Logo TIB AV-Portal Logo TIB AV-Portal

Settling the IM war

Video in TIB AV-Portal: Settling the IM war

Formal Metadata

Settling the IM war
Title of Series
CC Attribution 3.0 Unported:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor.
Release Date

Content Metadata

Subject Area
The world of Instant Messaging is populated with hundreds of providers - all incompatible with each other though history has shown that walled gardens are not sustainable. Why are we unable to agree upon a standard to communicate with each other?
Indexable Computer animation record
point Android Context services clients regular computational number goods naturally Conversation model form systems standards Arm Information Development Ordinary Differential Equations instant messenger bits meetings lines system call processes communication Right protocols
standards services views time share instant messenger clients part subset message-based Lecture/Conference level sort extent Routing protocols Windows tasks
point server non-existence services link files time print clients attributes image inference Facebook's rates different naturally encrypted box Conversation series extent Fingerabdr├╝cke multiple Development cellular horizon effects staff limitations several message-based means commitment statements Website sort progress
point implementation server services studies states time sources clients open subsets subset versions causal Conversation extent addresses classes install standards Information Development load The list mining message-based repo Drucke
server services Computer animation Information operations orders diagrams argument band extent systems
Computer animation operations The list Right help extent programs systems
point Actions implementation states time sheaf loss argument clients programs number Chat mathematics Forum different hypermedia testing model extent boss standards response Development The list physical plan bits lines indicators limitations system call processes Google case protocols Results write
point implementation Actions services time decision loss clients argument theoretical metadata number product Facebook's different terms naturally arrow Conversation model lies extent Security optimal plugin systems form standards multiple Arm Information Development cellular The list analysis infinity basis message-based environment case communication statements organization Go <Programmiersprache> progress protocols
point current Actions implementation attitudes sets theoretical fields Twitter odd mathematics Iris testing Office Conversation Security association Fingerabdr├╝cke standards Fokus Arm Development independent proof particle message-based processes Computer animation case logic retrieval mix Right sort protocols progress reading
and welcome to the session after lunch that is good is going to talk to us will tell us something is not messaging index of PP and very exciting world grounded of a reminder this is being recorded so the only problems of the advised recording this lots of fun who you don't you are I think you will come to my talk on the I M
war of creating an open federated is a messaging protocol I know something about me i'm doing a good job and from early known right now for our as being z primary developer of the Android Nixon pp planned conversations the all conversations just like regular information but it's relies on all the unferreted standards firm but this is not going to be an enticement talk what conversations so I'm just going to say this once all check it out it's northern clients are confident that I am um that the only after talk but I'm also a little bit involved in the are broader accepted the community arms I'm in interested context was little so the developers was other client developers was and I got a regular meetings there it's p community or it up and discusses the changes and if you want to know more about me I've side or called food sleep and yeah
so but enough about me on who uses more than 1 system is inclined at all OK so more than 3 OK I thought on the using conversations already are who are very slow the last the we will there's always been a lot of computing information and services they candidate for example create ICQ MSN and a couple of more however in recent years unchanged the but 1st of all information became more and more important instead of turning on our computers once a day and strategic ODE model see was online on ICQ but green all line 24 7 the on an instant messaging just became a very convenient form of communication but it's less obtrusive than the telephone call for example but it's allows for a fully natural way of communication the so the 2nd thing that changed was that the number of available instant messages exploded another point we have been trying to
me also this issue possible task of well that's not a problem you might think as long as everybody uses what that but in fact not everybody does on 1st of all of the ah everybody uses what's the our views of so the European view on its own if you look at our other parts of the world world of the market share of what it is not a significant and as a significant cents-off what that further diminishes if you start looking at what kind of instant messages solutions you are using it From at work 2 flights like to chat or even sky becomes but more important I horrible at the end of the day all those tools do pretty much the same thing on that's why there's a a huge on demand for all the so called multiprotocol clients like pigeon for example are truly this which is popular on Windows all the latest 1 could France routes which incorporates all their clients 1 you I the so this leaves us with the question on why isn't a standard protocol for and the messaging that would make all other services interoperable well actually there is on on such automatically exists and it has been for more than a decade on the if it's called BP was of which is sort for Extensible Messaging and Presence Protocol and so
on the not going to go into too many technical details the on the key thing to remember is that its extendable so at a very low level you are just able to exchange messages between 2 devices and the higher level you can put almost everything on top of it on the basic protocol is an IDS stand and the extensions of formalized as so-called accepts the buys acceptable standards foundation the but because of so those extensions X pp can keep up with the changing requirements that naturally occur over time on I've already talked a lot about the specific extensions in will it but a full-fledged international solutions I the for example at my or last year's talk the first one are called X and compute 2015 challenges of what is the messaging or up on and more recently wrote an essay on to called state of mobile between
2016 so if you are on want to know more about the technical details go check those hours on this elite from that site so who has read my essay or been to last year to again full of people yeah the rest of you should definitely check this out if you are interested in the technical details about like an end-user shot and using progression care about all this on the the the the so so I and my last year's talk on with a statement that was they pp gone now able to send an attribute messages which can also be an image to someone uses multiple devices and it's also survives can also be so this might not sound to revolutionary but for example the Facebook Messenger which was inference and Trent encryption all the effects users when you turn on transcription messages you have to select 1 specific device but to get to that point so we had to introduce him extensions on the no 1 was called HDP upload which shifted from the traditional our method of 500 is Nixon BP which was peer based where can only exchange messages from 1 specific to what device to another device to another message we basically uploads file an image for example to the server and then just attribute link to the other the the rest of the contact and the other extension introduced are capable of encrypting to multiple devices Scott communal the last August at the time of my talk only glide capable all from the know what organizations on and during the past year on the other extension introduced it to be lot has seen a pretty huge adoption rate several clients and service have non native support for the however co-mingled is in a slightly more complicated situation the good news for us the all last December really will be will use the very basic Plotkin for hazardous applied at which point the the multiple the guy says that of a booming will actually starts to make sense the but unfortunately of course suffers from 1 big problem Due to its very nature off involving encryption and multiple devices and fingerprints and stuff like this on implementing can take up to several days if not weeks the and the in a vault where development is mostly done the volunteers it makes a huge difference is something can be implemented in the afternoon of it's a week-long commitment the and this was something like it should be upload which can easily be implemented in laughter it gets adopted print trade the especially because you knew the series results the found the the the with the mean on the other hand you would simply just spent days just preparing staff and setting up the infrastructure before you are able to the center civil had world for example to and for a certain that seminal plucking forgotten only happens because we did we did the initial work all cells and I didn't mind putting an our the work because I commercial interest in it and try to put if it's frustrating for a couple of days and after that actually someone else took over because of the we have so bassists limits in they just have incremental steps and fixing Barker boxes doing UI improvements and sort the the and the only other mingled annotation on the horizon now is a 1 should secure which is also not being developed by volunteers but by all paid developers thanks to funding from the US government on the and if you look at the the progress conversations has made in the 2 years of existence of becomes clear that there aren't any unsellable technical problems and you can use a good of an extension of what of onyx assisting extensions work that some possible introduce you the the the the so it's not really about missing feature so on what it what about problems might they be the
1 so some say and the
it's it's the fragmentation with an open standards you always have on guidelines only implementing just
minus subset the I the the the 2nd the in the the the well in mining using open standards on there's no nothing stopping you from creating a very basic client that is Billy capable of sending messages and cause of Nixon pp client such clients will always exist and is again some popularity all the role of there's a chance to put Aksum pp into a better light the the however for the developers who are actually interested in creating a good state of the art client seller compliance were the as it had held a newcomer navigate the joint ovaries extensions so my text engines like C and compliance to 20 16 points the developer to only a handful of extensions and basically Fe implements those and you'll find so the when we the class the all start talking about fragmentation on the server side the following we have to make a distinction between the southern plantations and the actual installations the the 3 the Major our still implementations on actually that fragmented on the they also share our more or less the same feature set but if you look at the excellent installations that are out there and get a different picture we often see this problem are with user study by was support channel and so on after the question why don't feature X work for me and for a long time the answer to that has always been on the it is just a server that doesn't support can address conversations couple kid perfectly fine of handling of this just so for the his server but this is still the major issue because the end user really doesn't care the the if it's a several a client but did they just want to work on the so why why governs those installations are don't offers the future on most service are run by volunteers and it's safe to assume that they are not Billy holding makes features most of the our extensions don't put any Millwall load of pressure on the server and it's usually just a matter of enabling source extensions or simply not using z extremely outdated version from the data repository for example on the so how do you how do you fix this problem of well we 1st have to make the problem was about so a couple of months ago I the wrote a small tool that connects to your XMP piece over and simply checks for the future originally it was meant like some kind of a self-assessment tool we know that you can run on your own server and checks if you own so that's the so you operate support those extensions the but you can also all together more information from Russell service and and collect those thing from this information the and some kind of list on the so this is what I did
fine art by created accounts on various service and on In just at this
overview off from what extension is not enable on whatsoever so I made you this you get a lot of missing features which are marked in red answers but diagram but on a lot of service the so what do you do now how do you fix this how do you
convince a server operator to enable those who just well what I did was try asking them sir I spent an afternoon to to look up the contact information of a couple of a service and just something you know an answer have a bomb M. X. PPE involved a lot in the past and we now have extensions this needs to be unable server and this summer actually support this you just have to a bands the if you want to our better more convincing argument is an hour to give the system operator some incentive J. actually are enables us all trade unification if you own if your order this graph from by apply all supported features and not by not
a thematically the get of a nice little high school system and then we observe operator wants to be on top of the list right it on the so and those something we will every 1 of you can actually
help you don't need to have any programming skills just all good his list Picasso's from like going down there that doesn't support any of extensions and then sums on in all mention the much with our something and just ask them to enables those extensions FIL so what are the problems of the in I well
another argument so that is usually made again they simply P in particular more general on to other
standards are open protocols the is then it slows down your development process on but this is actually not true on it's very if you don't have to wait for all something to give standard to implement it you can implement something for and then try to stand of this size afterward I mean every half decent program I should write down what they're trying to do anyway and documented in in some way and if you modify that existing documentation you're already halfway To twentyfour legs extension and I mean the chances are that on during the process standardization on the protocol will involve a little bit and be adopted so are the people are gets the needs covered with this as well but usually the don't change in a way that makes the that makes it impossible to on I don't kill implementation on the we do what well and the and even if it's a little bit of extra travel on the the 2 2 stands as I those extension that's so Paul better than implementing and you an entire our I and political and a lot of extensions actually on came to be like this Google for example originally introduced all jingle z session so isation you need are to do boycott was of IP calls are video call the and this later evolved into the official standards into the official jingle extension the end the original Google Talk line also had an extension school to Q which are more limit at the battery consumption on of the client also generates when a background this extension was a few modifications of syntax all later became client state indication the on unfortunately this process can can become annoying if you're waiting for someone else to publish an extension on the primary example for this scenic story I xn pp always had extension for group chat codebook multi-user chat on and look looks a bit like I had the and has a very extensive permission model with on was moderators participants visitors and so on the and for some time now the abroad simply community on the agrees on the fact that book is it really suited for was those private group chat thought of what's there or hangouts and should be replaced with something else In fact last year during the Christian section of my talk when I was asked about on only 1 the group chat and my response was that you have to wait for makes the the on and it's at that time was in very early planning stages and without going into the into technical details illegally delete but the unit cell we still don't have a problem annotation well there are are a number of reasons for this and the number 1 reason this that makes us just a very boss extension the it's set out to cover the use cases for the next decade and a lot of people are are trying to get their use cases covered some but the more frustrating reason to some people was that makes was primarily written down by 1 or 2 people from the same company on why the requirements and the features that most makes a test on we are discussing a publicly and everyone could get their own wishes in basically on the actual on individual writing happened behind closed doors and and and the any offer to help out or up to speed up the process was rejected but so this less a lot of people who desperately needed a replacement for more uncertain about if and when it was going to happen the as a result we see we saw 2 different proposals of father the extensions that were basically variance of mock however it is rather unlikely at lenses there extensions will ever become standard because everyone else is still waiting for makes I'm so so what I did I was also waiting for makes the what I did the I should be we're at the books again and the came up with the solution to get my own communal group all covered physics this thing no I'm by on but yeah let's say using will in a way that no 1 else as before and stretching some of the definitions are in the standard around the but but it got to the point where it was actually a usable on with some limitations for example there's a primarily what's primarily on limitation of that on to to engage in a group chat you have to have everyone on in your dying out a personal contact list and but but this is still works fine on I mean especially for the loss small group chats nearly as disaster trading was your friends and have those individual countries in his anyway so I'm now at a point where on I can more patiently wait for makes to have because I I I don't know I no longer have to pressure her sexually media solution because already have a work around and and yes of course by using makes will make all those things more easy R and our the definitely more bulletproof the but it's yeah that's fine but somehow on and so just last week are the
people who wrote makes actually published a huge progress update to the standard the and we might actually now see 1st implementations of makes pretty soon on the so the if X and opinion a such a fine protocol on what is it used Malta the the well I am not our the 1 who makes the decisions that Google or Facebook or what that so I can only speculate on the reasons on the but yeah some companies for example Google go with the systems at some point actually should statements but unwise and not using some pp however those statements as every statement made by marketing department of course should be taken with a grain of salt so for example they are arguments from what was the systems for not using X pp effectually just incorrect I the so what are the reasons are there for a company not using can be p on but to answer that on let's take a step back for a and the look at the companies to actually do you think from pp from while they might not actively advertise this what's that ically doesn't sex pp the of course at times they introduce a couple of extensions and In the most prominent extends losses being are compression layer set optimizers X PPE is using on mobile devices which except that which but what that primely tries to do of but they decided not to publish publish incentivize extensions and before we going to speculate on the reasons on why it what that doesn't do have on let's look at some other companies or organizations that actually do use xm pp so for example later it uses a lot of x and 2 p and the and as 1 can imagine are government organizations and the army on general have some special requirements that I'm not already covered by existing extensions so what did they do did they the think Eriksson BP doesn't solve a problem then something not cross and then say the just came up with an extension on and this is why this non extension that all label of individual messages with security clearances and stuff like this because you do need for this in in our military environment the but this is a list of the question on what's the difference between nature and what's the well it's easy it's it's a revenue model NATO basically appears a company to develop an information solutions for them and there is a protocol standard his eyes the can only be beneficial to them because in theory after the concert is expires they can hire someone else to keep on developing since the missing solution it and so make sure as the user pays someone had to develop something In case of what's that the revenue model on it's entirely different I the user doesn't pay for anything and what's up is that the lights on the ideas inflated investment bubble and follow was what's awake argument that at some point when we get that's enough on metadata we can't sell very very targeted advertisement on that's during this arm inflating the business follow lies on girls on the generating more users every day In the federative world well clients and so are other interchangeable that's not possible I for example I I have no idea how many people using conversations because of course they do do have all the dollar numbers from home forms of the play start they don't have a donut numbers from effort for example on of from z the middle very few people but that's cells or I also the don't account for our own for the forks of commas that out there the so and by the way if you ever wondered why on but open with the system so so reluctant to have the signal at on F-Droid but is 1 reason because I mean everybody uses Google Play that abetted were overview of how many users into have the the so if you want to change all that the we would have to start paying for the service of or the softer directly well I remember when I said implementing on the Morris on is boring and on motivating work and to also implementation that out there that are being developed are they actually being worked on by all paid developers I what if we stopped relying on volunteers to develop our stuff down for us instead I'll pay someone what to do it like pay everyone else in our society and you don't even need to have a large user-base to make this work let's take conversations for example it which admittedly is so world smaller and so product compared to others this arrow news conversations would have bought on the place there for about 2 bucks the I would actually be able to make a decent living on this for i also terms you know a fairly small user base like 1 from conversations would be able to pay for our full-time developers yeah and well I guess if you fine with a few large companies are controlling our most important form of communication we don't have to do anything the so will continue to act in the interest of the Restil's and the current nest of incompatible in suppressing solutions will continue as there are consequences the model if you aren't satisfied with the current situation should stop relying on the work of if you volunteers and actually start paying for the softened services for use on a daily basis the thank you again there's not are will nothing actually on analysis on trigger a couple of days ago and well
it's been a very interesting year US INF introduced from newer last August on a with all the guys among your plugin cantilever December redress and important step because it actually allowed us to enjoy the a multi and aspect of I amino and I released for me more intuitive group
chats probable but the proof that smokers capable of handling and engendered for the group trends in March and we also sort of called some licensing issues all that stops the arch attitude developers from implementing only morphogens John and I ask the and there are actually working on that now and what I don't have an official ETA for our our only 1 iris I'm I'm I'm I'm I'm focus on that that will happen this year the I mean after all should security on I only made the huge progress this year and the implemented push-notification switch finally allowed us to the well reliably receive messages on on iris which was formerly impossible before on however 1 important thing before only was missing a and sets are independent security audit well it was missing until now and I'm very happy to announce that on can finally makes this are ordered public I've often sitting on the Star Audit for a couple of months now are actually and we had some arms this is sort hours with the company actually paid for this the and if you have a right upenn test on or before us is actually what we want to bet on and that's actually pretty nice read make a very competent people at a security company called on critically about security and which are based in on the Netherlands and that's was paid for by a very nice our company called the Pacific Research Alliance and it covers both a political itself as well as implementation in only by and but to quickly react reiterate on my earlier point this audit all we can to have because someone had a commercial interest in conversations and users wasn't paid for by OWL by donations or so wasn't done by someone wants you to do it the so yes and that's about it questions middle microphone I guess he said there were made 2 main developers of makes and that there were some of the some people there were some people who owns instead of you know waiting for makes so instead of the trying to get their own songs that their own ideas and and that the you developed of particles based on the on the kind of solution and it and so do use is there anything in the current of makes so constructed in the in the current makes the protocol that you think is not and may might be a problem in which might be because there it's only of produced by 2 developers from 1 company you who might have vested interests in the only doing on implementing a while the ideas and l I have sat on mix of the actual except itself like Lords entire takes account if on written down the 2 people and it was and had behind closed also you didn't have a T. A. for this but on the features that makes rail fully are discussed on on a public in a public place was there's a lot of different x indicates the they users from various companies so for example there are also called peace summit some it's twice a year are of the logic to be the community cattle odds of a conference and we actually discussed makes for a like entire day or some summonses 7 is so you don't think there are like any may bigger problems and makes which arise cool ideas which could have been included which weren't included because but it's so that the company of to develop which based retrieval office and my it's not want that to be included so you think it's a good an open process and the weeks protocol itself was to prove this is actually a very good good good good and no I I don't think so NO of and in any case once us it's public you could in theory of course gel and changes in if it were necessary in of course you can always if you no field you wouldn't necessarily have to fork it you just have to convince 0 Association pp standards on foundation to accept your change things the but there are very other Christians the can and thank you if you want to have stickers with me and I'm on 1 son just meeting at the at the cafe give thanks