 Before you can continue your session, I have a small announcement to make. The sessions after the lunch break about cell and cell art are free software by Enrico and auto building and experimental by Andreas and Martinus Faisalow will stretch their rooms. So if you would like to visit Enrico, go to the big audience T1 while the auto building and experimental session will be in the small room. So, thanks. Okay, so let's continue. The point I was beginning to talk about before the break was the whole topic of the layout and navigation and things like that. So what I would be interested in is should we try to develop a new layout? Should we try to make a competition out of this? I think some websites have good experience made with this in the last years. Let's take for example the Ubuntu page. I don't know if Mozilla was a competition or if this was something else. So, I would be interested in hearing your ideas about that. Yeah, I consider that the most important thing to say is form. My opinion is very important to say form follows functions and also follows content. So first I think we should really know what should this page content and information. Because you can't do a layout if you say, oh well, we're not sure what should happen with a page. Of course, a general layout also needs to be involved for the general layout. You might do it now and it might be a good idea, but really to say, okay, how the page should look like, especially for the front page, I think we need more than just a general layout. And that can only be done if you know what should be on it. Okay, so the point was, can we make a layout when we don't know what we want to change else on the side? And he said that you can only make a layout if you know the content. I disagree a bit. What you have to know for the layout is a bit about your navigation. So what do you need on navigational elements? So do you need only one list with some short links like we have on the top of the page? Or do you need a real menu like here at the left? That's things that have to do with the layout. For most content, I don't think that the frame in it is presented in really matters. The layout only has to make sure it doesn't waste too much space from the page so that the content is actually visible. But I don't think that it's one point of good layout that you don't have to change it for every page depending on the content. I just think you need to be sure how the main content looks like. The main content I'm being seen is like, okay, what's the navigation? What is the side structure? Should we move sub-projects or something else? These things need to be fixed before we make the layout. Of course, the layout is not a problem, but it tells the latest devs that this is the one or the other for work groups. We basically meet in the middle and say that some content is important, but the examples he mentioned were actually the navigation elements. I guess we some kind of agree in the end. You have to know what general links you have in your navigation structure and the real content doesn't matter. Okay, that's it. During the break, we've been discussing me and Fabian about, again, like user analysis as something which should become a bit more central. It's quite common in literature to see that if you find out that you have different kind of users, then you have to develop a different kind of interface and different kind of application to actually address them. I mean, so this user thing is so central that it should at least be part of the talking. And so we've been discussing a bit what could be the users of our website. And well, the developers we've been discussing before they're not addressed at the moment. There should be links to all the statistic page on people, NM people, and there's all the documentation around there. So they should be quite fine press. They look at the website, which is quite inappropriate for them at the moment. And the layout for press people, it would be absolutely the layout of the content. It would be absolutely different than whatever we are used to because you have to feed them the news, basically, which is short and to the point. And then after the news, all the pointers to the background, a little bit what DWM is doing actually. So DWM could be a source of news for the press, but could be expanded, I don't know. There could be some thinking about it. But we already have three web pages, three websites, sorry, or three parts of the website. And then there's Debian users. And we have to see if they fit there or not because Debian users is like anyone from a physical lab person doing quantum analysis to someone implementing Debian in a junior in a primary school, installing tux paint. And while we've been discussing if that isn't better left to the CDDs instead and maybe providing Debian stuff for developers or if the users could go in the press page. I mean, all this sort of discussion, if you think about it a moment, really change a lot what could come out of the website distribution, the website content. Okay, so he was making the point again that we really need to think about what users we have on our page and what do they seek to adapt our content to their needs. So we already discussed much about the developer aspect. But what he also mentioned was on the one hand the press, let's take someone from the press. He wants to have the news and he wants to have some concise information about the background, about general information about the project and things like that. And that's really difficult at the moment because the news are very split up and the background is on different pages. So that's an example for a user that's probably not self-good at the moment. And the last point are obviously the Debian users. So what do they want, but that very much differs because they are very different kind of users from system administrators who want to install large classes of Debian to people that only set up a small pool in their school or something like that or end users that only want to install their own computer. So the question is whether you can really make a page that gives them all what they want or if you have to guide them to some CDDs also that are more specialized for their needs and to make more emphasis on this. But yeah, so that was basically what Enrique was saying. And I guess the idea of trying to make such kind of analysis is really a good one. But how do we start with this? So do we make a discussion on the meaningless and say what are your users? Let's say something about this. How would you start with this? Let's try to make a discussion again. Should I answer? There is a raise hand down there. I don't know. Okay, let's make our first. I think the question is what should be on. One of the most difficult questions. One thing should be to take a look what the people are currently looking for. But it's only for the part of the picture because if it's not there, people will only look once for it. Because it isn't a work for me. So it's only half of the pictures. It's the other half. It's either to reconcile them, to speak with people and say okay, first to sort ourselves, okay, what could be useful? And then sort it, wait it in the end. Put some stuff in because we shouldn't have too many stuff on it. And perhaps also as some users you have seen have for example backed up for about the video saying okay, what do you expect? Because we know they are more interested to work with us than just average. But they are perhaps not as deep in as we are because everyone of us should be fairly easy to locate some information because we just know a lot about it. Okay, so the point was user analysis can consist of two parts. The one part is to analyze the current usage of the page and to see what users are using more and what they are using less so that you can find some information on which parts you should work because they are really important. But the other part is to find out what are people missing because they obviously can't use it but it's missing completely. So it's also important to get feedback from end users and trying to find a way to ask them. Yeah, well, in an ideal world you do all sorts of sociological stuff and interviews and which, okay, we can't. So we have to improvise. Either we do like some tiny fantastic thing like linking a small survey from every page and then see what's the profile of people answering from every page and so see what different kind of people are looking for. But that's, again, maybe more trouble than it's worth and there's a lot of risk of doing it wrong because we are not scientific knowledge researchers, basically. Or we can go really back driven. So suppose, well, we have some idea like press and M and D bells. I guess we have them. And so like start with something simple for them and maybe like keep the old website as like the everything else stuff. Like some like make some intro, like some, well, some homepage. There's no decent name. Like some personal page, some main page or main page for all these groups, for each of these groups. Starting simple and then see the request you get. Okay. And see not to always answer the request from the same person because that creates his own homepage instead of like the homepage. And then in that way, I mean, while doing that, you establish some kind of contact and you try to outline a picture of the person you have in mind which could be like even like an idealized character which you could describe and use as a reference. I don't know if people have read the inmates are running the asylum, which is a funny book as well. That's a book called... Didn't I send it acoustically? The inmates are running the asylum. The book is pretty nice and explains this technique a lot. It's a nice way of sharing the concept of the target user. And so like portraying this picture maybe as you go because we already have a website or like portraying an hypothetical one and discussing about that maybe like, oh, this guy makes the press release providing it and has little time for it because it's got tons of sources to track and you define it a bit in detail and you start designing for it. And then if that doesn't work out, then you try to find what's the shortcoming of that and you design another persona. That's the name of this thing and you refactor the website for this other persona and maybe that works better. That could be a process. So the main suggestion from Enrico was that what one could do was to create different starting pages for the different users we have in mind. So let's make a page if you're someone from the press and want to know something about Debian, go there. If you're a user, go there. And so that you have this as a starting page and link from there to the old content to the things we have right now and then to get some feedback about so that you have a starting point where users can comment on and give feedback about. Because other means of analyzing users like making surveys and stuff like that is much work and it's easy to do wrong if you're not experienced in it. So that's very problematic. And the other thing one could do is to make a discussion where you invent certain users and use cases you think of them and then try to go with them to the page and see where in their use case there are things that lack and problems so that you make different examples and try to analyze where they would run into problems. I don't think we should try to get people to feedback that probably won't work. I don't think people... Well, very few would leave feedback. Whenever I go to a website and see please leave feedback, I just ignore it. They pop up a window saying please fill out in our survey, I just close the window because I hate doing this. And a lot of people would probably do that. So we might get a few comments but we won't get... I think we'll get a representative view from trying to get user feedback. So I think it's better to try to figure things out to try to do the... try to think that use cases and build something and then maybe add something. You can comment here if it doesn't look good but you should try to... you should try to decide it before... I think. So the point was users are lazy usually so they don't give feedback and if you get feedback you get only a very selective view because only a few users provide it. So he thinks that the point about making hypothetical use cases and thinking them through is actually way better because that just works and doesn't require someone else to do work for you. So there was... Before doing use cases it's better to design personas which is like... I don't get the difference. No, well use case is what you offer them. Like the actions you offer them you support for them while the persona is who they are is them. And before designing use cases you want to design use case for whom. Because like getting news for a developer is very different than getting news for a press person. Okay, so the point was you don't have to create use cases but you have to create personas so that you have hypothetical persons that use the web page and to try to different kind of them like a press person that wants to make a general article about Vivian so what would he look for and thinking that through from their point of view and don't fall into the trap to just say all people want to have news because that's true but they want different kind of news so you have to think from their point of view and not only from the items you want to offer. Also looking at it from the other perspective from the people who are actually going to do this we probably need to adjust the process so that the people who are using it to do anything can actually make use of it. So what I'd like to ask is what kind of experience does this group have with this kind of development methodology? Do you all understand what a persona is and what a use case is or is this just mumbo jumbo that sounds fancy but doesn't really mean anything? So you are mainly asking the audience if they have already experienced with this kind of development so that you're trying to take the view of the user and try to develop from there or if that's really at the moment something that sounds sensible but you don't have any experience with. So let's take this point and then ask if someone wants to answer this. Yeah, just a re-question. Has anybody yet gone through the logs and tried to track down exactly how people use the web page? I mean it's good to think about hypothetical use cases and hypothetical personas but it would be ideal to figure out where people actually are entering the site from where they're going, what the commonly used information that we actually provide is. So the question was has anybody looked at the logs so has anyone tried to get information from them and the simple answer is I don't guess so. So, no. Sorry, www.devn.org is currently on Klecker which is a restricted host. Okay, so Paaske added that the logs for the main mirror for the www.devn.org not for the most are on Klecker currently which is a restricted host because it's also security.devn.org and so they are not really available but that's no problem. I can just give them to anybody who wants them and I even think that... I don't know, no normal website users don't have access. Okay, so the logs can maybe be available if there's a volunteer who wants to make some analysis out of the logs and he can even have other logs from other mirrors. We have some on deviant machines like www.us.devn.org which is on Glock so you can take a look at the logs there too. And so that's mainly about finding a volunteer who wants to do that. The technical stuff can be made possible. If you want to do that I suggest you get a friend who's in sociology and you feed them the whole logs and tell them to make a thesis out of it. That's the best way to do this stuff. But I don't want to wait for half a year for another site. Why? Okay, so the suggestion from Enrico was that one should try to find someone who is interested in sociology or something like that to help with that because it usually gives some results you miss otherwise. So the question was have people experienced the user-driven development? I don't know if that's the correct name or if it's the correct name for this kind of thing. Sounds fishy to you or... Put up your hand if you have ever started making a program without opening your next editor first in the writing program. Okay, that's a few. So... I guess what we would need for something like that would be someone collecting some useful resources about this. So let's say, okay, if you want to do this let's take a look at this web page or if you can get this book and read about it and then we can start with it so that we don't make every fault, every other people that have to try to do this before we also make again. So if you want to do this, you're welcome. I just sent to the web conf my English link about personas to show any article I just found on the web could be an interesting read. Okay, so Enrico will send some information he already has to the conf main list and I will forward this to the Debian website main list too. Yeah? I can also do some digging but I think it's important to get the doing started so we don't need to wait for everybody to get the PhD in user analysis before we do this. We can just start and fix things later once people have learned more. Okay, so he emphasizes that you don't have to read all stuff before you can start because it's just delayed it's not useful so you can start and then later when you know more just refine your results and fix everything you have made. In the meantime, if you want to get a PhD in user analysis this afternoon you can talk, that's the topic. Yeah, okay. I'm really interested in this we could even make it off at a later date at the conf, I don't know. I can just, that's it and if people are answering yes, I want that. I will write the domain list to the IC channel and see about feedback and I will usually request a spot if... Okay. So, other topics you want to talk about? Are there any? So before I just pick another one from my topics and ignore you? Well, I think we should do an only sort of start page but also there are a lot of how I should say, very depth licenses where you can go to for example if you are near the mirror list you don't need to go to a few steps away to get to the mirror list. It took me some time for asking me to find the actual mirror list. It should not happen actually. So, I think that's just an example and I don't have any idea how to do it but I think there needs to be a massive cleanup of the link pages. Yeah, okay. So the main point was that the links are really suboptimal at some places and one other example one user gave me last week, I think it was in Buckleyport or in the main was was about so you are getting Debian and... Yeah, the download link is also very good. Mega CDs, I guess. Oh no, that's not right. Okay, no. So, go on the download one. Download the next one. Oh yeah, see here it was. So you go to floppy disk because you, let's say you don't want to have the button I see. So you have another link. You have a page that is fully only footer and head up. There's nothing on it. Just a link to the install menu. The problem is there's no link so you have to go to the install menu let's say for something where there are actually floppies. So, now let's take the HTML. So, yes, now let's see let's see, it's actually made of floppies. Oh, another link. Floppies. Yeah. Yeah, okay, there's nothing. Let's try something else. Opting system installation media, let's try that one. So, maybe it's what we want. So, yes, that's a pretty much good example of what shouldn't happen. Go on a link that says get floppies. You want to get the floppies and not ten links to other pages where then we can actually get the floppies. So, that's a good example of something we want to fix. Other topics you want to discuss about? No, that's not actually the case. So, what we're still interesting about and we have left this topic I started because of the more interesting stuff we have then discussed. It's the whole layout question. I for myself, I can't make layouts. I'm really crap at this. So, how would you search for a new layout for the website? There are many people that have complained about this and so, would you just search for a volunteer? Would you make a competition? I would be interested in that because I can't do it. So, I have to find someone. How would you make this have someone experience with this? I think we have a set of people who make at least these layouts in the area of Debian and why shouldn't we just ask them if they want to provide a second side of my project and then we can of course discuss which looks better and who have good ideas about how the navigation should be and which of the sides the second step after fixing how the navigation should be. Okay. So, we just should just ask people but we need to have some information first to give them a better idea of what we are searching for. So, perhaps we should wait a little bit because we have really an idea about what do we want us to change before we search for a layout and then we later realize that doesn't fit us at all. Yeah. It would also be nice to get all pages that's more or less the same layout currently figured from here in the developer's corner at least the whole top navigation That's the left-hand side all that disappeared. Okay, good. We currently only have the extensive menu on the left-hand side only on the front page. So, the question was whether we want to make a more concise layout but the navigation elements don't change. Things like the CD page are also a question of course because it has a completely other layout currently and it would be interesting to see if we should merge this and I really think we shouldn't have a different layout for the CD pages but I don't currently want to impose on them the layout we have for the main page because it's crap so I invest the work into doing this and then better change it anyway. Okay, if there are no more things about that topic what I also wanted to hear is there are some people here that already work on the web pages some others that don't What I wanted to ask you is do you think it is particularly complicated to work on the web pages currently is there something we would need to change on this is to make it easier for contributors or something like that Split the users page Okay. One thing I find particularly hard but for other people if something goes wrong even if it's a main cloud file which will restrict the toast That's not true It's not obvious the documents that are available of course I myself can log into Black and look up and say, oh that's very good I think it might be better to document that from better I suspect you just have looked That's the thing it's hidden very well here on the web page Let's see if that's one good example for creating or editing pages There's a link here somewhere but it's obviously not on the page you would look for it The question was whether there are the build logs available since the page is regenerated with the VM error and you don't have the log files to look for problems especially if you change big things The point was that the build logs should be on the which is not true It's available via HTTP but the link is hidden It's right on the first where it's the first link has recommended reading next to it and it's a bit down at the how to help previous page previous page on the desk no, it's on the desk page the first link down to how to help how to help Ah ok, you have the build logs Yeah That was obvious Of course Ok We see we can find easily examples for what we don't want or what we want to fix So And in this link that's still not what I use most often are logs from the from the documentation sub-watcher and I don't think so That's not a real problem I'm just looking to click normally Yeah, that's how I find out about my rejects but on you But it's another question So Yeah We have only 5 minutes left so if there's nothing if there's not another topic which I think we can leave this here and just end this discussion I really want to follow up on the user on the idea of inventing personas and trying to fix some things and eager to read the links and we'll propose how to go with this Ok, so and I will try to summarize our discussion about this whole log-in stuff because we're not really getting to any decision or even a majority meaning or something that we should discuss this further There are obviously problems with CMS as the whole maintenance stuff and things like that so we should really discuss it and I will try to initiate this discussion on the Debian website maintenance so if you're interested in this point just go there and participate in this Ok, thanks I think it might be worse to make the use cases discussion because it's really not too hard to say everyone trust the whole use case we make a profit on Saturday or so and then we can summarize it up because if you're all here, it's much easier to discuss than by a main listener we have a salz if you believe that ok, so the proposal the proposal was to try to get the buff organized so that people that are interested in can start a little bit with this persona and try to get some first impressions and ideas about that I will propose a date do you have any no Saturday today is the formal dinner so that's not a good idea speaking of releasing everyone should just make one use case in about 2 or 3 minutes because then we have a lot of information yes there is an amazing find for a date we just got the weekend because of our time so they've come 5 at hot buck changes the changes link what did somebody do there? we have Thursday today it's already almost Sunday so we have Sunday no not really maybe we might hope we can just make a parallel to this ball from 8 to 9 and we just make that hot buck at the cabins place so your suggestion was to to make it a parallel to the start of the ball because the ball takes about 4 hours and we can just in the first hour parallel we can meet down the stairs down the stairs where also the Helsinki proposal was made so the suggestion was to make it a Saturday at 8pm because the linux ball is about 4 hours planned here so we could easily make an hour to discuss some of our stuff but of course also I don't know what's up with you but I don't like to keep it in parallel for example with web apps policy or with web apps policy or with the data and workflow management okay yeah I will write it on the mailing list and you can comment on this then um okay um yeah okay I like this suggestion to just make it at 8pm on Saturday is there anyone who has a problem with this okay so let's do that and who has to zoom go to stairs down stairs at 8pm stairs at 3pm and two bands and you just have to bring some chairs because you film also four people yeah okay I remember the photo Amaya posted it too okay so yeah okay let's stop this here and we hopefully meet some of you at Saturday some of you on the mailing list after this so thanks for the discussion