ORG-tech-vols IRC meeting log 2014 02 17

(19:33:54) graphiclunarkid: Right - let's get started. What have people been up to since we last met?
(19:34:28) graphiclunarkid: I was down in London at the ORG offices this time last week.
(19:34:46) graphiclunarkid: Turns out it was bad timing for this project since they were in the middle of launching Don't Spy On Us!
(19:35:22) graphiclunarkid: I did manage to talk to them about getting a Modx development instance up and running though.
(19:36:05) graphiclunarkid: This we now have running but it needs some minor configuration.
(19:36:06) vasilis: Well this censorship monitoring project started quite long time..
(19:36:42) vasilis: (Mobile networks censorship)
(19:37:45) graphiclunarkid: vasilis: Sorry, not sure what you mean. Are you talking about the name (as I was joking about it earlier?)
(19:38:24) vasilis: Yes
(19:38:55) dantheta: I hadn't realised that the domain had been registered quite that long!
(19:39:00) graphiclunarkid: vasilis: Ah right. Yes, it's been around a while.
(19:39:26) graphiclunarkid: If we're doing a lot of development, though, and some re-branding we can think about new names.
(19:39:53) graphiclunarkid: I believe some people want to find something snappy that catches the attention more.
(19:40:10) graphiclunarkid: I keep asking in case anyone has a sudden stroke of genius!
(19:40:47) graphiclunarkid: dantheta: how goes the middleware?
(19:41:28) dantheta: Pretty good - fixed a couple of bugs and put a vagrant VM up for people to use if they wish (probably of more use to probe developers further down the line)
(19:42:42) dantheta: There's still more spec to fill in for 1.2. I've started to look at the other wiki pages to add more endpoints. NetworkString's original thoughts were to have the API as the complete storage system for frontend data (user-contributions, URL histories) etc.
(19:42:51) dantheta: So we'll need more endpoints and database tables for that.
(19:43:17) graphiclunarkid: Yep. Did you get any feedback on your API spec updates?
(19:45:07) dantheta: I got a nod from Alex. I'm still a little undecided on the aesthetics of the call that takes user report metadata in addition to the URL to check, but it should work at least.
(19:45:27) graphiclunarkid: I had a quick look but I didn't see anything I understood^H thought worthy of comment...
(19:45:33) graphiclunarkid: :)
(19:45:49) dantheta: Korikisulda has also started a java client implementation.
(19:46:12) dantheta: No issues reported there so far.
(19:46:33) graphiclunarkid: dantheta: I had some feedback from someone off-list about the PHP implementation side of things. They had some useful comments about the structure of the API. I asked them to raise a GitHub issue to discuss it.
(19:47:03) graphiclunarkid: dantheta: They said they're really busy and couldn't really get involved but should be able to give a few suggestions.
(19:47:13) dantheta: Cool - that'd be great to see. Was that for a PHP client to talk to the API, or the PHP server code?
(19:47:26) graphiclunarkid: dantheta: It was the API server code.
(19:47:54) dantheta: All suggestions welcome!
(19:48:38) graphiclunarkid: dantheta: I think they wanted to explore a more extensible way of structuring the code - I'll remind them they promised me a GitHub issue on the subject ;-)
(19:49:36) graphiclunarkid: vasilis: Haven't heard anything from plett on the A&A infrastructure side of things this week. I know not having a UK connection to test is holding you up on the OONI side of things.
(19:50:14) vasilis: Well that somehow is a blocker at the moment for making any report.
(19:50:36) graphiclunarkid: vasilis: we could start asking around outside the project to see if we can find someone with (a) a censored line and (b) who is happy to let us use it for testing...
(19:51:12) graphiclunarkid: vasilis: The other thing that would unblock this would be if we win our bid for funding with Lancaster university.
(19:51:13) vasilis: Definetely, I guess there are a bunch of people that censored lines.
(19:51:40) graphiclunarkid: vasilis: We included money for enough raspberry pis, sim cards, and USB dongles to cover the UK mobile networks.
(19:51:44) vasilis: Still we need user reported data.
(19:52:25) graphiclunarkid: We need the probes working first. Then we can ask ORG supporters generally if they'll install the probes themselves.
(19:52:36) dantheta: I can do mobile internet testing (Vodafone), but ADSL is tricky for me. I was under the impression after the previous IRC that it was really landlines that you really needed.
(19:52:42) vasilis: The probes are working.
(19:53:04) graphiclunarkid: vasilis: I thought you had development to do on the probe side of things still - was I wrong about that?
(19:53:07) vasilis: ooniprobe could be installed via a script with ease as well.
(19:53:56) graphiclunarkid: I can do ADSL testing - the problem is my line isn't censored so you wouldn't get any results!
(19:54:11) vasilis: Well there would be some new feautures implemented as well, but the main functionality is there.
(19:54:47) vasilis: dantheta: Mobile internet is perfectly fine as well.
(19:55:14) vasilis: graphiclunarkid: Just try it, if you can spare some bandwidth.
(19:55:24) graphiclunarkid: vasilis: Raspberry Pi based?
(19:55:48) vasilis: graphiclunarkid: You mean the features?
(19:57:02) graphiclunarkid: vasilis: Just wondering if you mean people should try out the code in the OoniOnRasPi repository on GitHub or the stock OONI probe (installed wherever).
(19:57:31) vasilis: I mean the stocked ooniprobe.
(19:57:53) graphiclunarkid: Is it in the raspbian repo or something?
(19:58:00) vasilis: No
(19:58:32) graphiclunarkid: Compile from source?
(19:58:41) vasilis: There are no arm based packages at the moment.
(19:58:51) graphiclunarkid: OK
(19:58:54) vasilis: Installation via pip (python).
(19:59:11) graphiclunarkid: I see, ok.
(19:59:51) graphiclunarkid: Well, I can set up a stock OONI on my connection, which is actually FTTC not ADSL.
(20:00:51) graphiclunarkid: Did you want access to the machine running the probe, or a control interface to the probe itself (if there is one), or just some more probes submitting data?
(20:01:28) vasilis: Both :)
(20:02:02) graphiclunarkid: OK, I'll see what I can do.
(20:02:06) vasilis: Thx
(20:02:31) graphiclunarkid: vasilis: I guess you must be running a probe at home yourself?
(20:05:47) graphiclunarkid: OK, a couple of other quick things:
(20:06:49) vasilis left the room (quit: Ping timeout: 240 seconds).
(20:06:50) graphiclunarkid: 1) Wiki. Thinking about moving all but the general project description pages to github or another mediawiki instance. This is to avoid absorbing huge amounts of the general ORG wiki's namespace (e.g. wiki/android, wiki/stragegic_plan, etc.)
(20:06:54) graphiclunarkid: Any thoughts?
(20:07:18) vasilis [~vasilis@gateway/tor-sasl/vasilis] entered the room.
(20:07:42) graphiclunarkid: Quick repeat for vasilis: Wiki. Thinking about moving all but the general project description pages to github or another mediawiki instance. This is to avoid absorbing huge amounts of the general ORG wiki's namespace (e.g. wiki/android, wiki/stragegic_plan, etc.)
(20:07:56) graphiclunarkid: (badly timed time-out, eh!)
(20:08:09) graphiclunarkid: Any thoughts?
(20:08:26) dantheta: We could do, or we could rename the pages into a namespace of their own, I guess? Happy with either.
(20:08:36) dantheta: I use the category page to get at them, mostly
(20:08:42) graphiclunarkid: dantheta: Me too
(20:09:11) graphiclunarkid: It would be handy to have the detail of e.g. the API spec right next to where it's being developed though.
(20:09:31) graphiclunarkid: And also then people wouldn't need a github account plus an ORG wiki account and keep shifting between the two.
(20:09:40) dantheta: Yes, that's true. I think moving them into their own wiki instance might be a bit heavy - another edit account for people
(20:09:46) graphiclunarkid: (Trying to reduce barriers to entry for new contributors!)
(20:10:05) graphiclunarkid: dantheta: Yeah, that was my feeling too, though ORG's sysadmin Lee seemed to think it would be no hassle to host.
(20:10:24) graphiclunarkid: Not sure how well porting the accounts would work though.
(20:10:34) vasilis: graphiclunarkid: If I recall correctly you have requested an anonymous account for wiki editing?
(20:10:56) dantheta: graphiclunarkid: Yep, that sounds a bit more hasslesome to me.
(20:10:58) graphiclunarkid: vasilis: Yes. It hasn't arrived yet though - as far as I know.
(20:11:05) vasilis: oh
(20:11:35) vasilis: (Replying to your previous question) I do run some probes.
(20:12:56) vasilis: However it seems that I bricked one Pi today :(
(20:13:00) graphiclunarkid: vasilis: OK. Bit unsure what extra you get for development purposes if there are other probes submitting data is all. I guess if you have access to change their config etc. you have a wider base for testing?
(20:13:04) graphiclunarkid: vasilis: D'oh!
(20:13:46) vasilis: graphiclunarkid: Currently this is not 100% implemented for a wide audience.
(20:14:32) graphiclunarkid: vasilis: Sure.
(20:14:48) vasilis: That's why I try to keep things simple and not spend so much time with the images and probes.
(20:15:18) vasilis: Currenlty I'm trying to focus on how I can get more reports.
(20:15:43) vasilis: And data from censored websites so that I can create a technical report.
(20:15:50) graphiclunarkid: vasilis: Well we need it to be easy for people to set up and run a probe before we can ask the general ORG supporter base to consider installing them (which would get many more reports!) Chicken and egg...
(20:16:02) vasilis: which ORG could possibly used it as well.
(20:16:30) vasilis: True that's my concept as well easy to install for everyone.
(20:16:51) dantheta1 [~sailfish@dsl-217-155-42-217.zen.co.uk] entered the room.
(20:17:02) graphiclunarkid: vasilis: Are you talking about summary or other human-readable reports based on analysing the results from the probes to see whether we think the sites were censored?
(20:17:23) graphiclunarkid: Or about the raw reports (results) coming back from the probes themselves?
(20:17:30) vasilis: The ooniprobe reports are human readable.
(20:17:42) vasilis: I miss the data..
(20:17:59) vasilis: In ooni language reports == data :)
(20:18:11) graphiclunarkid: vasilis: Ah right.
(20:19:36) vasilis: I was thinking if there are any VPN services powered by UK DSL/Broadband (preferably filtered) connections.
(20:19:50) graphiclunarkid: So you want a greater volume of data in the database so that... you can draw some conclusions about censorship? Or so that you can improve what data is reported by the probes to make detecting censorship on UK lines possible?
(20:20:19) graphiclunarkid: vasilis: Yeah, VPNs occurred to me too. I thought providing SSH tunnels into coffee-shop wifi hotspots might work...
(20:20:31) graphiclunarkid: ...if they're not blocked that is!
(20:20:34) vasilis: Actually there are almost no UK based data...
(20:21:09) vasilis: or at least not from filtered UK broadband connections.
(20:21:18) graphiclunarkid: vasilis: OK - so we don't know what we're looking for to detect censorship 'cos we don't have any examples of it?
(20:22:46) dantheta: Our mailing list has indicators for BT Broadband, Sky and one or two of the mobile networks - specifically, the cameron filters at least.
(20:22:46) vasilis: Well If you look at an ooniprobe report (data) you could say if any type of censorship occurs.
(20:23:56) vasilis: dantheta: True but if you run an ooniprobe "scan" for 20k websites you would a clear picture of what is censored or not..
(20:24:20) graphiclunarkid: vasilis: So you just want more data so that you can say what websites are censored on what connections? Not for development?
(20:24:34) vasilis: Repeating the same test on a different broadback line and then you can do some correlation..
(20:24:50) vasilis: No not for development.
(20:24:59) graphiclunarkid: OK
(20:25:05) vasilis: I thought we had talked about it before..
(20:25:32) graphiclunarkid: vasilis: Not sure I quite understood is all! Just making sure I do now ;)
(20:25:35) vasilis: I'm pretty sure that a nice looking report would bring more people to the project as well.
(20:26:01) vasilis: ..and give a boost at ORG blocked project
(20:26:12) graphiclunarkid: Yeah, it would be good to have some early results to show off :)
(20:26:25) vasilis: Short of ;)
(20:26:36) vasilis: Adding some graphs maybe..
(20:27:54) dantheta: Really sorry, i'm going to have to head off in a couple of minutes ...
(20:28:23) graphiclunarkid: dantheta No worries - thanks for joining us and for your work on the API side of things :)
(20:28:45) dantheta: graphiclunarkid: All good! I'll mail you about commit access on the ORG github acct, if that's OK.
(20:29:05) graphiclunarkid: dantheta Yep, no problem.
(20:29:21) graphiclunarkid: Or just send a pull request :)
(20:29:21) vasilis: dantheta: Could you run a probe?
(20:30:19) dantheta: I'll definitely take a look at it. I picked up a vodaphone mobile connection for having a filtered line to test with, I should be able to connect the two together!
(20:30:42) vasilis: dantheta: Nice, you can always contact me on/off list.
(20:31:10) dantheta: Cheers! Sorry, gotta run! See you all next time.
(20:31:24) dantheta left the room (quit: Quit: dantheta).
(20:31:52) graphiclunarkid: vasilis: If you're hanging around for a bit I'll see if I can find a spare SD card and get an OONI instance running here...
(20:32:11) vasilis: Yep I 'll be around.
(20:32:38) graphiclunarkid: Cool. I'll give you a yell when it's up (or if I need a hand!)
(20:32:39) vasilis: If you an SD card I could even provide you with an image now.
(20:32:45) graphiclunarkid: Yes please - that'd be handy!
(20:32:49) vasilis: OK
(20:32:52) graphiclunarkid: How big?
(20:32:57) vasilis: 4G
(20:33:14) graphiclunarkid: For the RPi?
(20:33:31) graphiclunarkid: OK - found a spare 4G card.
(20:33:39) ***graphiclunarkid goes to find a spare ethernet cable...
(20:33:57) graphiclunarkid: Does the image have ssh and DHCP enabled by default?
(20:34:05) vasilis: Yes
(20:34:10) graphiclunarkid: Excellent.
(20:34:22) ***graphiclunarkid only has a laptop and doesn't have a monitor for the pi!
(20:36:00) vasilis: Nice
(20:37:30) vasilis: bbl in ~15min
(20:38:37) graphiclunarkid: OK. Thanks for the link. Getting it now.