Perl 6 - the future is here, just unevenly distributed

IRC log for #opentreeoflife, 2014-03-17

| Channels | #opentreeoflife index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:07 mtholder joined #opentreeoflife
00:15 josephwb joined #opentreeoflife
00:16 josephwb joined #opentreeoflife
00:52 josephwb joined #opentreeoflife
00:58 josephwb1 joined #opentreeoflife
01:13 towodo joined #opentreeoflife
05:57 jimallman joined #opentreeoflife
10:53 josephwb joined #opentreeoflife
13:32 josephwb joined #opentreeoflife
14:15 towodo joined #opentreeoflife
14:37 jimallman joined #opentreeoflife
15:02 kcranstn joined #opentreeoflife
15:55 travis-ci joined #opentreeoflife
15:55 travis-ci [travis-ci] OpenTreeOfLife/api.opentreeoflife.org#373 (sha-merge - 6601068 : Emily Jane McTavish): The build passed.
15:55 travis-ci [travis-ci] Change view : https://github.com/OpenTreeOfLife/api.opentreeoflife.org/compare/20375ef3ed69^...660106803bd2
15:55 travis-ci [travis-ci] Build details : http://travis-ci.org/OpenTreeOfLife/api.opentreeoflife.org/builds/20947236
15:55 travis-ci left #opentreeoflife
15:59 snacktavish joined #opentreeoflife
16:41 travis-ci joined #opentreeoflife
16:41 travis-ci [travis-ci] OpenTreeOfLife/api.opentreeoflife.org#374 (sha-merge - c2d22ad : Emily Jane McTavish): The build passed.
16:41 travis-ci [travis-ci] Change view : https://github.com/OpenTreeOfLife/api.opentreeoflife.org/compare/660106803bd2...c2d22ad8216a
16:41 travis-ci [travis-ci] Build details : http://travis-ci.org/OpenTreeOfLife/api.opentreeoflife.org/builds/20950824
16:41 travis-ci left #opentreeoflife
16:57 josephwb1 joined #opentreeoflife
16:57 snacktavish jimallman: I finally merged in the sha-bbbooookkeeeeppiinngg branch with some other minor changes and re-deployed to ot7. Should return parent sha's on GET now.
16:58 jimallman cool! thanks!
16:59 jimallman the curation app is already expecting this and will re-submit it (as a query-string argument) when saving changes.
17:01 towodo kcranstn, jimallman, et al: Google is failing us
17:02 towodo can't even see my hangouts list
17:02 snacktavish IRC hangout!
17:02 kcranstn ah, yes
17:02 towodo 'Things are taking longer than expected.'
17:02 kcranstn getting the same thing here
17:02 kcranstn I am happy to IRC hangout
17:02 jimallman same here
17:02 towodo we could meet by IRC if everyone were on it
17:03 mtholder joined #opentreeoflife
17:03 towodo but we only have 5 out of 10
17:03 towodo 6 out of 10
17:03 kcranstn option B is skype
17:04 jimallman towodo: can we skip the Plus sidebar and work directly from here?
17:04 jimallman https://plus.google.com/hangouts/active
17:04 towodo when i had the sidebar earlier i was getting failures as well
17:04 mtholder that looks promising
17:04 jimallman hm, nope. i tried to start a video hangout and it fails.
17:04 jimallman "There is a problem connecting to this video call. Try again in a few minutes.
17:05 kcranstn sending mail to software list re:IRC meeting
17:05 jimallman http://www.google.com/appsstatus#hl=en&v=status
17:05 towodo 'There is a problem connecting to this video call. Try again in a few minutes.'
17:05 jimallman status page (above) says G+ Hangouts is "condition yellow"
17:06 towodo I'm not getting the status page at all.
17:06 jimallman !
17:06 towodo Pinged stephen on skype, no response
17:07 towodo i'm not sure i have everyone's skype login… let me see
17:07 mtholder must be all the wild st. patricks g+ hangouts. That is a thing, right?
17:08 kcranstn of course it is
17:08 pmidford joined #opentreeoflife
17:08 kcranstn agenda item 1: implementing user feedback
17:08 towodo i don't have jim allman in skype...
17:09 kcranstn I already sent a message about using irc instead of hangouts
17:09 towodo there are only about 5000 jim allmans in skype...
17:09 towodo ok let's get started i guess.
17:09 jimallman towodo: FYI - my Skype name is jimallman_ibang.com
17:10 towodo https://docs.google.com/document/d/1F6O-RG-anyP_7_U083mJ500LBWq8TrteU89QswwUsho/edit#
17:10 towodo additions to agenda?
17:10 rree joined #opentreeoflife
17:10 lcoghill joined #opentreeoflife
17:10 kcranstn summary of current feedback situation: fielding many emails with feedback re: synthetic tree + taxonomy
17:10 kcranstn not sustainable
17:10 mtholder nor transparent
17:11 kcranstn have comment system on tree browser, but requires separate login from curation interface (= confusing)
17:11 blackrim joined #opentreeoflife
17:11 kcranstn bracket equals confusing)
17:11 jimallman yes, it's Janrain (federated social-web login)
17:11 blackrim whois kcranstn
17:11 kcranstn karen
17:12 blackrim ha, sorry, trying to see your client so i could tell cody --missed the /
17:12 kcranstn adium
17:12 towodo yes, two logins is an annoyance.  we had talked about reducing it to one, namely github
17:12 kcranstn I think having a single login for the webapp is a good thing
17:12 jimallman N.B. I could change the auth for commenting system to GitHub, but that seemed like an unneccesary barrier to participation.
17:13 towodo does commenting need to have an authentication barrier? what is experience of sites like ours with spam?
17:13 codiferous1 joined #opentreeoflife
17:13 kcranstn we get plenty on the blog
17:13 codiferous1 old school
17:14 towodo but blogs are standard targets, not one-offs
17:14 jimallman one-off?
17:14 kcranstn ok, so one question is "do you need to log in to leave a comment?"
17:14 towodo i mean, our site is bespoke, it doesn't look like others, while the blog site does look like many others
17:15 jimallman i will say from experience: w/o auth, some people will neglect to identify themselves (confusion, not malice)
17:15 towodo so blog spam is easy to automate, there is an economy of scale, while our site would need special attention
17:15 jimallman towodo: i see, so there's more room for confusion
17:15 jimallman too
17:15 kcranstn I can see having a feedback form that doesn't require authentication
17:15 towodo no, i'm saying that special sites might get less spam than standard sites like blogs
17:16 towodo put an 'identify yourself' field in the feedback form?
17:16 jimallman we can validate on this (non-empty, min 5 chars..). that takes care of user confusion
17:17 jimallman maybe stash prior input from this session, so it's not such a burden
17:17 towodo karen, is moderation of the blog an overwhelming burden? or manageable?
17:17 kcranstn its manageable, but annoying
17:18 towodo i guess it's like the 'edit' link in wikipedia...
17:18 towodo no auth there
17:18 towodo or should i say optional auth.
17:18 kcranstn can we agree that we don't want multiple authentication systems for different parts of the site?
17:19 kcranstn i.e. curation and feedback
17:19 towodo I can agree...
17:19 jimallman i think it reflects two different populations... but i could be wrong.
17:19 mtholder I think that it would not be terrible to have 2 if the curator pop is much smaller.
17:20 jimallman GitHub = active contributors/curators... vs. visitors and drive-by experts who might have useful comments
17:20 towodo the curator population will want to comment on the tree, so that's 2 populations commenting on the tree
17:21 jimallman not a problem, right? but they're the ones who might be annoyed by two auth systems
17:21 towodo so even if there are 2 populations, they're not collated by web2py application.  so yes, curators would be annoyed
17:22 kcranstn we also need to monitor / respond to feedback. Having github issues puts these in our regular workspace
17:22 lcoghill joined #opentreeoflife
17:22 towodo I haven't heard any objections to using github issues for tree and site feedback.  Can anyone think of an argument against?
17:22 jimallman +1 for moving storage and moderation to GitHub trackers (so many reasons). i assume it's a plus to have this happen in a more visible space?
17:23 towodo other than the change will make work for jim?
17:23 kcranstn and makes it simple to separate conversation about Issue #1 for node A from Issue #2 for the same node
17:23 jimallman i can keep the same UI (a la HerpNet) and simply re-route the data
17:24 snacktavish joined #opentreeoflife
17:24 kcranstn do we also want to have a general feedback form?
17:24 jimallman seems wise to sprinkle lots of "suggestion boxes" around, yes
17:24 jimallman Contact Us, Feedback, and in-situ commenting tools
17:25 jimallman (all feeding GH issue tracker, with different labels / milestones / embedded data)
17:25 towodo with a topic-selection radio - I think you've already done this.  maybe this could be used to route comments to different git repos.
17:26 kcranstn can you move an issue between repos on github?
17:26 jimallman that's my thinking, yes. the filters and location-specific metadata are already in the UI
17:27 jimallman kcranstn: doh! i'm pretty sure you can't.
17:27 jimallman duke and i had to move some issues between API and phylesystem, and it was manual copy-paste. ugh.
17:28 towodo i'm not aware of any way to move issues… i've been doing it manually.  there are some articles on the web about automating the motion of issues wholesale, but i never found a script for moving a single issue.
17:28 jimallman HerpNet just has dedicated repos for its issue trackers (several, if i recall)
17:28 towodo i.e. you can move them, it's a mere matter of programming
17:28 jimallman we could probably automate this somewhat using the GitHub API for issues.
17:29 kcranstn thinking about the user who comments on a relationship in the tree, but we won't know until later whether the problem is input tree vs synthesis vs ott
17:29 * jimallman means VertNet, not HerpNet
17:30 towodo Yes, I was thinking we could have a separate repo for otherwise unclassified comments. Small issues could just stay there, you wouldn't necessarily have to move an issue at all.
17:30 jimallman the current UI lets the user "qualify" their comment (best guess), but stores metadata for all contexts (neo4j node, ott id, source-tree id)
17:32 jimallman we could route comments based on this, if there's just once choice to make (by source tree, or taxonomy...)
17:33 towodo (logging in to opentree comments now to look… had to look up my yahoo user id)
17:33 kcranstn we probably want to have a version 1.0 of the feedback system that simply replicates the vertnet functionality (click submit -> log into github -> create issue) before getting fancy
17:33 jimallman fyi - here are some (all?) of the issue-tracking repos for VertNet: https://github.com/rom-vertnet
17:34 jimallman kcranstn: i think we could do a lot in one repo, with clever assignment of labels and milestones
17:35 towodo yes.  so part of this plan is to set up a new catchall repo?
17:35 towodo first issue at vertnet i looked at is 5 months old and unanswered...
17:36 kcranstn jimallman - that link is only for issues with data from the ROM (royal ontario museum)
17:36 jimallman yes, it's all reeeaally quiet in there.
17:36 jimallman kcranstn: ah, that makes sense. they use multiple orgs to facilitate third-party moderation.
17:36 kcranstn they are also trying to push issues to data providers
17:37 jimallman yep
17:37 towodo (do they have an authentication barrier?...)
17:37 kcranstn you need a github account
17:37 mtholder isn't that what filteredpush is supposed to do? Do the FP folks have a solution?
17:37 kcranstn good question
17:38 kcranstn vertnet also just has a generic feedback form (no auth required): http://www.vertnet.org/feedback/contact.html
17:38 mtholder I think the answer is "no, they don't" though
17:38 kcranstn or "yes, they do, but it is complicated"
17:38 towodo SILVA just has an email address, FWIW
17:40 kcranstn does it makes sense to have a separate repo for feedback, versus using the webapp repo?
17:40 towodo that's what I was proposing… a new catch-all repo to be used by feedback submission form
17:41 jimallman i think webapp repo is probably best used for bugs/features on the webapp itself
17:41 kcranstn ok
17:42 jimallman but we might route comments tagged as "bug report" to its issues.
17:42 towodo "bug report" could mean anything - tnrs, treemachine, api, ...
17:43 jimallman the UI offers the option "Bug report (website behavior)"... though often a user can't know for sure which they're seeing.
17:44 kcranstn do we have a "submit issue" form that requires github login and a "general feedback" form that doesn't?
17:44 jimallman i'm pretty sure we can do this, using the API and our "robot GitHub user" to submit issues.
17:44 jimallman anonymous issues, i mean
17:45 towodo not sure how login some places and not others would help...
17:45 towodo what if we just made it open, like wikipedia, but be ready to put up barriers if we have spam problems?
17:46 mtholder the other route is to have a opentree user profile. If you have your github info there, then you can have your info in non-anonymous issues.
17:46 jimallman true, we could still detect an authenticated GH user (logged in) and paste in their contact info.
17:47 jimallman this is effectively what we do now. any web2py query on the user is passed through GitHub API, so we'll know if they're logged in.
17:47 towodo I like the idea of dispensing with the persistent opentree user database…
17:48 jimallman if we do run into spamming problems, we can simply add the login requirement to the feedback form (or CAPTCHA, or whatever).
17:48 towodo I'm wondering if we should use the last 12 minutes to talk about the second agendum?
17:48 kcranstn sure
17:48 towodo regarding feedback form, i think we've just put this on Jim's long to-do list...
17:49 towodo I have a spiel about indexing.
17:49 kcranstn go!
17:49 towodo I am assuming that OTI will serve our purposes for launch / paper.  Maybe I am wrong, correct me.
17:49 towodo So work on indexing can be put off a bit. I'm assuming I'll be the one to do it, but someone else can take it on if they like.
17:50 towodo I anticipate a design phase where technology platform options are considered
17:50 towodo We could evaluate four or five options, maybe do some experiments, and pick the best.
17:51 kcranstn sounds good, as long as your first assumption is correct
17:51 towodo We have the use cases for indexing written down. I don't see any of these as must-haves for launch.
17:51 jimallman i think OTI is working well for most purposes. my immediate concern is "freshness" of the index as studies change.
17:52 towodo If I understood Cody correctly, there is a service that will replace the current version of a study with a newer version. Yes Cody?
17:52 towodo Even if not, we could refresh OTI nightly - it doesn't take that long.
17:53 codiferous1 yes
17:53 codiferous1 the indexSingleNexson service will replace any existing index entries for the study in the nexson
17:54 towodo So, version 1, a cron job that refreshes OTI periodically. Version 2, some contraption that updates OTI as studies are updated (git commit hook maybe, or a poke from the curation tool)
17:54 towodo Does that work for you Jim? Or worried that high latency will be a turn-off to curators?
17:54 jimallman re: indexSingleNexson, i can start triggering this form the curation tool, but i think it would be best to set up hooks on GitHub.
17:55 jimallman yes, i'm concerned that some things like the study list will be confusing if they don't respond quickly.
17:55 kcranstn agreed
17:55 jimallman but if indexSingleNexson is a pig, we could target just specific events (eb, study creation and deletion) from the curation app
17:55 towodo You could always maintain a local version of the study list, and update it with changes made by that client
17:56 jimallman eww
17:56 towodo ok.
17:56 jimallman :D
17:56 towodo I doubt indexSingleNexson is a pig.
17:56 codiferous1 it shouldn't be
17:56 jimallman agreed, i'd love to try the git hook solution to keep it in sync
17:56 kcranstn how much of that work is on the webapp side vs oti?
17:57 jimallman git hook would be entirely in GitHub, perhaps with helper logic in the Open Tree API.
17:57 mtholder joined #opentreeoflife
17:58 snacktavish That makes sense.
17:58 jimallman but i think it's a simple script triggered by commits to master, and pings the indexSingleNexson service with one or more study ids
17:58 kcranstn nice
17:58 pmidford yes
17:58 towodo I think this is manageable.
17:59 towodo karen, is the question about indexing answered to your satisfaction?
17:59 kcranstn yup
17:59 towodo am i waving my hands too vigorously?
17:59 kcranstn no
17:59 kcranstn ok, gotta run convince some folks that putting CC-BY-NC-SA on a database is a bad idea
18:00 towodo good luck.
18:00 towodo ok, anything else?
18:01 jimallman i'm happy.
18:01 blackrim not from me
18:01 towodo ok, I think we're adjourning.  Think about agenda for Friday.
18:01 mtholder ok
18:01 pmidford ok
18:01 blackrim see you later
18:02 blackrim left #opentreeoflife
18:03 lcoghill exit
18:04 josephwb joined #opentreeoflife
18:11 josephwb joined #opentreeoflife
18:26 pmidford joined #opentreeoflife
18:37 josephwb joined #opentreeoflife
18:47 kcranstn joined #opentreeoflife
20:07 josephwb joined #opentreeoflife
20:57 mtholder left #opentreeoflife
20:57 mtholder joined #opentreeoflife
21:03 mtholder left #opentreeoflife
21:04 mtholder joined #opentreeoflife
23:06 josephwb joined #opentreeoflife
23:37 jimallman joined #opentreeoflife
23:45 kcranstn joined #opentreeoflife

| Channels | #opentreeoflife index | Today | | Search | Google Search | Plain-Text | summary