Tom Morris

A blog post on the importance of keyboards. I concur! I’m reasonably happy with my pre-Aluminium Apple wireless keyboard, but I’m still waiting for something that’s RSI-reducing, has decent tactile response (despite being a child of the eighties, I used to own a typewriter and have a lot of push in these fingers of mine). A nice noisy clatter is fine.

I have a huge problem with most laptop keyboards - the one on my MacBook Pro is pretty good - I fat-finger the keys a little, and it’s plagued with first revision Apple defects (fuck you, Steve Jobs, I just want a computer that works) that have cost me tons in repairs that anywhere else than Apple would be covered by warranty. Most PC laptop keyboards I’ve used are a pile of utter shite. They keys are off just target enough to be unusable.

Despite people wishing for neural and body gesture interfaces, keyboards are still going to be with us in fifty years. I want a keyboard that doesn’t get dirty, doesn’t require cleaning, that I can drop off my desk and it’ll still work (hey, I could probably drop my old typewriter on the floor and it’d do more damage to the floor than the typewriter), is secure (ie. the Bluetooth connection between the computer and the keyboard is encrypted), has a reassuring clatter and is something that I can write hundreds of pages with comfortably.

I’ve been looking recently at guides on essay writing - since it’s something that I will both be doing a lot more of very soon, and something I will have to assess in others if everything goes to plan. My local library had Writing an Assignment by Pauline Smith (Amazon UK) in the rat nest that is their non-fiction collection.

It’s a very good guide - the sort I would give to someone trying to figure out how to get by in academic writing. Particularly worthwhile are the exhortations to know thyself - figuring out how one learns and manages time and then using this to reflect on how to do it better. It covers style and tone, although not quite with the level of detail I would have liked. The ‘case studies’ were ever so slightly grating, although not intolerably so.

Until schools teach this kind of basic level of competence to pupils, this book and others like it will be necessary. I guess I’m just a traditionalist fuddy-duddy in wanting schools to adequately prepare those in their care before sending them out into the world.

I’ve been having fun with the Fire Eagle integration on my blog - yesterday I posted my first real geotagged post (here, while I was at Senate House Library, University of London; my previous ‘Eagled posts were me pretending to be in Paris).

I have just been styling them. I just have to write a brief ode to CSSEdit - which makes editing CSS so simple for me (I have a terrible memory for CSS properties) - and for CSS itself. Love it.

Anyway, the design I was shooting for was actually based on datelines from news reports. At the beginning of news reports, you’ll often have “NEW YORK CITY” or “BASRA” or whatever. For short posts, I now have a little italic dateline before the post and a dash to separate it from the post itself. For longer posts, I decided that having the dateline at the end would be better, and justifying it to the right. I’m very much a design amateur, so I hope it’s not too tatty on the design front.

I’d like to write (or better yet, have someone else write for me) a J2ME MIDlet application that uses the Sony-Ericsson Java geolocation API to update Fire Eagle with my current location - either using the Sony-Ericsson proprietary API or through JSR-179. Now that would be cool.

I am also planning to add some extra functionality to my site soon using JavaScript and the sexy framework known as jQuery - specifically, new functionality related to the microformats and other embedded data - letting you lookup geolocations on maps, export KML and let you browse tags. All of this will be unobtrusive, will not affect you if you don’t have JavaScript turned on (for instance, if you use Noscript in Firefox to filter JavaScript and Flash on sites you don’t trust or opt-in on sites you do like).

What I am hoping will come out of things like this is that people will start treating their blogs as playgrounds to experiment with new forms of content, and new ways of linking that content together - using Semantic Web technologies and so on. We don’t have to wait for Google or a VC-backed startup to do something - we’ve got all the technology to do it for ourselves. I’m sure that’s a neat vision for personal publishing. And it’s a good way for me to learn jQuery.

I must say, I’m glad that there’s a negative reaction to TechCrunch being stirred up. It’s an utterly dull website, peddling uninteresting stories about Web 2.0. I unsubscribed about a year ago, and I don’t miss it at all. I was very pleased to see IsMikeArringtonADick.com surface on the Twittersphere yesterday. The problem isn’t actually Arrington, it’s the fact that he is the de facto King of the new Web, which means everything becomes about him, and he’s not a good gatekeeper. Add Ajax to a turd and he’ll rave about it, while really interesting stuff like Fire Eagle will just whiz past him. Similarly, Arrington is very much against SXSW, calling it a pathetic excuse for a conference. Funny that. Personally, I much prefer reading what the SXSW-attending crowd have to say than the Web 2.0 Expo crowd - mostly due to the fact that the good ideas and hard work that goes on behind the scenes are done by the sort of people who go to SXSW while the incestuous VCism goes on at Web 2.0, and the fact that designers and developers - the core SXSW market - tend to talk in plain English, call things by their real name and not bullshit, habits which the Web 2.0 Expo types could learn from.

Reporting on Web 2.0 and startups is barely even necessary anymore. I don’t need a reporter to tell me what Twitter or Facebook or Google are doing - they’re on the fucking Internet! They have blogs, and I can read. This is not Iraq - I can see for myself. If this is the decline and fall of TechCrunch, bring it on. As web 2.0 winds down, perhaps it’s time that TechCrunch winds down with it.

I now have FireEagle built into my blogging software. Pretty cool, no? I’m working on building it so that my blog posts will have my geographical location built in (if I’m not at home, that is). This is a sample post just to test the storage of this information. I’m not actually in Paris (although it’d be jolly nice if I was).

Almost works. Made a silly mistake (one of those not reading the documentation mistakes).

I’m still not in Paris, but I am pretending to my software to be so. If it all works, we should be able to get the website and the feed pushing out my geolocation in no time at all.

Today, I’ve been working on hooking up Fire Eagle to my blog. I’ve got it doing some fairly smart things. First of all, when I post on my blog, I now have a ‘Location’ field in the form so that I can bash in a new location. This is just simple HTML with a bit of PHP back-end to handle updating Fire Eagle with my new location. Fire Eagle is one of those fantastic things that gets better the more you use it. And having it in my blogging software means I end up using it all over the place.

But there’s more. When I post to my blog - whether it’s through the web-based interface, or through an XML-RPC client, my server will ask Fire Eagle where I am. If I’m not at home, it stores the name, latitude and longitude of where I am in the XML files that holds my blog content (SQL? Pfft!). This is then published on my blog in the geo microformat. It looks ugly at the moment, but this evening I’ll be tidying up the stylesheet. I’ll also be adding it to the RSS 2.0 feed I publish (yes, I use RSS 2.0 not Atom, because Atom requires that posts have titles - and, as Twitter proves, not all posts have titles) using GeoRSS.

Does your blog know where you are?

I’m all in favour of good user experience design. Just in case you haven’t heard, I also am all in favour of cute puppies, saving the planet and soft French cheeses. Hey, someone’s gotta grasp the really controverisal nettles.

But this morning, I got to the station in a state of barely disturbed slumber only to find that someone had redesigned the ticket vending machine. I have programmed the operation of the machine into my muscle memory - punch the ‘London Cheap Day Travelcard’ button (or Standard Day if it’s before 8.31am - don’t ask), then punch ‘Young Persons Railcard’, then put my card in, enter the number, wait a moment and then collect printed ticket and receipt.

Seems fairly easy. Only the geniuses who design these things decided that they should change all the menu layouts so that the simple task of buying a train ticket to the largest city in Europe should be made unnecessarily complex.

Now you have to choose from ‘Destinations’ or ‘Travelcards’ on the main screen. Well, I want to go to London, and I want a Travelcard. It pisses me off at a formal and a practical level, since you might say “I want to go to Brighton, but I’m not sure whether I need a Travelcard or not” - when in fact, Travelcards are only for people going to London. I chose Travelcards, and then I was presented with a list consisting of ‘All Zones’, and then each zone ticket from 1-6 and combinations of different zones. I bashed in Zone 1. No go. You can’t buy a ticket for anything other than All Zones - which makes sense. If you are travelling from outside of London, you are going through all the zones anyway and the price of going into London is rolled into the ticket. Again, don’t ask. But, why exactly is the machine offering Travelcards plural when there is actually only one option you can choose?

Having chosen the only option that works, it then presents me with a highly complex page listing all the different options I might want to add to my ticket. Hey, I haven’t got time for this! My train goes in two minutes! Just leave me alone and let me buy the ticket! I guess I should be thankful that it’s less complicated than booking a flight. Eventually, I manage to add my railcard to the ticket and get the thing to print with enough time to comically sprint across the bridge.

Here’s my vision of the same procedure. I get to the station, push my credit card into the machine, it then says “Good morning, Mr Morris. The usual, then? That’ll be 11.75 then. Enter your PIN. Done. Here’s yout ticket. We’ve e-mailed the reciept to the address we have on file. Enjoy the journey - and, by the way, the Circle Line is right royally fucked, so steer clear.” Okay, it’s slightly creepy and Big Brotherly - but it means I don’t have to think early in the morning.

At the very least, reduce the number of options. If you cannot actually buy a particular ticket, why offer it and then, when it’s chosen, tell the customer that it’s not actually available.

What’s particularly irritating about this is if I get a decision wrong, then I’m breaking the law. Surely, in designing a machine which increases the probability of me cocking up what should otherwise be a simple procedure, the designer of the system is aiding and abetting frustration and failure.

Techdirt reports that NBC Universal wants to force Apple to cripple iTunes so that it sucks so much that nobody will use it, just like all the other attempts to do DRM. Whoops, I meant to say to enforce digital piracy prevention. You know, because an MP3 you download off the Internet is very easy to distinguish from an MP3 you’ve ripped from a CD. This kind of shit makes me look forward to the inevitable failure of these dinosaurs.

I’ve been rewriting my static site in Webby, a Ruby static website management system. Hopefully, if it all works, you won’t notice. The only way you’ll notice is if you peek under the hood. I’m using “Options +MultiViews” in Apache to do URL rewriting, and there’ll be more about that later.

Using Webby is pretty easy: webby site-name makes a new folder. Inside, there’s a content/ folder where you can put your content. These are .txt files which contain some YAML to include the title, what filters it runs through (ERb and Textile for instance) and what layout it uses.

It comes with Rake tasks to build and deploy. Here’s how I’m using mine. The basic procedure is rake build and then rake deploy to SCP or rsync it to your server (I use SCP). I’ve introduced another step - rake validate.

What does the validate task do? Well, the validation I’m doing is using oNVDL, which is a Java tool to validate XML. I’m using James Clark’s XHTML modular schema, specifically xhtml-strict.rng. There is a downside with this approach, and one I’ll probably fix soon - that is, if you use a mixture of HTML 4 and XHTML, or you use different types of XHTML - either 1.0 and 1.1, or Strict and Transitional, or even the new modular DTDs that the W3C have put out for SVG and RDFa. The reason I used the RELAX NG schema is because I had it there and was in a hurry, I knew how to use oNVDL and I’d used the modular RNG schemas before.

The Rake task is quite simple:

desc 'validate XHTML'
task :validate do
  Dir.glob('output/**/*.html').each do |f|
    print "validating " + f + "\n"
    sh "java -jar /Applications/oxygen/lib/onvdl.jar /Users/tommorris/bin/xhtmlrng/xhtml-strict.rng " + f
    print "\n"
  end
end

This will iterate through your output folder (the place where built markup goes) and validates each one against the schema. It should also print the result to your screen. You’ll see nothing if it’s all good, but you’ll get errors like this otherwise: /Users/tommorris/code/HTML/test/output/profiles/nsfw.html:3:71: error: attribute “lang” not allowed at this point; ignored

The nice thing is that the Java process exits with status 1 if there are problems, meaning the rake task itself is aborted. You can then fix one error at a time. And you can also make your deployment (or repository commit) dependent on the whole site validating. Hey, it’s compile-time error checking for XHTML! I’m sure some people are now furious since HTML is Not A Programming Language, but for me, this is important. I want my computer to tell me when things go wrong, and prevent me from sending invalid markup out on to the web. I’m sure one could even replace the validator with a stricter one that does some basic accessibility tests - and it’d be neat if we could add similar tests for CSS and so on.

Once you validate all the markup as XHTML, you probably want to then get Apache to serve the files as application/xhtml+xml, but not do so when Internet Explorer comes knocking on your door, since Internet Explorer is so utterly shite that it can’t read XHTML properly. Adding the following to your .htaccess file does exactly that:

RewriteEngine on
AddType application/xhtml+xml html
RewriteCond %{HTTP_USER_AGENT} ((.*MSIE.*)|(Lynx.*))
RewriteRule .* - [T=text/html]

This will serve all your .html files as XHTML, unless Internet Explorer or Lynx is being used, in which case it should get text/html instead. XHTML 1.0 can be served as application/xhtml+xml or text/html - but if you are using XHTML 1.1, or maybe the custom XHTML+RDFa DTD, then the validator and specification says you should be serving it as application/xhtml+xml. It’s probably okay if you still allow people using retarded browsers like Internet Explorer get your XHTML as text/html though.

By the way, one last thing - I mentioned earlier in passing ‘repository commit’. If you are using Git to manage your Webby sites (and frankly, you should be, since Git is made of win), you should not check your output directory in. Just add output/* to .gitignore. I’d highly recommend keeping your static site in version control. It’s very useful to be able to roll your site back, or get a list of all the changes you made to the template or the CSS…

Rock Star Programmer has an excellent comparison piece showing the differences between Git, Mercurial and other distributed version control systems.

And if you want to use Git (and, hey, it’s cool and sexy and made of epic win), two posts worth reading: New Bamboo has a post about using Git and Github with Capistrano for Rails deployment, and Gitmagic is a nice detailed tutorial on using Git.

I’ve just been trying to help unbox and setup an HP Photosmart Pro B9180. It’s not detecting the ink cartridges, despite the ink cartridges being there. The support is almost totally useless. Here is what the booklet which accompanies the printer has to say:

Let HP’s award-winning support and services show you the way with assitance online or over the phone, 24 hours a day, 7 days a week. (emphasis mine)

The manual includes a telephone number - +448700104320. If you phone that number, you will be informed that technical support is available online through a live chat service on the HP website. The website does not allow you to get online chat support for the make of printer.

The website also lists telephone numbers which quite plainly aren’t +448700104320, and which are listed as being not open 24x7. Funny that, you’d think that the booklet that comes with the printer stating that you get 24x7 support would, you know, match up with things in reality. As Colbert put it, reality has a well-known liberal bias.

The lady who did answer on the aforementioned number told me that the technicians don’t work on the weekend. Jolly good. People need time off. But why exactly are you advertising 24x7 service in the documentation and then not providing it?

I’ve always thought printer manufacturers were a bunch of assholes out to take people’s money for tiny little cartridges of over-priced ink. Guess today proved me right again. Now the owner of this printer needs to call back on Monday, between 9am and 6pm - when at work. Interesting definition of 24x7.

I’ve got a LazyWeb request: I was wondering if there’s any possibility of there being a Firefox extension for setting up different browser use profiles (not the same profiles as in Firefox) - so, for instance, when I’m on a mobile connection I might be able to specify rules of what to download. When I’m on a mobile connection, I want to be able to use twitter.com but without it loading all the images of people. The idea is to reduce bandwidth usage so that I don’t hit the “fair use policy” of my provider.

The sort of rules I’m thinking of would be things like turning off JavaScript or Ajax, auto-redirecting to a mobile or low-bandwidth alternative, turning off large graphics downloads or not downloading sections of the page (XPath, CSS selectors etc) - or maybe have certain Greasemonkey scripts activate or deactivate.

Sorry to continue wasting your attention.

There’s no other way to test this…

I have recently been participating in the Citizendium, a new wiki encyclopedia that was set up to try and build an encyclopedia without some of the problems with Wikipedia. Citizendium is a project started by Larry Sanger who co-founded Wikipedia.

What’s different about Citizendium? Well, one of the primary differences is that you have to register in order to edit, and you have to register with your real name and provide a brief description of who you are, which is verified before you can participate. This typically takes 24 hours. Another difference is the split between three classes of user: authors, editors and constables. Authors are fairly obviously the people who edit the wiki.

Editors are people with a specific, high-level, real world expertise in a subject - usually signified by having a Ph.D in the subject (equivalent expertise is accepted for non-academic subjects). Editors can check over articles and then mark them as ‘edited’. This is basically freezing their contents - creating a stable branch which they validate. On those articles, work is then conducted on the ‘Draft’ branch, which the editor then moves over into the stable, public-facing branch periodically. This is to prevent you going to the wiki and looking up, oh, Immanuel Kant and getting back “Kant was a big willy-head. Yours, x_LaRouche_x!”

Constables are there to solve dust-ups in the Citizendium community. They serve much the same function as administrators, but they do not make any decisions about content. There are rules about separation of powers - constables can make decisions only about pages they aren’t involved in writing.

The community also has pretty strong rules about not engaging in personal attacks. The site encourages people to avoid writing “encyclopedese” - preferring well-constructed and non-boring prose.

It’s certainly a fun place to be. Yesterday we had the monthly Write-A-Thon, where people have an online party trying to crank out a lot of new pages in a day. I’ve started pages on Yochai Benkler, Humanism, Tim Berners-Lee, Capital punishment, Recieved Pronunciation and Prayer. Before the Write-A-Thon, I’ve also written pages on Continental philosophy, Unicode, Alternative medicine and more.

It’s tremendous fun contributing to Citizendium, and it’s something I’d encourage you to do. Request an account and join in. I’m trying to organise the Citizendium Local Library Storm - if you are in or near London and are interested, please do shout.

Ashley Highfield has blogged about a report arguing that the iPlayer has changed how broadband should be offered. The Telegraph says that television should be streamed rather than offered via download. Bullshit!

Let’s take a common use case - you want to watch a programme again. If you have a download sitting on your desktop, you can watch it as many times as you like. Okay, there are problems for the BBC with rights-holders, lawyers and other corporation bureaucracy. I don’t really give a shit. The user experience of downloaded files is a good one. You can watch them as much as you like, back them up to a DVD and keep watching them forever. Give me that over streaming anyday, where you have to download the same file over and over again every time you watch it.

I think that Trading Standards and other consumer regulators need to step in and slap the “unlimited” broadband peddlers with a giant wet fish. They are lying to their customers. If you shape P2P traffic, you are not providing an unlimited service, you are limiting the speed for one application. If you have a “fair use policy”, you are not providing an unlimited service, you are limiting it - but just doing it vaguely. I have membership of a library in London, which I pay an annual subscription for. This subscription entitles me to access the library at any time it’s open. Imagine the “fair use” version. You are allowed to come into the library, but if you use it more than an arbitrary and undefined amount each month, we will Have Words With You. Fuck that!

The ISPs have entered a difficult business, and are whining. They need to invest in infrastructure, rather than bitch to journalists that users are using the unlimited services which they have been solved. Until they significantly increase their infrastructure investment, everything the ISPs say is crocodile tears.

Don’t worry. The iPlayer will still be irrelevant in all this. Anyone who’s got more than three braincells will get their television through the Medium That Of Which We Cannot Dare Speak Of.