15:59:50 <hellais> #startmeeting OONI Commmunity meeting 2017-02-28
15:59:50 <MeetBot> Meeting started Tue Feb 28 15:59:50 2017 UTC.  The chair is hellais. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:59:50 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:59:59 <hellais> agenda pad: https://pad.riseup.net/p/ooni-community-meeting
16:00:11 <slacktopus> <hellais> Hello this is Arturo from OONI. Who is here?
16:00:18 <slacktopus> Action: sbs is here
16:00:23 <slacktopus> <agrabeli> Hello! This is Maria from OONI.
16:00:46 <slacktopus> <dbyko> Hi! This is Dan Bateyko
16:02:01 * landers is here
16:02:02 <slacktopus> <sukarn> Hello, this is Sukarn from SFLC.in (Software Freedom Law Centre, India)
16:02:04 <slacktopus> Action: darkk is also here, affiliated with OONI as well
16:02:09 <slacktopus> <sukarn> I am a Counsel
16:02:34 <slacktopus> <tafiti> Moses here from CIPIT
16:02:50 <slacktopus> <agrabeli> Great, thank you all for joining us! Today we're having a more community-oriented meeting and the agenda can be found through the following pad: https://pad.riseup.net/p/ooni-community-meeting
16:04:27 <slacktopus> <agrabeli> Generally the OONI team has weekly meetings on IRC that anyone from the public can join. But last week we decided to switch these meethings to slack and to invite broader members of the community to join us so that we can hear your feedback and thoughts.
16:04:47 <slacktopus> <agrabeli> The first topic of the agenda: 1. How to suggest URLs to be tested for censorship, without having to use Github (i.e. how to suggest URLs for testing directly through a web interface).
16:05:12 <slacktopus> <alin> Hi all
16:05:45 <slacktopus> <alin> I have a general question regarding the URL collection (scrapping or whatever we call it)
16:06:00 <slacktopus> <agrabeli> Would the individual who added the topic like to share a few words so we can kick off the first topic? :slightly_smiling_face:
16:06:11 <slacktopus> <tafiti> I had the thought of broadcasting a form that captures the url, category and description - just as they are on /citizenlab repo
16:06:24 <slacktopus> <alin> 1- is it page based or domain based?
16:07:01 <slacktopus> <alin> if so then you are in a trouble in countries like Iran. It is the current problem of OONI with its Iran's report
16:07:55 <slacktopus> <hellais> So on this I wanted to ask a clarifying question to who put it on the agenda (it’s also perhaps useful to put a name next to each topic in the pad): By 1. we mean to submit URLs to the global URL testing corpus and not that of instrumenting ooniprobe to test a specific URL, correct?
16:08:37 <slacktopus> <alin> 2- How should we find the popular sites in a country including the blocked ones? Alexa won't work, obviously.
16:08:43 <slacktopus> <hellais> @alin I am not sure I understand your question. Are you talking about how we detect that censorship is happening?
16:10:00 <slacktopus> <tafiti> I proposed the agenda on that. The idea is to have a form on the backend that collects suggested urls from people out there. Then after cleaning out possible errors or inconsistencies, they can be added to the global list or country list (if we decide to add a form field asking for which country it is being suggested from
16:10:03 <slacktopus> <hellais> If that is the question we use a variety of different techniques: 1. We check to see if the A record lookup for the domain in question is consistent with a lookup from a control vantage point 2. We check to see if the HTTP response we got is consistent with control vantage point 3. We check to see if TCP connections to the IPs the domain resolves to succeed or if they fail if they are also failing from the control
16:10:34 <slacktopus> <hellais> These 3 are all done in the client and provide an immediate indication of possible censorship to the user upon running the test, but can sometimes lead to false positives.
16:10:49 <slacktopus> <alin> @hellais no, I am talking about the strategy of OONI in countries like iran. In Iran, if google,com is blocked, all subdomains and all pages are blocked as well. in the OONI's report about Iran you see may pages and subdomains
16:12:16 <slacktopus> <alin> I have one more question as well. How does OONI categorize the sites into different categories?
16:13:25 <slacktopus> <agrabeli> @alin the URLs are categorised based on these category codes: https://github.com/citizenlab/test-lists/blob/master/lists/00-LEGEND-new_category_codes.csv
16:13:54 <slacktopus> <sweemeng> are we discussing on 1) how to suggest a URL?
16:14:07 <slacktopus> <hellais> @alin yes our current approach is that of considering as a unit a specific URL (i.e. we don’t group different URLs based on common base domain name). We are considering potentially enriching this grouping in the future, by being able to present a view that is more aggregate by grouping together, for example, all google related services or all pages on wikipedia, but that is not something we are currently doing. Is that what you are asking?
16:14:28 <slacktopus> <agrabeli> @alin: These categories were developed by the Citizen Lab in its attempt to provide a resource for network measurement researchers to examine a broad range of categories. These categories include mainstream types of sites, but also others that are more provocative or objectionable. You can find more information about this here: https://ooni.torproject.org/get-involved/contribute-test-lists/
16:14:58 <slacktopus> <agrabeli> @sweemeng: yes we started to discuss 1... but 1 is quite related to @alin's question...
16:15:05 <slacktopus> <alin> @agrabeli I have read it but how do you decided to put a link in each category? manually or by machine learning or by peer checking with other link directories or ...?
16:15:06 <slacktopus> <sweemeng> ok thanks
16:15:29 <slacktopus> <sweemeng> alin decided by a person.
16:15:53 <slacktopus> <agrabeli> so @tafiti suggested creating a sort of web interface that allows individuals to submit URLs for testing without having to use Github
16:16:24 <slacktopus> <agrabeli> I think that's a great idea and we definitely have in our long-term plans to create a sort of URL management system which allows individuals to do precisely that
16:16:41 <slacktopus> <tafiti> Something like this: https://docs.google.com/forms/d/e/1FAIpQLScHG_K3KXaR84wsBkVed_3zdYgJG8T2whkCEcEDcooRVX0_GA/viewform
16:16:52 <slacktopus> <sweemeng> actually, that is one of the idea i want to pursue. For a good reason, because people keep asking us "Are example.com blocked?"
16:17:03 <slacktopus> <agrabeli> in the shorter term, we aim to add support to ooniprobe mobile and OONI's web UI that will allow users to test single URLs of their choice, though this won't necessarily be synched directly with the citizen lab repo
16:17:06 <slacktopus> <sweemeng> it give an idea what people want to know and actually being use
16:17:30 <slacktopus> <darkk> @alin AFAIK, current categories markup is manual, no ML there for the moment
16:17:46 <slacktopus> <sweemeng> so our (aka sinar project original block or not site) is popular because people are asking the same question
16:19:15 <slacktopus> <sweemeng> the catch is, what people use might not reflect censorship level. Say people not interested in political site and it get blocked, we don't really know about it
16:19:25 <slacktopus> <agrabeli> @alin: That's a great question! We generally categorize URLs manually....we don't have any machine learning tech for that. Of course, this is not perfect and certainly prone to error. It's also often quite difficult to accurately categorize URLs in all cases. As such, to my understanding, the Citizen Lab developed a new set of categories that would allow more types of URLs to fit under one category, and thus potentially prone to less error. B
16:19:26 <slacktopus> is an iterative process, and we're definitely open to suggestions. We'll actually be hosting a workshop about this at the IFF that you might be interested in attending.
16:19:32 <slacktopus> <alin> We are working on Iran's censorship specifically. Our idea came from OONI and we'd like to do it with OONI but we received no reaction from you guys :slightly_smiling_face: maybe this conversation helps both and exchange the experiences we have. you can find more about our project here: biband.com
16:20:49 <slacktopus> <alin> I think, I talked a lot. I will be quite for the some mins.
16:20:57 <slacktopus> <tafiti> :slightly_smiling_face:
16:21:45 <slacktopus> <hellais> Yes so on the topic of adding URLs for testing an important constraint to consider is that we don’t want unvetted URLs to end up being automatically tested, but instead we want it to go through the normal review process that involves submitting a PR to the citizenlab/test-lists repo and having somebody approve it. The reason for this is that we don’t want to be testing from probes: a) Sites that are more likely to get them into trouble b)
16:21:45 <slacktopus> very relevant or useful to be tested c) Sites that can lead to users being identified (for example: http://evil.showmyip.com)
16:21:46 <slacktopus> <agrabeli> @alin: Awesome thanks for sharing! I admittedly didn't know about the project... I'll have a look at it and I'd love to explore ways of collaborating (we can also continue discussions on private channels later on).
16:22:16 <slacktopus> <sweemeng> hellais any idea on the criteria?
16:23:37 <slacktopus> <sweemeng> for example, do we want to test pornographic site? why?
16:23:39 <slacktopus> <hellais> @alin that looks like a really cool project (I also was unware of it), we should for sure discuss more how we can help.
16:23:56 <slacktopus> <agrabeli> @sweemeng: Some of the general criteria that we follow include:
16:24:08 <slacktopus> <agrabeli> 1. Never include child pornography or sites that contain malware
16:24:40 <slacktopus> <agrabeli> 2. Add sites that are more likely to be blocked (such as pornography), because that will allow us to develop heuristics on identifying censorship in a network
16:24:42 <slacktopus> <tafiti> @agrabeli the idea is to make it easy for both parties to enrich the test list (for the contributers, it will be a simple form taking like a minute to fill, for OONI, the structured data at the backend will be ready to paste on /CitizenLab
16:24:53 <slacktopus> <alin> I am going to present our finding at IFF on 9th. BTW, the project has been launched some weeks ago even though we have been working on it for a year
16:25:10 <slacktopus> <sweemeng> agrabeli alright, got it
16:25:15 <slacktopus> <darkk> I'd also add that we had idea to mine URLs from actual measurements to be able to find new sites people *care* about.
16:26:16 <slacktopus> <sukarn> @hellais Thanks. That actually answers the fifth point that I added to the agenda
16:26:26 <slacktopus> <alin> The we can share our ideas. we have started it a bit. Right now we are working on the data analysis and visualization with a great guys
16:26:49 <slacktopus> <sukarn> @agrabeli you too.
16:27:05 <slacktopus> <agrabeli> @sweemeng: We generally try to add as many different URLs as possible so as to reflect diversity in the research and to limit bias. That said, websites which provocative material can be useful as part of the testing because they can help us identify how censorship is being implemented, but it's also important to include a variety of human rights websites (because if they are found to be blocked, that tells a more compelling story).
16:27:20 <slacktopus> <sweemeng> right.
16:27:43 <slacktopus> <agrabeli> @alin: Awesome! I'll take a note to attend at IFF :slightly_smiling_face:
16:27:59 <slacktopus> <sweemeng> one reason to have a user submitted site is do that we don't have to mind URL from measurements darkk
16:28:07 <slacktopus> <sweemeng> that's the theory anyway
16:29:21 <slacktopus> <sweemeng> anyway if there is a database of url collected how to release it? add it to citizenlab list?
16:29:56 <slacktopus> <agrabeli> @sweemeng yes please. You can find a guide on how to do so here: https://ooni.torproject.org/get-involved/contribute-test-lists/
16:30:14 <slacktopus> <dbyko> @agrabeli is there an easy way for OONI-users to check URLs of neighboring countries in addition to their own country's list? Are there regional lists?
16:30:17 <slacktopus> <sweemeng> i mean for the web frontend thing for agenda 1 agrabeli ?
16:30:25 <slacktopus> <alin> I am sorry that I mess again. There is a website: blockediniran.com which is going to test sites in Iran if they are blocked or not. We got the idea form them to let users check the sites and it automatically adds the site to our list. By that, I meant it is a great idea to see what users like to check
16:30:32 <slacktopus> <agrabeli> @sweemeng: It's important that URLs are added to the Citizen Lab repo because that allows many network measurement researchers (beyond OONI) to test those URLs.
16:31:29 <slacktopus> <agrabeli> @dbyko Yep! All of the country lists can be found here: https://github.com/citizenlab/test-lists/tree/master/lists
16:32:28 <slacktopus> <agrabeli> @dbyko You can find all of the country lists that we have available through there. They're included as csv files with the country code of each country. If you're interested in adding a list for a country that is missing, you can find instructions here: https://ooni.torproject.org/get-involved/contribute-test-lists/
16:32:56 <slacktopus> <dbyko> @agrabeli thanks!
16:33:15 <slacktopus> <agrabeli> @dbyko If you're a github user, you can make a pull request with the new csv for the country you're interested in (following the data format of the other csvs). If you're not a github user, you can share your spreadsheet with the OONI team at contact@openobservatory.org
16:33:27 <slacktopus> <hellais> @sweemeng no there is currently no such web interface. I started writing a simple proof of concept of how it could work here: https://github.com/hellais/tazo, but it still needs a bit of work.
16:34:08 <slacktopus> <sweemeng> hellais turns out that i have the same idea https://github.com/Sinar/blockornot2
16:34:10 <slacktopus> <hellais> The basic idea of the web service was that of automating the process of committing to the relevant country test list and opening a pull request on github with the added URLs
16:34:12 <slacktopus> <sweemeng> except I forget ot commit
16:34:32 <slacktopus> <agrabeli> @sweemeng Awesome! :slightly_smiling_face:
16:35:20 <slacktopus> <sweemeng> the big idea is, 1) get user to query what they want to know. then count it for priority
16:35:27 <slacktopus> <sweemeng> 2) get user to submit a url, which we filter
16:36:04 <slacktopus> <sweemeng> i kind of busy that I didn't finish the poc
16:36:22 <slacktopus> <sweemeng> 1) is let user find out whether a site is block
16:36:35 <slacktopus> <sweemeng> then see how many interest to add to list etc
16:36:48 <slacktopus> <darkk> @alin we're a bit afraid of automatic test of submitted URLs as ooniprobes doing actual tests are run by volunteers and  we assume that it may be illegal to _visit_ specific URLs in some countries, that's another reason for OONI to avoid runtime checks for user-submitted URLs unless the user tests the URL himself. That's the good question if that risk is real or not.
16:38:02 <slacktopus> <sweemeng> it is true for thailand darkk and alin
16:38:04 <slacktopus> <sweemeng> fyi
16:38:09 <slacktopus> <dbyko> agrabeli: I see potential for x country to block neighboring countries' websites/news media and also for those websites to be missing from x country's list, so regional lists might also be helpful
16:38:48 <slacktopus> <alin> I was not aware of it. Thanks @sweemeng and @darkk. We work on Iran and there we don't have such as this problems.
16:38:56 <slacktopus> <agrabeli> Yep you're right! We've seen that in certain African countries where the website of a Ugandan NGO was blocked in a neighbouring country.
16:39:14 <slacktopus> <hellais> in addition to what @darkk said :ooni: tries to avoid collecting identifying information about users. If an attacker interested in discovering all ooniprobes they could/would instruct all probes to check http://attacker.example.com/ and learn all the IPs of ooniprobes.
16:39:22 <slacktopus> <sukarn> This is not legal advice: According to Indian law, even 'seeking' child porn is punishable, so a website that contains links to child porn could also get someone in trouble.
16:39:31 <slacktopus> <dbyko> ahhhh
16:40:05 <slacktopus> <agrabeli> I think it could definitely be a good idea to create regional lists as well. My only concern is the user's bandwitdth - in the sense that they would then be testing the global list, the country list, and the regional list....all of which would be a lot of URLs! :P
16:40:12 <slacktopus> <sweemeng> so is there a conclusion for 1) yet?
16:40:15 <slacktopus> <darkk> @sweemeng can you share more context on TH? is the rule actually ever applied? is it widely enforced?
16:41:19 <slacktopus> <agrabeli> I should mention though that what we're writing here is appearing publicly in the ooni IRC channel.
16:41:22 <slacktopus> <sweemeng> I learn through one of the contributor for the thailand citizenlab list, asked that we commit and update the list. It is through him that I learn about the law
16:41:38 <slacktopus> <sweemeng> why knowing about a link is enough for persecution
16:42:15 <slacktopus> <sweemeng> he shall be anonymous
16:42:57 <slacktopus> <agrabeli> @sukarn Indeed, which is also why we never include such sites in the testing.
16:43:21 <slacktopus> <hellais> I think the conclusion is that having something that would allow non github people to suggest URLs for inclusion into the test-lists repo would be a great idea, however from our end it seems like we will not have much time to work on it in the immediate future. If somebody were to develop it though I think that would be great!
16:43:42 <slacktopus> <sweemeng> i try to find time to finish my little poc then
16:43:45 <slacktopus> <sweemeng> POC
16:43:51 <slacktopus> <darkk> disclaimer: I don't suggest to break the laws, but it's always interesting if some rule is just a non-enforcable paper-rule, something that really matter to people at risks (selectively enforced for "troubled" people like activists) or something that matters to general public (either widely-enforced or randomly-enforced). :slightly_smiling_face:
16:43:54 <slacktopus> <agrabeli> @sweemeng thanks :slightly_smiling_face:
16:44:12 <slacktopus> <julie> That'd be great indeed
16:44:14 <slacktopus> <sweemeng> darkk got it
16:44:26 <slacktopus> <agrabeli> So time is running up: Let's move onto the 2nd topic? (we can also continue discussions on the 1st later on if people wish to)
16:44:48 <slacktopus> <sweemeng> can ellaborate on 2)?
16:45:02 <slacktopus> <hellais> Relevant tickets on the citizenlab repo about this are: https://github.com/citizenlab/test-lists/milestone/1
16:45:08 <slacktopus> <agrabeli> #topic 2: How to ease transfer of material for shutdown testing in countries far from Europe and the US, with questionble postal services?
16:45:31 <slacktopus> <agrabeli> So the context around this involves shipping Raspberry Pis around the world to run ooniprobe
16:45:44 <slacktopus> <agrabeli> And there have been various challenges associated to this
16:45:54 <slacktopus> <agrabeli> Would the individual who added this topic like to share some words?
16:46:01 <slacktopus> <sweemeng> raspberry pi itself is not illegal, and many country know of welcome it. When I last check, I could be wrong
16:46:05 <slacktopus> <sbs> good point :slightly_smiling_face:
16:46:12 <slacktopus> <sweemeng> the image on the other hand :S
16:46:17 <slacktopus> <sukarn> It's not about legality, it's about reliability of postal services
16:46:21 <slacktopus> <sukarn> (I think)
16:46:57 <slacktopus> <agrabeli> @sukarn: Yes, though I think the legality component could be interconnected potentially
16:47:40 <slacktopus> <sweemeng> if using the pi's why not rely on local user group that might have extra pi's at hand?
16:48:04 <slacktopus> <agrabeli> As many of you know, ooniprobe can be run across platforms, including Raspberry Pis
16:48:05 <slacktopus> <julie> I did: was thinking of countries like Chad where we had huge issues shiping material in the country. Is it possible to reach out to OONI team in countries contexts whre a shutdown could happen( let's say 70% chances), so that material can be shipped way in advance of the shutdown?
16:48:06 <slacktopus> <sweemeng> if image is considered sensitive, encrypt it?
16:48:45 <slacktopus> <agrabeli> One of the main reasons why we created a distribution for RasPis was to enable less-technical audiences who don't feel comfrotable with command lines to also be able to run ooniprobe.
16:49:27 <slacktopus> <agrabeli> @julie: Yep! I think planning the shipment in advance (especially in countries where censorship events are more likely to occur) is important.
16:49:28 <slacktopus> <sukarn> I have received dozens of international packages in India. As long as they have a tracking number and don't contain anything illegal, they are almost sure to be delivered (1 out of ~45 was not delivered to me), although there may be a delay or there may be customs charges on the package. Packages without a tracking number are totally hit or miss. They may be delivered, or they may disappear along the route.
16:49:44 <slacktopus> <dbyko> @sweemeng had trouble with this in Ghana - only one local user group with Pi's in Accra, couldn't find others in neighboring towns.
16:50:11 <slacktopus> <sweemeng> oh
16:50:18 <slacktopus> <darkk> @agrabeli does not mobile ooniprobe solve the issue of delivery ? Cheap android device is probably ~same in terms of cost as RPi + SD + case + PSU.
16:50:49 <slacktopus> <tafiti> With ooniprobe-mobile, I see a much more easier way of collecting measurements
16:50:56 <slacktopus> <sweemeng> yeah ooniprobe on android works well.
16:51:02 <slacktopus> <agrabeli> @julie: That said, shipping hardware might not always be the best solution -- particularly in countries where challenges with postal services have occurred. We going to start moving more away from RasPis, and we would like to enable users to run ooniprobe more easily from their devices. To this end, we launched the mobile app that can easily get installed, and we have (longer-term) plans of creating web applications for Windows and Mac device
16:51:34 <slacktopus> <julie> @agrabeli that'd be more than awesome
16:51:48 <slacktopus> <sweemeng> how many country don't have access to app store or play store?
16:52:09 <slacktopus> <sukarn> Do you have plans to get ooniprobe-mobile to people if the app is removed from play store in a particular country, and the website gets blocked?
16:52:16 <slacktopus> <julie> would solve postal issue + legal issue (we xanted to run tests in Cameroon, but people being arrested and charged with terrorism for no reason, so we refrained, thinking that it'd involve shipping pis to the country)
16:52:34 <slacktopus> <hellais> I think Iran has their own play store thing, because of US sanctions.
16:52:42 <slacktopus> <hellais> not sure what the situation is there now.
16:52:45 <slacktopus> Action: darkk wonders if he should say Russia because of Crimea :slightly_smiling_face:
16:53:04 <slacktopus> <agrabeli> @julie: Here you can find the various platforms that ooniprobe can be installed and run from: https://ooni.torproject.org/install/ This includes the mobile app and running ooniprobe from a web UI. Currently, the easiest way to run ooniprobe is via the mobile apps, but going forward we would like to create applications that users can easily install on their Windows or Mac machines.
16:53:05 <slacktopus> <tafiti> After oonimobile was released, I saw a spike in countries testing in Africa (places I never even thought people would be interested - like Somalia
16:53:06 <slacktopus> <alin> In Iran people have access to Google Play but there is Cafe Bazar as well
16:53:55 <slacktopus> <julie> I thought the app was only for site censorship tests, didn't know it worked as well in case of total internet shutdown, which is great
16:54:15 <slacktopus> <agrabeli> So I guess we've now already transition to topic 3: How to have a network of OONI testers ready in countries likely to shutdown the Internet?
16:54:27 <slacktopus> <hellais> > Do you have plans to get ooniprobe-mobile to people if the app is removed from play store in a particular country, and the website gets blocked?  We are working on getting it included in F-Droid and we also always share the pre-built APK as well. I don’t know of a way to get an app on a device on iOS without jailbreaking.
16:55:13 <slacktopus> <hellais> > I thought the app was only for site censorship tests, didn't know it worked as well in case of total internet shutdown, which is great  On this note I would actually like to ask if we can clarify what is meant by the term internet shutdown.
16:55:30 <slacktopus> <agrabeli> @julie: The app doesn't examine total shutdowns. Rather, it tests the blocking of websites, it searches for middle boxes in networks, and it measures the speed of a network.
16:55:51 <slacktopus> <julie> @hellais  International bandwidth is cut
16:55:53 <slacktopus> <agrabeli> @julie: We have plans to develop a methodology for examining internet shutdowns in an automated way...but we haven't developed this yet.
16:56:12 <slacktopus> <julie> @agrabeli ok thanks
16:56:28 <slacktopus> <hellais> Intuitively I thought people meant when they say “Internet shutdown” a complete blackout of the internet, but reading more carefully access now website it seems like they use the term also to describe internet censorship
16:57:05 <slacktopus> <sukarn> This question is interesting to me as I work with SFLC.in which has recently launched https://internetshutdowns.in/ that tracks internet shutdowns in India. SFLC.in works with Access Now on this issue. We define an Internet shutdown as "a Government-imposed disablement of access to the Internet as a whole within a particular locality or localities for any duration of time". There are two key-components to this definition:  (1)    An Internet sh
16:57:05 <slacktopus> always Government-imposed i.e. Internet Service Providers serving the locality in question are ordered by an agency of the Government to cut-off Internet services to that area (2)    An Internet shutdown always imposes a blanket ban on Internet access, where access to the Internet as a whole is disabled, and not a surgical ban, where access to particular content/services is disabled leaving access to other content/services unaffected.
16:57:34 <slacktopus> <hellais> "An internet shutdown is an intentional disruption of internet or electronic communications, rendering them inaccessible or effectively unusable, for a specific population or within a location, often to exert control over the flow of information” - which I guess could include also internet censorship as a form of “disruption"
16:57:47 <slacktopus> <tafiti> Complete blackout is also censorship - on the extreme end of the spectrum.
16:57:48 <slacktopus> <agrabeli> @sukarn: Thanks for sharing your project and congrats on the excellent work!
16:58:28 <slacktopus> <sukarn> @agrabeli Thank you for your kind words. Ooni and Tor are essential for our mutual aims, which is why I'm here today.
16:58:35 <slacktopus> <julie> @hellais indeed
16:58:58 <slacktopus> <darkk> complete blackout is hard to confirm without variety of vantage points. local electricity and/or ISP blackout should not look like "Internet shutdown" if we have only one or two probes in the country.
16:59:04 <slacktopus> <agrabeli> @tafiti: I guess the confusion is that OONI generally refers to an internet shutdown when there is a total blackout and the internet as  a whole is inacceesible, in which case it's pointless to run ooniprobe because it wouldn't be able to connect to a network to perform measurements.
16:59:32 <slacktopus> <darkk> BGP stream & aggregated traffic opendata may be better sources for that kind of issues.
16:59:47 <slacktopus> <sweemeng> yup
17:00:06 <slacktopus> <sweemeng> looking glass and what not
17:00:14 <slacktopus> <sweemeng> even ripe-atlas is a better tools on this
17:00:33 <slacktopus> <sweemeng> maybe not atlas
17:00:35 <slacktopus> <julie> @darkk thing is it's interesting to know technically what was used to proceed to the blackout, because some gov say it's because ate a cable or something like that, when they actually ordered ISP to proceed to blackout
17:00:52 <slacktopus> <julie> *a rat ate a cable (true story)
17:01:00 <slacktopus> <tafiti> @agrabeli When you think about it, any intentional disruption is censorship. It is the extent of content intentionally blocked that defines it. I agree, there has been an assumption on this.
17:01:50 <slacktopus> <sukarn> @agrabeli but there might be situations where the internet is shut down using techniques that interfere with DNS lookups, instead of turning off the entire infrastructure. In that case, a bypass should be possible.
17:02:08 <slacktopus> <agrabeli> @julie: Indeed! though as @darkk mentioned it's usually important to be able to collect data from a variety of vantage points. As part of our priorities though over the next year, we would like to develop a methodology for examining shutdowns in an automated way.
17:02:23 <slacktopus> <tafiti> The Rats Ate The Internet (An apt T-Shirt message for Internet Shutdown government officials)
17:02:38 <slacktopus> <agrabeli> @sukarn We usually refer to those cases as "censorship events", rather than "shutdowns".
17:02:47 <slacktopus> <sukarn> Oh, okay. Thank you
17:02:53 <slacktopus> <agrabeli> @sukarn: Those cases (DNS blocking) are detectable by ooniprobe
17:03:23 <slacktopus> <sweemeng> so are we still on this? or we should go on to 3)?
17:03:32 <slacktopus> <sukarn> Yes, I know. I read through all the test-specs. There's only one that I could not understand, and that one is on agenda 6.
17:03:41 <slacktopus> <agrabeli> @sukarn: And ooniprobe mobile app provides tips on how to circumvent the specific type of censorship that is detected (for example, if sites are blocked based on DNS blocking, then the app suggests circumvention by using a Google DNS resolver instead).
17:04:06 <slacktopus> <agrabeli> @sweemeng: I think we've kind of transitioned to topic 3...in a way :slightly_smiling_face:
17:04:11 <slacktopus> <julie> Linked to 3) : OONI will be at rightscon, right? Are you doing a session to explain your work, and potentially get in touch with CSO interested in running tests?
17:04:42 <slacktopus> <agrabeli> @julie: Yep! And @hellais and @sbs will be hosting a session related to shutdowns I believe
17:05:01 <slacktopus> <hellais> @julie yes we are going to be at rightscon and there we have a session specifically on internet shutdowns (with the definition of shutdowns as in full internet blackout)
17:05:14 <slacktopus> <agrabeli> @sukarn: Oh wow, great to hear you read the specs! Perhaps we can discuss agenda 6 on a private channel?
17:05:40 <slacktopus> <hellais> Do we want to move onto point 4)?
17:05:56 <slacktopus> <sukarn> @hell @agrabeli Nice @agrabeli Sure. That would work for me, if you do not have any issues with doing that.
17:06:18 <slacktopus> <hellais> Can the person that wrote 4) say a few words on it?
17:06:44 <slacktopus> <agrabeli> @sukarn: I'd love to! Just being conscious of the time and some people being at 1am...
17:07:02 <slacktopus> <sweemeng> sorry it is 1am here.
17:07:22 <slacktopus> <sweemeng> someone will see a panda moving around in the office in the morning
17:07:36 <slacktopus> <agrabeli> @sweemeng: So so sorry about that, and THANK YOU for joining us! We'll try to set a better time for the next meeting!
17:07:36 <slacktopus> <sbs> lol
17:07:50 <slacktopus> <sweemeng> so who is on 4)?
17:08:37 <slacktopus> <hellais> @sukarn re: 6) meek_frontend_requests test is a test that checks to see if the domains used by the Tor pluggable transport are working. The meek Tor pluggable transport uses a neat treat that involves connecting to a popular CDN (such as amazon cloudfront) and inside of the encrypted TLS request the `Host` of the domain they specify the domain of where they really want to connect to and the CDN will transparently route it to the correct domain
17:08:38 <slacktopus> checks to see if the front domains used by meek work.
17:09:11 <slacktopus> <hellais> For the record 4) is "Visualizing censorship, surveillance, tests and results as part of resources for OONI training and partly data presentation"
17:09:12 <slacktopus> <sweemeng> there is a hackathon on rightscon right? can we do 4) for the hackathon?
17:09:25 <slacktopus> <agrabeli> OK so regarding 4.... it sounds like this topic is about creating visualizations based on the OONI data that can be useful to trainings and presentations
17:09:42 <slacktopus> <agrabeli> @sweemeng: +1
17:09:56 <slacktopus> <hellais> I guess what I can say on that in the upcoming months we are going to be working on some better visualisations to be integrated into explorer also based on better mining of the OONI data with the new pipeline.
17:10:21 <slacktopus> <hellais> If there are some specific visualisations or questions you would like to have answered visually please let us know and we can see what we can do :slightly_smiling_face:
17:10:29 <slacktopus> <hellais> @sweemeng that’s a great idea!
17:10:42 <slacktopus> <sweemeng> i need more visualization work thanhttps://sinar.github.io/popit_visualizer/?entity=persons&entity_id=545e446f5222837c2c058cad
17:10:56 <slacktopus> <sukarn> @hellais Thank you. There are still some parts of that description that I do not understand. I'll ask @agrabeli about those parts, so that this discussion can continue without further interruptions.
17:13:08 <slacktopus> <agrabeli> So please contact us with any suggestions/ideas/wishlists for visualizations that you would like to see OONI do
17:13:21 <slacktopus> <agrabeli> You can reach us here or via email at contact@openobservatory.org
17:13:34 <slacktopus> <agrabeli> We'll also be attending IFF next week and RightsCon at the end of the month
17:13:45 <slacktopus> <agrabeli> During which we're happy to brainstorm further with you in person
17:13:47 <slacktopus> <sweemeng> awesome
17:14:27 <slacktopus> <agrabeli> Given that we're past time and it's extremely late in @sweemeng's time zone, perhaps we can start to end the meeting?
17:14:35 <slacktopus> <agrabeli> We are also available on slack afterwards
17:14:41 <slacktopus> <agrabeli> And you can find us here everyday
17:14:53 <slacktopus> <agrabeli> Feel encouraged to pop-in and ask questions or engage in discussions with us
17:15:05 <slacktopus> <sweemeng> good night folks. nice chatting with you all
17:15:12 <slacktopus> <sweemeng> see you in rightscon!
17:15:14 <slacktopus> <agrabeli> We're thinking of having community meetings which are either monthly or bi-weekly. What would you prefer?
17:15:24 <slacktopus> <sweemeng> monthly is good
17:15:28 <slacktopus> <agrabeli> @sweemeng thanks, good night! :slightly_smiling_face:
17:15:42 <slacktopus> <sukarn> Thank you, people. This was fun. I'd like to participate again next time.
17:15:47 <slacktopus> <hellais> Most excellent, thank you all for attending!
17:15:52 <slacktopus> <eiko> physical meet up in RightsCon would be great :slightly_smiling_face:
17:16:08 <slacktopus> <agrabeli> @eiko: +1 :slightly_smiling_face:
17:16:31 <slacktopus> <agrabeli> So let's say that community meetings are monthly?
17:16:35 <slacktopus> <agrabeli> Votes before we leave?
17:16:46 <slacktopus> <julie> +1
17:16:51 <slacktopus> <hellais> Or it can happen asynchronously
17:17:00 <slacktopus> <sukarn> @agrabeli Monthly or bi-weekly; either works for me.
17:17:45 <slacktopus> <tafiti> julie: Hello
17:18:01 <slacktopus> <tafiti> Which country is that that said "The Rats ate a cable?"
17:18:24 <slacktopus> <agrabeli> Great, thanks everyone!
17:18:55 <slacktopus> <agrabeli> I'll circulate an email (and the pad for next agenda topics) for the next community meeting next month
17:19:22 <slacktopus> <tafiti> Thank you @agrabeli for making that happen
17:19:41 <slacktopus> <agrabeli> @tafiti thank you and everyone else for attending :slightly_smiling_face:
17:19:44 <slacktopus> <sukarn> Are the emails based on our slack accounts?
17:20:34 <slacktopus> <agrabeli> @sukarn argh good questin. I was thinking of circluating it on the #ooni-talk mailing list
17:20:47 <slacktopus> <hellais> ooni-talk: https://lists.torproject.org/cgi-bin/mailman/listinfo/ooni-talk
17:21:00 <slacktopus> <agrabeli> Please subscribe to #ooni-talk: https://lists.torproject.org/cgi-bin/mailman/listinfo/ooni-talk :slightly_smiling_face:
17:21:03 <slacktopus> <sukarn> Thank you. I'll join that list.
17:21:07 <slacktopus> <darkk> TWIMC https://lists.torproject.org/pipermail/ooni-talk/2017-February/thread.html is quite low-traffic :slightly_smiling_face:
17:22:04 <slacktopus> <hellais> Great, thank you all again for attending
17:22:06 <slacktopus> <agrabeli> @darkk: Indeed....well, people please feel encouraged to share news/thoughts on the mailing list, and together we could make it higher traffic! :slightly_smiling_face:
17:22:09 <hellais> #endmeeting