacing the 2012 workplan?
11:17 <sly> If we are not willing to go public to ask contributors what priority they would like adressed, then yes, just let is like this
11:17 <woodpeck> suggest to add specific mention of handling dmca takedown notices (where necessary in team with lwg)
11:18 <sly> *is/it
11:18 <pnorman> I'll add that to copyright
11:20 <woodpeck> instead of "re-start regular meetings", put a wording that says what we agreed on above - meetings+minutes or perhaps just regular reports
11:20 <pnorman> k - I'll amend it based on the meeting and re-send
11:20 <rweait> Shall we move on?  Time's a wastin'
11:21 <pnorman> OTRS alternatives. sly, you mentioned that there were people in france trying out trac. how's that working out?
11:22 <sly> I'd say, well enough
11:22 <sly> still, that isn't the perfect tool for that either
11:22 <sly> but we have what we need :
11:23 <sly> anonymous report (people are often shy)
11:23 <sly> mailing list forwarding
11:23 <sly> close/re-open/attribute to/etc.
11:24 <sly> and, well, the tool was allready up on our servers, we just added a new component to trac
11:24 <sly> but we wanted to be public from the start
11:24 <pnorman> at this point, we're absolutely failing at using OTRS. heck, I can't find an old ticket half the time, even when I know its been inputted
11:24 <sly> so I have no clues about how it performs in private needs
11:25 <sly> pnorman: I must admit I didn't know OTRS nor trac, the learning curve was much faster with trac
11:26 <sly> but well, the tool is secondary, we must just ask ourself (IMHO) who can report, who can read and what are our needs, the tool might follow
11:26 <sly> we don't use OTRS, I think, because it just doesn't bring much more that the mailing list
11:27 <sly> *than
11:27 <sly> only when a problem does I linger I find the need for keeping track of it
11:27 <pnorman> it'd be nice if we could look up a ticket and actually get the status/history of it
11:27 <woodpeck> i think we use OTRS little becuase we don't really care to learn it ;)
11:27 <woodpeck> of course otrs does give you a proper history 
11:28 <woodpeck> only thing that we could use and that it doesn't do (or at last i couldn't find out how) is to communicate with different external parties within the same ticket
11:28 <woodpeck> also, but that is a shortcoming of the osm website, we can't email users directly, i.e. i cannot send something from otrs to an osm user
11:29 <woodpeck> unless i had prior communication and therefore a temporary email alias
11:29 <woodpeck> we could just ask to have access to the username<->email database of the api
11:29 <woodpeck> we don't have that access currently
11:30 <pnorman> yes, that's a pain - all of OTRS' email communcation stuff is pretty worthless for 80% of cases where we're contacting through the OSM site
11:30 <woodpeck> should i inquire with the admins? it is possible that this would require us to sign some data protection thingie since osm has to protect that email list.
11:30 <sly> There are 2 reasons why we didn't connect our trac to the osm user base :
11:31 <sly> 1) lazyness of hacking
11:31 <sly> 2) wish that external parties could still produce ticket with complaints
11:31 <sly> Like : "Why is my name displayed on the map"
11:31 <rweait> sly you mean trac-users != osm-users?
11:32 <sly> rweait: positive
11:32 <woodpeck> i have suggested that we create some kind of permanent email-to-user gateway (i.e. you email to and it comes out in my inbox) but tom has pointed out that user names can contain characters that are not valid in email local parts.
11:32 <sly> woodpeck: I suppose, by reading you, that there aren't API for message sending ?
11:32 <sly> *api calls
11:33 <woodpeck> not really, of course you can use the same http post message that is sent when you message a user
11:34 <sly> ok
11:34 <pnorman> the OSM messaging system has no ability to CC either, if we had that we could CC an account that would go to OTRS or trac.
11:34 <woodpeck> but having such an email gateway would mean i could easily include someone in e.g. otrs without having to hack up a module for it
11:34 <sly> I guess it would be very fun to mess with cookies and POST requests
11:34 -!- _sev has quit [Quit: This computer has gone to sleep]
11:35 <woodpeck> i think it would not be difficult to fiddle with the interface but you'd have to write a trac/otrs/... component and that's what i am not experienced with
11:35 <sly> woodpeck: that is a perfectly valid remark
11:37 <woodpeck> i wonder if everyone who currently doesn't use otrs. or uses it little, 
11:37 <woodpeck> would suddenly use trac much more ;)
11:37 <woodpeck> because otherwise it's not worth discussing
11:37 <rweait> Sounds like we want something other than otrs, but don't know exactly what.  Should we conintue on mailing list?
11:37 <sly> that is a good question ;-)
11:38 <sly> What I'm missing with otrs is the otrs->mailing liste gateway
11:38 <rweait> +1 sly
11:38 <pnorman> I'd like to use OTRS more but I always end up struggling for 15 minutes trying to do what I want and then giving up after the first item on my list
11:38 <sly> that would save 1 out of 3 copy/paste of a message
11:39 <woodpeck> i suggest that, one the mailing list, we make a list of things we wouldlike to have from our ticketing system, then we try to find a trac guru and an otrs guru or whatever other gurus there are and ask them: do you think your system could do this with some fiddlng if ncessary?
11:39 <rweait> woodpeck +1
11:39 <pnorman> k
11:39 <pnorman> +1
11:39 <sly> (1 to copy the message you sent to an osm user, 1 to send the dwg mailing list you have done so and opened a ticket, and 1 on otrs itself)
11:40 <woodpeck> if we KNOW that trac or otrs or ... can do what we want and we just have to find out how, then that's a good incentive to work with that system.
11:40 <rweait> sly, okay to move to next topic?
11:40 <sly> 'k
11:40 <sly> +1 for the guru hunt
11:41 <pnorman> so, we have a few items where some policy or guidence work should be done
11:43 <pnorman> DMCA - we got our first take-down on the 27th or so and the response was a bit uncoordinated
11:44 <woodpeck> before that there were tons of spam messages via the form, which means that it is possible a real dmca notice might go unnoticed
11:45 <woodpeck> i think as long as the volume is low like now, we could simply carry on like this - whoever takes the ticket, takes it and says so on the list
11:45 <woodpeck> but if people prefer we could also introduce a rotating DMCA duty, where everyone has the DMCA hat on for a week or a month or so
11:46 <rweait> pnorman, could you prepare a DMCA checklist to show the steps and checkpoints for each dmca matter?
11:46 <sly> I'm okay with a few spam
11:47 <sly> but not okay to have the rotating hat

[portions of DMCA discussion removed]

11:54 <woodpeck> so what kind of policy/guidance work is required do you think for dmca?
11:54 <woodpeck> eager to move to next topic.
11:54 <pnorman> probably me drafting something, sending to board+lwg?
11:55 <sly> I have no opinion at all since I'm far from understanding anything about abroad laws
11:56 <pnorman> +1 for pnorman to draft text to send to board+lwg?
11:56 <rweait> +1
11:56 <sly> +0
11:56 <woodpeck> +1
11:57 <pnorman> bulk edits (imports and mechanical edits)
11:57 <woodpeck> please send to lwg first and then to board when we are happy btw.
11:57 <rweait> Easy.  Block them first.  Ask questions later.
11:58 <woodpeck> that is indeed what we do if we see an ongoing, problematic import 
11:58 <sly> +1
11:58 <pnorman> the current state of policy/guideline is messy, to say the least
11:58 <sly> easy, just write that as a new guideline ;-)
11:59 <pnorman> sly: so you're willing to defend a policy of block them first, ask questions later to everyone in france? ;)
12:00 <sly> yes !
12:00 <sly> but don't repeat that to anyone
12:01 <sly> back to serious
12:01 <sly> the problem is tricky
12:01 <rweait> woodpeck, are you logged in as woodpeck, and as sly ?
12:02 <sly> when trying to write in an "open" manner I get queries that the content isn't what x or y wants
12:02 <sly> whish mean the guideline page isn't easy to move forward
12:02 <sly> *which
12:02 <pnorman> the problem with an unwieldy policy is that it doesn't distiguish between reasonable and unreasonable
12:03 <rweait> our open community includes many people and groups who do not wish to see beyond their special interests.  Even if that harms the greater OSM community. 
12:03 <sly> maybe
12:03 <pnorman> I had started on some stuff with my classification of imports stuff but its thankless discouraging work
12:03 <rweait> As representatives of the project as a whole, we must be able to set aside our biases, and act for the good of the project. 
12:04 <sly> or maybe they feel hurt when someone pretend to know better than they do (see, I'm not woodpeck )
12:05 <sly> rweait: nice target ;-)
12:05 <rweait> So if one company is acting to increase their market share as the expense of OSM, we have to act against that.  even if we like the company and their employees.
12:05 <sly> but who pretend to tell appart good and evil ?
12:05 <pnorman> thoughts on how to move forwards?
12:06 <sly> pnorman: yes, move forward no matter what
12:06 <sly> write something
12:06 <sly> discuss a bit
12:06 <sly> and broadcast that it is the new policy we are enforcing
12:06 <sly> then allow time to help us improve it as problems raise
12:07 <pnorman> i mean, in the end, we're going to end up enforcing a policy that not everyone is happy with. no way around that really
12:07 <sly> If we are too much willing to write a law like stone written policy, we'll just end having people playing with it or finding flaws
12:08 <pnorman> I'll try to write some more then?
12:08 <sly> what about the one I wrote ?
12:08 <sly> Is that so far to what you have in mind ?
12:08 <sly> should we keep 2 splited policy ?
12:09 <sly> should we merge imports and automatic edits
12:09 <rweait> The prolem is similar.  Edits without sufficient human attention.
12:09 <sly> should we allow more fuzzy borders to let us handle unexpected cases ?
12:09 <pnorman> I think so - mechanical edits aren't attracing the same problems as imports. with mech. edits its really just people bulk-changing tags based on their view or the wiki without discussing it
12:10 <pnorman> (i think so for keeping it split)
12:10 <woodpeck> i think it is easier for users if we have 2 separate documents for imports and automated edits
12:10 <woodpeck> even if they share 80% of content
12:10 <woodpeck> people can then select the right document for their use case
12:10 <sly> why not
12:10 <rweait> okay
12:10 <sly> we currently have 3 similar wiki pages
12:11 <sly> moving to 2 is still a progress
12:11 <woodpeck> s/automated/mechanical - lack of better wording but if someone makes a bulk xapi+josm change that neets to be covered too
12:11 <pnorman> I think automated and mechanical should be combined
12:11 <woodpeck> currently i say to people that they should look at all three pages because i am not sure if everying is covered everywhere
12:11 <woodpeck> once we have that tidied up we can have everything covered in one place
12:12 <rweait> I'm happy to have the person drafting it refine the format(s)
12:12 <pnorman> okay, so I'm to go away and lock myself in a room until I come up with some kind of draft?
12:13 <rweait> :-)  YOu don't have to lock it.
12:13 <rweait> or you can share portions with the list?
12:13 <sly> I fail to see a difference between automated and mechanical...
12:13 <sly> while we agreed the keyword was "unreviewed"
12:13 <woodpeck> someone from the US community has worked a bit on the import page and he wrote it well and invested some time, but he made it sound like "your easy 10-step guide to a successful import". i think that was losing some of the "so you want to do an import. pause for a moment and think if you REALLY want" component
12:13 <sly> or "uncompared to other sources"
12:14 <pnorman> it's about the language used, I think mechanical is best suited
12:15 <sly> pnorman: about you, drafting that in a locked room, I'm of course against
12:15 <rweait> can we wrap up?  we're over time.
12:15 <sly> As ever, I'd defend a community approach
12:16 <pnorman> drafting text by committee doesn't work well - we need something to present first, which is what we don't have yet
12:16 <rweait> editing to complete the document, then presenting it is a community approach.
12:16 <sly> I feel we should allow them to envolve in the process
12:17 <sly> for the same final content, you have more people understanding what if you allowed them to participate
12:17 <sly> even if you refuse their contributions
12:17 <rweait> And way more bike shedding
12:17 <pnorman> yes, the community needs to be involved, but if we just have it as some vague target nothing will ever get written
12:18 <rweait> The bike shedding is exhausting and counter productive.
12:18 <sly> on the short term, yes
12:18 <woodpeck> sly you have already presented your draft to the list and i don't think you've got much meaningful feedback. instead you have lots of people who say a minority opinion but say it loud, and then you feel you have to do something. i'm against involving the public at a too early stage
12:18 <rweait> What do you call the short term, sly?
12:19 <woodpeck> also the discussion is not structured enough.
12:19 <sly> rweait: I'm struggling to find what "bike shedding" to see if understood you, and then answered what I had in mind
12:19 <woodpeck> if we say: here's our draft, in four weeks we'll decide on the final version, your input please - then that can work better.
12:19 <sly> *means
12:19 <rweait> sly: okay.  another time. 
12:19 <woodpeck> bike sheedding is when people lose view of the big picture and discuss without end about a minor issue
12:20 <sly> ha okay, so my answer is a failure
12:20 <sly> sly: about the draft I wrote, I didn't start from a blank page asking : "please fill" it
12:21 <sly> so, of course, there is a first step to reach before asking for community contributions
12:22 <sly> and a dead line is a good idea to speed people up without them feeling the "gun on their head"
12:22 <sly> with an allready finished document version
12:22 <woodpeck> i'm dropping "my" agenda point of "bully editors", i'll write something about that on the mailing list
12:22 <pnorman> anyways, we're not going to solve this today. 
12:22 <sly> no
12:23 <sly> but I can move on to two draft with you if you want
12:23 <sly> community is also "us"
12:23 <sly> seeing that more than one guy is involved makes people more believeing that democracy is involved
12:25 <sly> maybe for another time, but I find the idea of the "10 point you need to check before importing" a good thing
12:25 <sly> however, priority should be decided
12:25 <sly> mine would be : discuss first, and take your time
12:26 <sly> I'm currently experimenting a mass edit with the french community
12:26 <sly> and it has the positive effect that many people asked me "why don't you just do it ?"
12:27 <sly> where I can answer that "with great power comes great responsability" and mass changes are to be discussed for quite some time before jumping in
12:27 <woodpeck> suggest to end the meeting now. pnorman, are you happy to invite sly to your locked room?
12:27 <pnorman> sure
12:27 <sly> I prefere to let the lock open
12:27 <sly> if you don't mind
12:28 <woodpeck> so you two try up to come up with two concise policy documents that we can present to the community for comment
12:28 -!- _sev has joined #osm-dwg
12:29 <woodpeck> or, for improvement
12:29 <woodpeck> and then we'll see where that goes.
12:29 <rweait> and thanks.
12:29 <sly> I guess that instead of a locked room, at a least a few emails to the dwg list could get feedbacks from dwg members as well
12:30 <pnorman> yes. 
12:30 <woodpeck> sure i'm happy to comment on anything. just not now as i'm eager to go ;)
12:30 <woodpeck> meeting closed?
12:30 <sly> yep
12:30 <woodpeck> thank you all for coming.
12:30 <pnorman> when should we schedule the next one?
12:30 <sly> thanks to all for your work
12:30 <woodpeck> sorry to be in a rush bit i am literally sitting on a construction site ;) 
12:31 <woodpeck> next meeting in ~ 3 months but of course other stuff on the list as before?
12:31 <sly> let's keep going be email
12:31 <sly> *by
12:31 <sly> ++
12:31 -!- woodpeck has left #osm-dwg []
12:35 -!- sly has left #osm-dwg [Quitte]