Ask Me Anything: MailPile
The prospect of a highly usable and user friendly PGP email client continues to elude us. The participants of our most recent Ask-Me-Anything, however, are setting out to change that. Is MailPile really the answer to our PGP woes?
Read the transcript to learn more about MailPile AMA including: it's interesting origin story, the reasons behind its search-based design, and how the developers of MailPile value the input of translators from around the world. Want to know more about MailPile? Check out www.mailpile.is or get involved in localizing it on Transifex.
Localization Lab’s Second Ask Me Anything took place on May 21, 2015 with the team from MailPile. Below is a transcript of the questions and answers.
Localization Lab: Hi everyone, it is 11:30 ET (3:30 GMT) so we are going to get started.
Localization Lab: Please note that people often come in throughout the AMA so you will probably see me repost instructions a few times
Localization Lab: I am your friendly facilitator today. There are also a couple other Localization Lab team members here who will also be helping out.
Localization Lab: We will be taking questions from the room throughout the chat, and I will also be sharing questions that have come in over the past few days and weeks.
Localization Lab: We want to make this as relaxed and casual as possible. If you have a question, feel free to post it right here in the chat.
Localization Lab: If it gets lost in the shuffle, I will call attention to it later on.
Localization Lab: Before our first question, I want to send a big THANK YOU to Bjarni for joining us today.
Localization Lab: You can see an example of how we process the logs by looking at the GlobaLeaks ama here: www.localizationlab.org/globaleaks-ama
Localization Lab: I also want to say a big thank you to all of the folks in the chat on time. Thank you for joining us!
Localization Lab: OK! Bjarni, I want to start off with a simple question:
Localization Lab: QUESTION: Bjarni, why did you decide to start developing MailPile?
BjarniRunar: Well, Mailpile started twice. The first time, it was an experiment to see if I could write a search engine for e-mail.
BjarniRunar: I have, as hobbies, tried to write mail clients a couple of times before and failed. One day I had a realization that modern computers had gotten so big and powerful, that some of the old assumptions were no longer needed and it would actually be quite easy to write a simple search engine that handles all of one person's mail - a lifetime worth - on consumer hardware.
BjarniRunar: So I wrote a proof of concept, and called it Mailpile. I gave a 5 minute lightning talk about it at FSCONS in Sweeden, I think that was in 2010. And then I set the project aside.
BjarniRunar: My friend Smári was impressed by the idea and unhappy with the free software mail clients he had available to him. So he kept badgering me to improve Mailpile.
BjarniRunar: We started to toy with the idea of running a crowdfunding campaign, to work on it full time. And then we met Brennan, somewhat randomly, and he was the missing link - someone with design and communication skills to bolster our tech skills.
BjarniRunar: So Mailpile had a second start in the summer of 2013, when the three of us launched the campaign on IndieGoGo. Out of pure luck, the timing coincided with Snowden's revelations, so we got a lot of support and met our goals.
BjarniRunar: That was the second beginning... and Mailpile has pretty much been my full time job since then.
Localization Lab: I love hearing origin stories like this it takes so many pieces to bring things together
Localization Lab: That is wonderful, thank you for sharing Bjarni
Localization Lab: QUESTION: Can you briefly explain how MailPile works and who your intended users are?
BjarniRunar: The really short answer is: Mailpile is an e-mail client that you can run on your own computer, and we hope as many people as possible choose to do so. We do not specifically target technical users or activists.
BjarniRunar: But there are a few things about Mailpile that make it different from other mail clients.
Participant A: how does this make email encryption easier
BjarniRunar: (I will get to that)
Localization Lab: thanks, Participant A! and thanks bjarni for getting to it
BjarniRunar: One of the things that makes Mailpile different, is that it is built around a search engine. Similar to Google's web-mail, one of the core concepts is that an e-mail client that makes it easy to search will allow people to interact in more productive ways with their mail than a traditional folders-and-mailboxes approach.
BjarniRunar: Another thing that makes Mailpile different is the way we approach the user interface. Most desktop applications use windows and widgets provided by the operating system. Mailpile users web technology instead, so although you run Mailpile on your computer, you use your web browser to interact with it.
BjarniRunar: This allows us to take advantage of the advances that have been made in web design, and makes it easy for web developers to contribute to our project, even if it is (usually) a desktop app.
BjarniRunar: Finally, Mailpile is different because we decided to think about encryption from the very start and use it wherever applicable.
Participant D: BjarniRunar: can you talk more about the productivity side benefits given MailPile's basis on a search engine?
BjarniRunar: Participant D: ok, I'll add that to my mental queue
Participant B: Does that make it more secure, to keep everything on your hardware or improve accesibility when there is insecure connection
Localization Lab: Bjarni, i am tracking these questions as well so if you are done answering a previous one, i can share what the most recent unanswered question is
BjarniRunar: As an example about encryption being baked in from the start, we for example take care to encrypt the search index itself on your disk. This in turn allows us to make it possible to search the contents of encrypted mail, which most e-mail clients can not do. And in turn this makes encrypted mail more convenient and pleasant to work with - you can encrypt without sacrificing the ability to search.
BjarniRunar: So these things work together at a very low level. The goal is to make working with encrypted mail as seamless and convenient as working with unencrypted mail.
Participant C: ^ Which is the holy grail of encryption
Participant B: yes
BjarniRunar: Regarding the productivity benefits of a search engine - here I don't have research to back me up, I just have gut feelings and instincts.
BjarniRunar: But I use the search engine all the time, and it relieves me of the burden of having to sort and organize my mail. I am not a very organized person when it comes to paperwork-style things, and e-mail is in that category.
Localization Lab: The next question in the queue is from Participant B (“Does that make it more secure, to keep everything on your hardware or improve accesibility when there is insecure connection?”) - Bjarni, feel free to answer that after the search enging question when you are ready. No rush, plenty of time
BjarniRunar: Being able to just let my mail pile up and search for people or topics has made my life much simpler, and I expect that will apply to many others.
BjarniRunar: Related to that, Mailpile does offer tools for organizing - but instead of folders, Mailpile users tags. Tags are what GMail calls labels.
Participant B: In tunisia it has become very important to learn how to protect email messages, would this make it easier
Localization Lab: Also a reminder to everyone that follow-up questions are encouraged! If you have a question in response to an answer Bjarni gives, feel free to jump in and ask it.
BjarniRunar: Instead of requiring that a message be in just one folder, you can assign as many tags as you need. So it's a more flexible tool.
BjarniRunar: People who use GMail will find this very familiar, but it turns out not many Free Software applications offer these features. So Mailpile is filling a void there.
BjarniRunar: ok... back to security and your own hardware
BjarniRunar: Security is really, really tricky. Different people have different security needs and what improves security for one person, may do the opposite for another.
Participant C: Re: Tunisia, Mailpile is at 25% and we could use help with making it available to Arabic speakers around the world! https://www.transifex.com/projects/p/mailpile/language/ar/
BjarniRunar: One of the hardest parts about security is just helping users understand their own needs.
BjarniRunar: But if you can reach a point here someone's e-mail is on a physical device that they control, then they can start to understand at least some of the security in a much more natural way.
BjarniRunar: We all know how to keep valuable possessions safe.
BjarniRunar: If it's your mail, and it's on your computer... and you want to be sure nobody has access to it, you just turn it off and lock it in a safe. Or carry on your person.
BjarniRunar: This is security which is very easy to reason about and understand. It may not work for everyone, but it is at least accessible.
BjarniRunar: (gathering my thoughts a bit)
Localization Lab: no problem bjarni
Localization Lab: take a break, you’v ebeen typing nonstop!
Localization Lab: while bjarni is gathering his thoughts ill share a few things
Localization Lab: first, please continue to ask questions and i can add them to the queue. i have pre-submitted quetsions on the list but i am prioroitizing chat questions
Localization Lab: the next question i have in the queue is from Participant B: “In tunisia it has become very important to learn how to protect email messages, would this make it easier”
Participant C: BjarniRunar: Agreed, I think with training it’s true also
Participant C: that the first thing, and arguably most important thing
Participant C: is to understand HOW we communicate, and then where the gaps in security are, and how to protect oneself
Localization Lab: second, bjarni you make an interesting point on physical devices and “we all knwo how to keep valuable possessions safe” - soem of us at SecondMuse have done research that emphasizes that as well, such as our report on vietnamese digital activists (internetfreedom.secondmuse.com)
BjarniRunar: OK, the Tunisia question actually dovetails nicely with what I was saying about physical security. In some cases, your physical person may be an unsafe place for your data. So carrying your Mailpile around is safer for some, but more dangerous for others.
Localization Lab: After this question, I have a pre-submited question. Let me know when you are ready for it, Bjarni.
BjarniRunar: In the case where it's dangerous to carry it, because Mailpile's user interface is based on web technology, you actually have the option of leaving it behind somewhere you consider safe, and accessing it over the network.
Participant C: These are characteristics of the TSyrian and Iranian users too—physical safety
Participant C: and the danger of your hardware being confiscated
BjarniRunar: I honestly do not know if that will help people in Tunisia, as I do not understand their risks well enough.
BjarniRunar: But I imagine that reporters and journalists entering dangerous areas, will want to leave their Mailpile behind and access in that way.
BjarniRunar: We have not yet implemented the code required, but it is planned - we want to make it very easy (just a click of a button) to make your Mailpile accessible over Tor, as a hidden service.
Participant B: during ben ali the information ministry was spying on people
BjarniRunar: We also want to make it easy to access it over the normal web, for people with lesser security concerns.
Participant B: since there has been more transparent
BjarniRunar: I don't know if I have answered the question - what do you think?
Localization Lab: I think you answered very well and honestly! I appreciate your honesty. It is good to be able to say “i dont know the risks well enough to know for sure”
Participant C: Totally!
Localization Lab: the next pre-submitted question is somewhat related
Localization Lab: and perhaps difficult to answer, but let’s see!
Localization Lab: QUESTION: If you could learn more about users of Mailpile in any region or community around the world, where would it be and why?
BjarniRunar: perhaps one thing first
Localization Lab: (answer whenever you are ready, you can keep answering the last question if you’d like)
BjarniRunar: The other way Mailpile *may* help improve security, is hopefully our interface will make it easier to encrypt without making mistakes. This is an explicit goal in our UI design and development. How well we do, well only time can tell, but we intend to do our best and want to do usability studies and such to that end.
BjarniRunar: So on to the next question - learn more about users of Mailpile, where and why?
Participant C: Also how can WE help improve Mailpile
Participant C: (answer when you can) sorry to jump in
Localization Lab: good question, Participant C! ill add that to the queue as well. you can answer that next, Bjarni
BjarniRunar: Since Mailpile is still under development and isn't really useful to non-techies, we don't have many users yet. I expect that when we do have users, I will be very curious to know which communities like the program and why.
Participant A: why not for non-techies
BjarniRunar: I am really looking forward to that. And I don't have any particular preferences as to where the program finds a home. Not yet anyway!
BjarniRunar: To clarify, it is not at the moment useful to non-techies, simply because it is still under development and has a lot of rough edges.
BjarniRunar: Our goal is to fix that so everyone can use it, but we're not there yet.
BjarniRunar: Hopefully we will be able to release a 1.0 at the end of the summer which everyone can use.
BjarniRunar: ... ok, next question, how can people help improve Mailpile!
BjarniRunar: At the moment, Mailpile is at the stage where programmers are the most important resource. Lots of code needs to be written.
BjarniRunar: Later this summer, there will be a third beta release. At that point, testing the program and sharing opinions becomes very valuable, so we can identify flaws and fix them before the 1.0.
Localization Lab: (reminder to the room: we have about 20 minutes left so please be sure to get your questions in! i have another question in the queue now that will be asked next)
Participant C: Interesting BjarniRunar, how can we “share opinions”?
BjarniRunar: And then of course, as the 1.0 approaches, I will make sure the strings in the application stabalize, so people can translate.
Participant C: Is there a place we can provide feedback?
BjarniRunar: Participant C: Currently the only interface we have for that is the project's github issue tracker.
BjarniRunar: Participant C: or informal chat on IRC - we have a channel named #mailpile here on Freenode.
Participant C: —aside from translating Mailpile for our respective communities, that’s always a way how we can improve
Participant C: and we can find you on #mailpile channel regularly?
BjarniRunar: And finally, once Mailpile 1.0 is out, I would love to get help from people documenting and explaining. Videos teaching people how to use the software, blog posts, thhings likek that.
BjarniRunar: yes, I am on #mailpile all the time
Localization Lab: Great!
Localization Lab: Ok, the next question is from me! It is a follow up to the earlier point about non-techies and development
Localization Lab: QUESTION: What are the rough edges that need to be worked out to make MailPile useful for non-techies?
BjarniRunar: Sometimes I'm "idle", but if I'm not hiding away to concentrate or asleep I'm usually active on the channel.
BjarniRunar: Localization Lab: There are so many rough edges... where to start!
Participant C: Start with the roughest edge
BjarniRunar: The things I have identified as critically needing to be fixed before the 1.0 are, roughly as follows...
BjarniRunar: 1. I want to make it easier to get the software up and running, and easier to configure it.
BjarniRunar: 2. The user interface is extremely slow at the moment, it needs to be optimized so it works well even on slow computers.
Participant A: slow computer or slow conection?
BjarniRunar: 3. There are problems with how data is currently stored to disk, which makes Mailpile very slow to start up if you have a lot of mail, increases the risk of data loss if the app crashes, and makes it hard to run Mailpile on shared computers.
BjarniRunar: re Participant A's question about 2. - both, but mostly slow computers, remember our primary use case is people running Mailpile on their own hardware, so it needs to work well on what they have.
BjarniRunar: Developers often have very fast beefy machines, that's not necessarily the case of the general public.
BjarniRunar: 4. We have defined a security model, which we need to finish so Mailpile can live up to its promise of helping people communicate securely.
BjarniRunar: So those are sort of the big tasks for the summer, for me.
Participant E: Will you post on the Localization Lab Google group once MailPile 1.0 is released so that we know to go and try it out?
Participant C: re internet speed, I feel like we also get used to fast internet, even though we dont think it is sometimes— internet is slowed down purposely in iran and syria
Localization Lab: Those are some big tasks, thank you for sharing them with us
Localization Lab: And i agree iwth Participant E’s point - be sure ot post it on the localization lab google group! (if that is not configured for you, let me know and i can get you sorted out)
BjarniRunar: Participant E: I should join that group and talk to you guys more often. I hope someone can mail me a link to it aafter this chat.
Localization Lab: ill be sure to do that!
Localization Lab: our next question is a pre-submitted one and is similar to an earlier one about feedback
Localization Lab: but it is more focused on translators:
Localization Lab: QUESTION: What is the best way for translators to communicate with you with questions about the text and/or to check on integration of translations in new releases?
Localization Lab: by the way, the localization lab google group can be joined here: https://groups.google.com/forum/#!forum/otfl10n just submit a request to join and we will approve it!
BjarniRunar: That is a really good question. The BEST way, is to talk to me on IRC. But that's me selfishly suggesting what is best for me.
Localization Lab: your preferences matter too, bjarni
BjarniRunar: I am afraid I am so overloaded with communication on most other channels, that there is a significant risk of delays or me missing posts if you use Github for that sort of thing. E-mail is actually a pretty good option, if people have lots of questions and are not in a big hurry.
BjarniRunar: One of my personal e-mails is email@example.com, and I am very happy to hear from translators.
Localization Lab: Bjarni, you made a point earlier (before the AMA) about some thoughts on translation for mailpile. i think they relate well to the next question:
Localization Lab: QUESTION: If there is one thing you want translators to know about MailPile, what would that be?
Localization Lab: (also this is a 10 minute warning! the AMA will be over before we know it, so please get in any additional questions now. we expect to ask at least a couple more)
BjarniRunar: I want them to know that I really, really value their work. Coming from a small country like Iceland, I do not take it for granted that software is available in my native tongue.
BjarniRunar: So Mailpile would like to make translators' job as easy and pleasant as possible, and to that end we have created a very primitive plugin for Mailpile that tries to help find the context of strings.
BjarniRunar: So there is actually code in Mailpile already that is meant to help translators do their job. It needs more work, and we probably also need to make sure that there are demo installations of Mailpile accessible that have the plugin enabled.
BjarniRunar: I have a lot on my plate, but if translators would like to request features that make their job easier, I am keen to implement them if they're not too difficult.
Localization Lab: i just think that is so cool that you guys are building that plugin
Localization Lab: and a short follow-up question to that, also submitted with it, is: How do you honor (or how do you plan to honor) the volunteer translators that work on this project?
Participant C: It would be great if we could replicate this for other projects
Participant C: it’s hard to translate out of context, we usually have a reviewer “test out” a project after localization is complete to catch any of these
BjarniRunar: That's a great question. I would like to have a credits page somewhere in the app. Currently it's very easy for me to generate a list of people that have contributed code, by using git's history, but that would ignore the contributions of translators. Does Transifex make this easier?
Localization Lab: Bjarni, that is a good question. Participant C, do you know?
BjarniRunar: Basically, if you guys can give me a list of people to thank, I will make sure they get mentioned.
Localization Lab: we wIll work together to make sure that happens!
Participant C: Oh yeah!
Localization Lab: ok, we have just five minutes left. i want to ask one more question and if we have time for another we can squeeze one in.
Participant C: We can definitely make that happen, we want to give credit to all those folks who spend hours and hours making these tools available in their language
BjarniRunar: We also have a thank-you page on the www.mailpile.is site, and thanking contributors and translators there should also happen. I need to revisit that and make sure people are not being left out.
Localization Lab: that’s great, thank you bjarni!
Localization Lab: QUESTION: What security tool would you like to build or work on next? Why?
Participant A: so should we use mailpile now or wait until next release
BjarniRunar: The next thing I want to work on... is Mailpile. The 1.0 will by no means be the end of the road, I have so many ideas I want to implement that this is going to take years if only I can figure out how to pay myself a salary to do so!
Participant A: in latin america it is also dangerus to carry laptops with you
Localization Lab: Ha, good answer!
Localization Lab: OK, if there are no more questions at this time I am going to start wrapping things up.
BjarniRunar: Localization Lab: what's nexxt, shall I answer Participant A? "
Localization Lab: sure!
Localization Lab: answer Participant A then i will wrap up
BjarniRunar: Participant A: I would not recommend non-technical folks use Mailpile right now, there are too many bugs.
Participant A: we can work on translation still
BjarniRunar: The Beta III release will hopefully finalize the on-disk storage formats and things, so testing the app at that point should be safer, and allow for a seamless upgrade to 1.0 when it is ready.
Participant A: how will you tell us when we can use it
Participant C: You can also post on out OTFL10n Google group page, BjarniRunar
Localization Lab: Yes! That is the best way.
BjarniRunar: Yeah, that's really the question - whether I should be posting announcements in your forums.
Localization Lab: OK, everyone, it is now one hour in!
Localization Lab: First, I want to say a huge thank you again to Bjarni for typing until his fingers are probably sore
BjarniRunar: We currently announce releases on Twitter and our blog.
Localization Lab: And second i want to say a huge thank you to everyone who has joined and listened ro asked questions.
Participant C: BjarniRunar: it’s a good way for everyone to keep tabs of announcements
BjarniRunar: Thanks for the interest guys!
Localization Lab: We will be putting this whole transcript up on localizationlab.org within the next week. And we will anonymize everyones names (except Bjarni who said we dont have to!)
Localization Lab: Bjarni, can you share your email address in case anyone has any follow up questions?
BjarniRunar: yes, firstname.lastname@example.org or email@example.com (they go to the same place)