Well, I got my patched Orinoco driver working, and I also got Kismet working and so it outputs what it finds by speaking to me with the Festival speech synthesis system. It is pretty damn good – I still have no idea how to use it to its fullest potential, but it is fun to play with. I can’t wait to get the GPS now.

Speaking of GPS, I was browsing around the net and stumbled across a project called OpenStreetmap. This project seeks to create a completely free and open set of street maps that will solve the problems that Europeans face with non-free and non-editable maps. This is not so much of a problem for our American friends as they have access to a bunch of free maps that can be downloaded, but even then, these maps are not free as in speech. Rather annoyingly, one of the chaps behind the project talked about this at EuroFoo when I was there, but I evidently missed it. Damn.

I find this project interesting for a number of reasons:

• We need free maps – firstly, we need free street maps. We need maps that can be loaded into free software that can be used to indicate where we are and plan routes.
• We need maps that have open data – many of the current maps we have access to are simply huge images with little useful data. With an open map, you ideally want to have each place and location stored in some form of open and accessible data format that can then be used in other applications. If you made this data available with web services for example, a stack of useful information could be available online.
• We need maps that represent current data – if I go onto the M6 and there are queues between J6 and J8, I should really see this when I look at the map. Likewise, when I am driving around a town, I should ideally know where the car parks are and even how much the fees are. If you take it much further and to a point where we have countrywide wireless net access, you should be able to click on something and access current information or a website about the resource.

Now, this is all a pipedream in my head after seeing the site, and to be honest, a lot of this is way off yet. I think that the key in making this work is by creating tools that are so easy to use and so accessible, anyone can contribute to the project. If this project involves people downloading complex software, taking a class in GPS positioning and spending hours of their life writing repetitive data, the project will never catch on – in any case, very few people have GPS units anyway. I think the key is in requiring as little information from general users as possible and putting the software in the hands of everyone (including those with no GPS units) and everything (apart from courgettes, they are shit). A few ideas:

• Make it simple – the GPS software really needs to automatically indicate when information is needed from the user. As an example, if the user is driving along a road and takes a sharp left turn, this could possibly indicate that a new street has started. In this case, the startpoint right through to the endpoint should be logged and flagged up as needing a street name. Generating the road map should not be a problem as you can just turn on the receiver and it will track wherever you are driving to, but these changes in direction should automatically log that a new place name is needed. This then gives those with a knowledge of the local area and no GPS unit the opportunity to go online and add the place name in some kind of map based Wiki.
• Resources should be defined and then placed – it would be useful if Joe Bloggs goes online, sees a couple of street names in his area and then knows there is a car park near by. He then has a stab at locating the car park by adding it to the map by clicking in the general direction, but it is flagged as potentially not accurate. When someone with a GPS unit is roaming in the area, the contributor could then drive into the car park, select it from the map Wiki and add the coordinates.
• Voice recognition would be useful – obviously voice recognition is not necessarily as accurate as we would like, but it could be used to get a quick approximation of a street that could be fixed later. This would allow roamers to add data hands free – a key requirement in the UK where handling any kind of device while driving is an offence.
• Make available capturing software on everything – it is no use if the only way to capture GPS data for OpenStreetmap is by having a laptop and GPS unit. There needs to be capturing software available for new phones with GPS facilities, PDAs, actual GPS receivers and more. This will make it more accessible to most people – if someone can sit on the bus to work and contribute to the project without looking like a dork, this would be useful.

I think the main point is that GPS roamers and web contributors will need to have different interfaces that are separated in usability and ease to facilitate these different users. Web contributors should not need to know about long/lat coordinates – they can probably add place names due to their general knowledge of the area If the system is easy to use, I am sure that walkers, ramblers, hikers, explorers and others could potentially contribute to the project, particularly if they can add meta-data that is useful to their hobby too (such as ramblers rests, points of interest etc). Imagine the potential – you could even indicate wireless hotspots that people are increasingly leaving open to allow you to log on when needed. Cool.

Oh, and last night I spent a few hours recording a new tune. It is called Given Time and is available, as usual, in the Music section. It is dedicated to my girlfriend, Sooz. 🙂

This week has pretty much reflected exactly what being a freelancer is all about. I have had late nights, early mornings, a variety of different things to work on, challenges speaking to vendors who can’t their arses from their elbows, and I did the whole lot from the comfort of my dressing gown – all day, every day. My trousers are probably wondering what happened…

I finally placed an order for a GPS unit, and I plumped for a Rayming TripNav TN-200. It seems like a pretty solid GPS unit, has built in support in 2.4.x and later, and it has got a magnet so you can stick it to the roof of your car when driving. It should be delivered here while I am at the expo next week. The magnet thing sounds excellent – I wonder if I can get a shirring police light that will stick to the roof too. Maybe not.

I am planning on using the GPS with both Kismet and GPSDrive, but I was informed that the mapping situation for GPS units is a bit dire in Europe. Over in the USA, there are sites where you can download maps, and GPSDrive allows you to download these maps within the program if you want. You can set up maps manually if needed, and I am going to look into this. There is another GPS tool written by a guy in the UK called GPSS and it is free software (as in beer) for Windows. I mailed him to discuss about making it available for Linux or possibly contributing maps to GPSDrive. If all else fails, I can make use of the 1.4GB of NASA maps – I am really after UK street maps though.

What I find especially cool is that you can hook GPSDrive and Kismet into Festival for spoken output of what the programs are doing. I find this just so damn cool that I have to play with it. People are doing some pretty cool things with technology, and I am looking forward to blogging from different parts of the world and showing a map of where I am. Yes, this may seem sad, but I think it is cool. Some other bloggers have done similar things, and people are also encoding photos with long/lat coordinates. Someone at EuroFoo did this but I lost the link. Damn.

Oh, and thanks for the kind emails from Donika and Michael. It is always nice when you get some nice emails like that. 🙂

I stayed up pretty late last night trying to patch the Orinoco driver so it can slip into monitoring mode. It turns out that the version in 2.6.8 is 0.13a and version 0.15 has the interesting monitoring code in it. Anyway, dragorn from Kismet hacked together a patch for the version in the kernel that not only implements the code but also fixes a less than standard frame layout that the original feature uses. Apparently this non-standard frame is causing it to be a pain for some userspace software such as Kismet, and despite constant emails, the original authors of the Orinoco driver are not doing anything about it. Ho hum.

Anyway, I downloaded the kernel source on my Powerbook, patched the kernel and compiled the modules. I tried to insmod the module separately, but it was having none of it. I decided to install the entire kernel as a seperate kernel on the system, waited until 4am for it compile while chatting to Aq in Gaim and finally restarted the Powerbook. In a fate that I have experienced a few times before on a Powerbook, the fecking thing did not even make a humble attempt to work. I finished at about 4.30, frustated and tired. I hate yaboot.

I got up today and had a chat with Colin Watson in #ubuntu where he explained how I could fix it. I have now restored my previously working kernel and still have no Orinoco module working. I will play with it later to get it to work. In a related note, I want to get a GPS unit that will hook up easily to Linux and work – many seem to still come with Serial as opposed to USB connectors. Does anyone have any suggestions about which one to get?

Part of this research has been for something I am writing and also t help get the monitor modepatch into Ubuntu. Check the bug report page for the latest.

Kudos to sparkes by the way for his GTK based system updater. With a bit of eye candy it could be a truly useful piece of software, particularly if merged into GNOME.

Finally, my sympathies go out to the family of Hans Bakker who died in a car accident, Hans was in the accident with Edwin Hermans, Sebastian S. and rumoured to be RMS. It turns out RMS was not involved in the accident and Edwin and Sebastian are fine. A sad day for free software.

Got up early today and started writing up some SuperDisc pages for PC Plus magazine. This kind of work is always interesting because I often get to see a broad subset of quality from different vendors. This is no Microsoft, Adobe, Macromedia or other big name software, but lesser known niche applications, one of which (AliceStreet) that uses some Open Source components to provide a 3D virtual meeting room. The software is a pretty cool idea, and maybe this kind of concept could work well within Project Looking Glass.

Today I also looked at getting Kismet running. I wanted to get it going on my Powerbook with its rfmon supported Airport card, but for some strange reason it seems that the Orinoco driver in the Linux kernel does not come with the ability to switch on rfmon (essential for sniffing networks). I started to download the driver source and look to patch the kernel and then I thought “Ubuntu should really support this – I am gonna file a bug”. It turns our Daniel has beaten me to it and already submitted a bug report. Fingers crossed this should be fixed sometime soon. There is a fix out there (the feature is actually in the driver, it just needs turning on). Also, the bug is being handled by Herbert Xu who handled my previously submitted bug to include the ADM8211 wireless driver. Herbert managed to implement the fix in a remarkably quick period. 🙂

Sparkes dropped in last night for an hour and we had a bit of a natter which was cool. When I first met sparkes when we started Uni, we used to chat about all kinds of tech stuff for hours on end. Last night, it was mainly revolved around how good Ubuntu is doing and the impressive responsiveness of the hackers. We also discussed LUG Radio and some other bits and bobs such as my idea for a dialog abstraction layer. One thing I said to sparkes which I stand by is that if I had some disposable money to chuck at a developer to write something, it would be this dialog layer. If KDE/GNOME applications used the native file/directory open, save/save as and print dialogs as well as unifying the theme and possibly icons, the desktop would look so much more integrated. Sure, you can’t really unify which way around OK/Cancel buttons are, but I think it would add tremendous value to the desktop. Anyone have any disposable cash to chuck at a hacker to write this? Or, more specifically, anyone want to write this for nothing more than a kind email from your appreciative users?

Recently, I have been reading Alan Cooper’s The Inmates Are Running the Asylum, and it is quite evident that although Cooper has worked in software and usability for 25 years, he has a bit of a hatred towards the familiar box of circuitry that most of us use. probably built up from his 25 year long sentence. Admittedly, I am not that far through the book, but it seems to take on a pretty pessimistic angle towards programmers. Cooper seems to think that all programmers are simply code workhorses who have no idea of interaction design or usability. You know, about four years ago I would have agreed with this concept, but in recent times there has been an evolution of the programmer, particularly in Open Source circles.

Take for example hackers like Joe Shaw, Jeff Waugh, Robert Love, Mike Meeks, David Faure, Edd Dumbill and others. All of these guys have shown great ability in not only constructing large chunks of code in their sleep, but also taking this code and applying a usable and considered interface to it. From reading many of the hacker blogs, it is evident that a lot of these guys are well versed in usability engineering principles. I am pretty confident that the importance in Open Source usability has encouraged many hackers to read up on the subject to get at least a cursory knowledge of how to create reasonable interactions. Yeah, there are still some abysmal interfaces hacked together by the inmates, but I think the tide is changing. The work done by these highly public (in the Open Source community, that is) figures is no doubt inspiring the new breed of hackers.

Incidentally, I am currently hating computers right now. As I type this into Ubuntu on my Powerbook, I am going through the grisly installation of Windows 2000 again. What a long and drawn out process this is with endless reboot cycles. I would rather chew on my own testicles that install Windows 2000. It seems that another person has a real seething hatred of Windows – he/she posted this comment to my Dumbing down the decision blog entry:

Who is stupid?
2004-09-27 20:58:46  ClueGiver

Well of course it can be determined in two ways:

1. If you are George W. Bush or like him, YOU ARE STUPID.
2. If you are a Windows user, regardless of excuse, YOU ARE STUPID.

Neither of these conditions affords one any right to sympathy.

Both ensure the reduction of humanity to insipid, whining dullards. 

I wonder if ClueGiver would spontaneously combust if he/she saw George W. Bush playing Solitaire in Windows XP in the oval office with Darl McBride and Bill Gates clapping at regular intervals. 😛
—–

Well, I just wrote an article about some things that were knocking around in my head for a while. The result of my key tapping has been shoved onto my O’Reilly Blog in the form of Dumbing down the decision. Anyone who is familiar with UK loans adverts will probably see some truth in what I am saying.

What a busy time it has been recently. I spent much of last week out in meetings with people. I submitted some content over to Linux Format and Linux User & Developer as well as getting Ubuntu onto every machine I can find. I suggest you don’t leave your laptop lying around in my presence – you may find something new on it when you get back. Those folks at Canonical have done a great job. It really is quite cool. 🙂

I spent some time adding a load of blogs to my Bloglines account, and you can see which ones I have subscribed to at the bottom of the front page on jonobacon.com. I am looking towards using RSS as my primary feed to read news as well as blogs.
—–

Things seem to be going pretty good at the moment – got some articles sent off and I am finishing up some other bits and bobs. I went over to Birmingham yesterday to meet with Exposed Magazine to start a new column with them on the local metal scene. I submit my first effort tomorrow. 🙂

Ubuntu is currently rocking the Bacon household and it is pretty damn impressive. I have submitted another couple of bug reports that have been met with enthusiasm from the good bunch behind the distribution, and I am pleased to see that the ADM8211 driver will be added to the next release. I am also pleased to hear from Colin Watson that the six monthly time releases will keep the repository open while development happens – this means I can incrementally upgrade software I care about and still keep in with the regular six month way of doing things for the big bits. This is a wise move and I hope the mainline Debian hackers are keeping their swirly eyes on this. 🙂

The iRiver goodness continues to spread and I have been adding more tunes to the shiny little black box. When I first got it, I uploaded some MP3’s and didn’t get round to building the database as it is created with a windows based application and I had not got round to compiling iRipDB for my Powerbook. Anyway, I eventually got round to compiling iRipDB, and the iRiver has taken on a whole new level of ease of use in terms of picking tracks and creating playlists. My hat goes off to Miguel A. Arévalo for writing this handy little tool. 🙂

Aside from my joy at listening to Deicide and Cannibal Corpse when out on the road, I have encountered a few issues in ripping tracks for the shiny box of delight. Firstly, I can’t rip to ogg format as the iRiver does not understand any tags within ogg files and it buggers up the internal database. I was tempted to sacrifice the database, but the iRiver really is a pain to use without it – you select a track and it has to flick through each track name in the album before it plays it – this is a bind when in random shuffle mode. The other problem was that no MP3 encoder (read lame) is included with Ubuntu due to understandable legal issues. I was using the Sound Juicer to rip to ogg, and I when tried to rip to MP3, it complained about the lack of an MP3 encoder. I went and got lame from the usual place and then Sound Juicer complained that it didn’t have some form of gstreamer plugin. At this point I became impatient and carefully aimed a large brick at my monitor. I then put the brick down and installed the faithful RipperX instead. This application did not need any gstreamer fluff I thought, it should be fine.

I tried to rip a track and it created a file called:

The Forsaken - A time to die.mp3

Not good, I want the track in the format of (number) – (band) – (song).mp3 – this means that my tracks are then in the right order of the album. I fired up the Settings dialog and filled in the cryptic combination of symbols to get the right format, hit the Go!! button, and the same format came back. Damn, I filled in the symbols in the WAV section of the dialog – not exactly intuitive, but nevermind, so I tried again in the MP3 section (as Spolsky says, users never read dialogs, and I am a great example of this). I fired it off and the same format came back. Why on earth was it not working…I bent over to pick up the computer room brick and then I flicked through the dialog box and it seemed there was a tab called Files where I needed to specify the format. What on earth was that MP3 section for? I have no idea. Shocking user interface design at its best.

As you may have noticed over the last few blog entries, I am really getting into usability mode. I am noticing silly usability problems that have really simple solutions. I still believe predictive usability is a flawed concept, but common sense usability is very much an art of practise that can be mastered. This just goes to show that what was intuitive to the hacker who wrote the great RipperX was not entirely intuitive to me. I am not saying what I say is the right way, but when there is a problem with a user interactive with a tool, there must be some form of issue. If I am the only person who has experienced this problem, it is probably just an anomaly on my part (I don’t believe any application can be 100% usable by everyone), but I would love to see how many people experience similar issues in Open Source software.

First of all, thanks must go out to everyone who passed on their birthday wishes. Thanks so much, I am really grateful. I had a bit of a birthday bash on the Friday night, and I managed to put away a remarkable amount of lager. It was a great night and it took most of Saturday to recover in front of the TV.

This weekend gave me the opportunity to give Ubuntu a decent try. I first installed it on my Powerbook; a device that has been giving me earache with my X server. With stock Debian, the process of getting it to play on a Powerbook involved the normal Debian install, configuring your kernel to support specific features, downloading special packages to get X working and finally trying to sort out the oddness that is the Mac keyboard and the infuriating single button mouse.

When I installed Ubuntu, most of these problems went away. There was no frustrating row of artifacts when you log out of X (which always prompted a reboot), and all of the hardware works great. Another nice point is that I can put the Powerbook to sleep. All of these benefits are nestled within the wonderful world of GNOME 2.8 and the other pretty new software included with the distribution. Everything else is available in the ‘universe’.

Although this process has been enjoyable (so much so that I just installed it on my main PC), there has been a downside to using Ubuntu. In total, I have experienced four bugs:

• Powerbook – right mouse button key combination (usually Ctrl-Click) does not work
• Powerbook – the function key symbols do not work on the keyboard – as such, no # key.
• PC – ADM8211 driver not included
• PC – Issues with sound – the driver seems to be loaded, but I get sound errors

Anyway, I have these bugs, so what is the right thing to do? Naturally, I should file a bug. I headed over to https://bugzilla.ubuntu.com and started to file the bugs. I then realised that Bugzilla is possibly the most horrific user interface I have ever used. Every bug I filed needed at least two attempts to get it to work. Here are the problems:

• First of all, the international language of the-image-in-the-top-left-of-a-web-page-is-a-link-to-the-homepage has not been adhered too. Why not?
• There is no clear way to just get a list of bugs. I don’t want to search for bugs – I just want a list.
• When you add a bug, there is a text box where you can type in the package that the bug applies to. As you type the package name, a nifty little drop down combo box appears with a list of matches. You would expect the down arrow to be able to select an item in the box – oh no – the box frustratingly disappears. Also, you can’t actually click on the scroll bar to scroll down in the box. A broken UI is not good.
• When you select a package, it automatically fills an ‘assigned’ box with the email address of the person who looks after the bug. This is great, but if for some reason it fails to put an email address in there (and just puts ‘unassigned’), the bug report will not submit. When you check the help to see what to put in there, it says – This is the person in charge of resolving the bug. Every time this field changes, the status changes to NEW to make it easy to see which new bugs have appeared on a person's list. The default status for queries is set to NEW, ASSIGNED and REOPENED. When searching for bugs that have been resolved or verified, remember to set the status field appropriately. I read this and figured that I could add NEW to the box. No, that failed to work.
• Just to top this off, when you make a mistake and you submit the report, it will indicate where the mistake is. You press the back button and it loses everything you typed in. At this point you stare at the screen in disbelief and consider methods of inserting bricks into your monitor.

How on earth can regular users be expected to do the right thing and submit bug reports if this ordeal has to be encountered. I just don’t understand why two key bits of software have horrific UIs – bug reporting systems and wikis. Strange.

Yesterday I headed down to London for the Voluntary & Community Sector Open Source Meze. It was a great day and I had the opportunity to meet some really interesting people, many of which I have spoken to over email before. I think Dan, Adrian and the rest of the team did a great job putting together a day that can hopefully mark the start of an annual event. The day was certainly highly attended – far more showed up than I was expecting. Good stuff.

I am just downloading Mandrake Linux 10.0 to do base a review on and I am also uploading a WAV file to a mastering company – my connection is a little maxed out at the moment to say the least. I am quite looking forward to giving Mandrake a run – it has been a while since I have used it and I am sure there are some interesting improvements in there. It is a shame apt is not included though (no, not apt-rpm, real apt 😉 ).

The Improving Linux Driver Installation article is still getting comments made to it, and there have been over 20,000 people read it. I have also had some more interesting emails from people regarding the problem. One email mentioned the DKMS. Hmmm, I will have to look into that.

It seems the programme is up for the Blender Conference too. 🙂

I can’t believe I will be 25 tomorrow. Crazy.