worthydolt's blog

Don’t make me register

worthydolt — Fri, 19 Aug 2011 11:30:40 +0000

I recently attempted to buy some panniers for my wife’s bicycle. I already have some Ortlieb Back Roller Classics and am very happy with them so she decided that she’d like more of the same. So far so good.

So, I go to Google and find the best price. There are two shops offering them at the same price. I choose the one that’s highest in the list and add the panniers to my basket. Then I hit ‘checkout’. Oh dear.
A form appears. A vertical scrollbar appears. It’s a big one. I look for the Paypal button. Nope.

So, I go back to Google Shopping and go for the other result at the same price. They have Google Checkout. Well done, Bike Republic! You win.

When I look at the list of passwords I’m already managing with the excellent KeePassX I certainly don’t need any more accounts with some random website I’ll probably never use again.

Don’t make me register!

Ripping iPlayer programmes under Ubuntu

worthydolt — Sun, 28 Feb 2010 21:18:29 +0000

I was using Ubuntu 9.10 Karmic Koala when I wrote this.

I installed flvstreamer and get-iplayer using APT. You may need to add software sources as I have universe and multiverse enabled. I then had a go at figuring out how get-iplayer works. Of course, it takes loads of options but passing -h, identifying what looked likely, and a couple of failed attempts got me a copy of the Saturday play as broadcast on Radio 4.

By viewing the source of the player popup I found the programme’s PID hidden in a JS block -

iplayer.semp.setVisualisation("http://www.bbc.co.uk/iplayer/playlist/b00qynvv");

I then passed this PID to get-iplayer thus:

get-iplayer --pid b00qynvv -o /media/entertainment/ --modes flashaudio1

I ended up with an MP3 named “Saturday_Play_-_Othello_b00qynvv_default.mp3″ in /media/entertainment.

I’m going to play with get-iplayer some more to see if I can stream the whole of In Our Time but I’m mightily impressed with how simple-to-use and effective it is.

MP3 player heaven and hell

worthydolt — Sun, 26 Jul 2009 22:38:20 +0000

I’m a big music lover. I have hundreds of CDs and more music playing gadgets than you can shake a stick at. But I’ve never had an iPod. I’ve had my fingers burnt by shiny things before (virtually unused MiniDisc player anyone?) so my first ever MP3 player cost less than £10, had no display and played your music in the order you’d uploaded it. I used it loads and proved to myself it was worth buying a decent one.

On a recent trip to China I was wooed by the traders of the Beijing Pearl Market. In a frenzy of bartering I failed to notice that the 8GB iPod Nano I was buying was actually a GCSE Design Technology project in a pretty case. Having 8GB of music on a device that was equipped with software that only allowed access to 5 albums made me really yearn for a proper MP3 player.

Saturday was a nice day. So I cycled the 16 miles to Goole Sports Centre then swam 1km (0.62 miles sounds so unimpressive). Such endeavours always leave me flushed with feelgood, and I was in this spirit – and thus susceptible to spending money – when I dropped into Argos before coming home. The Creative Zen Mozaic has some good reviews and in 16GB flavour would hold plenty bangin’ choons. I found myself with £45 on my Nectar points card so only had to hand over £35 to those nice Argos people.

It’s early days yet but initial impressions are impressive. It’s tiny, pretty and sound quality is decent with the included earphones. I think it’s even better with some Sony Ericsson earbuds.

Creative is negligent in its lack of support for Linux. The management app is only available for Windows XP and Vista. I don’t have anything in the house that runs these OSs but you can easily drag ‘n’ drop into the Zen’s music directory. But what’s this? Albums that were ripped using Windows Media Player are properly categorised by album and artist. Albums ripped under Ubuntu using SoundJuicer are all under Unknown Album by Unknown Artist. Yet inspecting their tags using Kid3 shows they have ID3v1 and ID3v2 tags.

Then I spotted the problem. When SoundJuicer has ripped albums, it applied ID3v2.4 tags. It seems that the Mozaic (and it’s not alone in this) doesn’t read ID3v2.4 tags. Nope, I didn’t realise there were so many tag versions, either. So I loaded my entire MP3 collection into Kid3 and chose the Tools>Convert ID3v2.4 to ID3v2.3 option and lo! when I deleted all the ‘Unknown’ tracks and uploaded them again, the Mozaic began displaying the correct artist and album information. Now to figure out how to get SoundJuicer to rip them with the right version of the ID3 tags…

What! Another Java MVC web framework?

worthydolt — Fri, 27 Mar 2009 15:03:09 +0000

I’ve been enjoying myself writing Java code lately. This is pretty refreshing. The cause of all this sweetness and light is Groovy on Grails.
I followed a couple of tutorials on IBM’s developerWorks site then got stuck in creating my own webapp. In the space of an afternoon I had an app to manage motorbike services. I can create service items like spark plugs, oil and filters or consumables like tyres all with costs. I can also record the odometer mileage when they were fitted.
Validation is scant and there is a number of things I could do to improve it. But compared with a traditional Java webapp, which I’d have needed an afternoon just to setup this was pretty good going.
Whilst perusing the documentation, of which there is plenty, unlike some other web MVC frameworks, I got the feeling that this was a mature, capable framework which has benefited from a lot of thought.
I intend to document briefly some of the things which I’ve discovered.
Here’s a starter:
A metaclass allows you to replace an object in your unit tests. I don’t think it’s as sophisticated as a mock object. I don’t think you can create expectations of methods being called and verify that indeed they were. But to allow me to get a taste of writing Grails code and tests this technique was fine.
I was trying to test a simple tag-lib method:

def thisYear = {attrs, body -> out << Calendar.getInstance().get(Calendar.YEAR) }

in a unit test thus:

void testThisYear() { def output = new DateTagLib().thisYear(null,null) String expected = Calendar.getInstance().get(Calendar.YEAR) assertTrue "years don't match", expected.equals(output.toString()) }

The test was failing because the use of the property out in the tag lib couldn’t be resolved. Presumably it’s an implicit property.
The solution was to use the metaClass technique to replace out in the code under test. This is done by adding one line to the setup method provided for you in tests generated under Grails 1.1.
DateTagLib.metaClass.out = new StringWriter()

Now, when I run grails test-app -unit I get my unit tests run and passing. Hurrah!

A Good Read

worthydolt — Mon, 09 Mar 2009 15:16:57 +0000

We don’t have a telly. We haven’t had a telly for over 4 years. “So what do you do in the gap between shovelling down your microwaved dinner and bedtime?” I hear you ask. Well, a variety of things. But on this occasion I want to hold forth about reading.
I don’t think we, as a nation, read enough. I think that’s why our literacy is pretty poor. As a person who spends a lot of time at a computer, I spend a corresponding lot of time reading people’s outpourings on forums and blogs. I admit that sometimes people just bang something out, for example on Twitter, and they may be keen to convey a message without worrying about the accuracy of their grammar.
That’s absolutely fine. As long as people know the difference between “they’re”, “there”, and “their” I don’t mind a bit. As long as they understand when to use “you’re” and when to use “your”, I’m unperturbed. But a number of people don’t.
Gail Trimble, the Oxford University student, came to prominence because she performed so brilliantly on University Challenge (apparently – clearly I couldn’t have seen her performance), but has been derided for something that is beneath the pale in this great country. She is intelligent.
It’s quite alright for footballers to use their natural talents to become massively rich and famous but woe betide anyone who has the audacity to be clever.
I’m currently reading The Suspicions of Mr Whicher by Kate Summerscale. It’s a joy when such a book comes into your life. During the time that you’re reading it, you feel a little twang of excitement when you think of being reunited with it after work. You feel your life is ever-so-slightly enriched by it and I’ve never felt that way about any television programme.

A stitch in time…

worthydolt — Fri, 06 Mar 2009 14:42:15 +0000

Toss. Utter toss. I broke my MIG welder. I was working in garage doing a bit of metalworking to hack up some auxiliary light brackets for my motorbike. I’d been having all sorts of problems getting the wire feed to work. The feed wheel was spinning happily but no wire was feeding through the torch.

The manual says this can happen when your wire is rusty. Damn! My wire is rusty. This is what comes of keeping the welder in an unheated garage for 15 years. I use the welder so infrequently I’m in no danger of using up the 5Kg reel in my entire life.

So what do I do? I unroll some wire and clean it with wire wool. A very long-winded and time-consuming process. Still, no feed. So I make my final mistake. I increased the pressure of the pinch roller and kept increasing it. Still the wire wouldn’t feed. So I cleaned the wire some more. So I manually pushed and pulled the wire through the liner. Then I pulled the trigger on the torch. A large relay clicked but the wire feed motor didn’t run. Oh no! I can’t have burnt the motor out, surely? So off to a mig welding forum for advice. If I apply 12V DC to the motor it runs. Good. So it must be the PCB. No so good. A replacement costs £50. Bad.

So here I am waiting for a confirmation from Machine Mart that they’re quoting for the right version of the PCB with a job on the bike half-finished. If I don’t get the job finished soon, there’ll be no need for the lights until next back-end, when the days get shorter again.

This time I’m going to buy a new, clean, reel of wire. I’m going to keep it in the house when I’m not welding. And I’m not going to resort to brute force when I’m keen to get something finished. Not ’til the next time, anyway.

Torrid times

worthydolt — Fri, 06 Mar 2009 13:57:00 +0000

I have recently had a close brush with our much vaunted credit crunch of late. I was in a meeting with a couple of colleagues where we were discussing a Google app engine/Django app that we were about to begin. I was plenty looking forward to it as I’d been playing with GAE and Django and found them great fun.

A colleague steamed into the room and told us that an email foretelling redundancies had just arrived. When I got back to my desk I discovered that I was one of those selected for redundancy.

Well, a week of turmoil ensued. I was subjected to uncomfortable meetings during which I was given updates on my employer’s efforts to find me work elsewhere in the company.

Finally, at the end of the week’s consultation process, I rolled up at work feeling pretty good. We have a bit of moeny laid by for such an eventuality. The weather was getting better and I would be able to spend plenty of time outdoors cycling and gardening. The house needs painting too. And I can devote time to learning Python and Django etc etc. Well, when I was called into my final dismissal meeting I was informed that some work had been found for me and that I wouldn’t be made redundant after all. I was relieved that my salary was going to keep on coming in but very disappointed to have been jerked around for the whole week. The well-documented feelings of anger, rejection, depression, worry etc. were all present.

Perhaps, before an HR person is legally permitted to deal with making staff redundant they should have to have been made redundant themselves…

Upgrading Google app engine’s Django

worthydolt — Tue, 24 Feb 2009 10:46:40 +0000

Google app engine is great for quickly knocking up simple web apps but the version of Django it comes with (0.96) is a bit out-of-date. I already had Django 1.0.2 installed on my machine and wanted to use it with app engine.

I watched a really interesting talk on YouTube by Guido van Rossum in which he explains, among many other things, how to upgrade the version of Django included with app engine. He also recommends using the Google App Engine Helper to remove some of the boilerplate code you need to write for each site. Unfortunately it didn’t turn out, in my case, to be quite as simple as Guido described it, at least on Windows XP. I haven’t tried it on my home Ubuntu box yet. A bit of reading led me to the following.

Say I want to create a website called mywebsite. I first create a directory called mywebsite to hold it.

Checkout the head of the app engine helper project.
svn checkout http://google-app-engine-django.googlecode.com/svn/trunk/ app-engine-helper

Copy the contents of the app-engine-helper directory into the mywebsite directory.

Change main.py, adding

if os.name == 'nt': os.unlink = lambda: None
immediately before the line InstallAppengineHelperForDjango()

On Windows, where you can’t symlink (AFAIK), copy the contents of the django directory found within the Django version installed on your machine into the mywebsite directory. In my case, the directory I copied was C:\Django-1.0.2-final\django.

Finally, update appengine_django/__init__.py in the app engine helper directory, overwriting its contents with the contents of the file here at Google Code

Once you’ve done all this you’ll be able to follow the remainder of the tutorial and get the ‘Hello World’ app running.

Addendum
I’ve recently been playing with Nose and NoseGAE. I couldn’t get any tests to work without setting a PYTHONPATH variable (in Windows XP). I set it to include the libraries mentioned in the instructions for using PyDev with Eclipse for app engine development. Then I came along to start a new app engine/Django project and manage.py runserver was failing with an import problem. Removing PYTHONPATH solved it. I still have some learning to do…

Free headlamp upgrade

worthydolt — Tue, 03 Feb 2009 22:17:54 +0000

I’ve long been underwhelmed by the lighting on my venerable old ’96 BMW R1100R. As a year-round biker, I really felt the need of some improvement when travelling home from work along some very dark country lanes.

I subscribe to the mailing list at www.ibmwr.org. It was there I first discovered what a difference replacing the wiring feeding the headlamp can make. Instead of relying on the stock wiring you use it to power relays (one for dip and one for full beam) and the power comes direct from the battery via some much heavier wire. Less resistance = less voltage drop and I ended up with around half-a-volt extra at the terminals. Doesn’t sound much but there isn’t a linear relationship between voltage and light output. I now see a dramatic improvement in the headlamp’s output though do expect to see bulb life reduced.

Thanks to Ebbo for the howto.

Of course, I’m still not happy so am planning further lighting upgrades. Watch this space!

Freedom of information too expensive?

worthydolt — Sat, 17 Jan 2009 23:19:29 +0000

I’m livid about the government’s latest shenanigans over freedom of information. At the same time as announcing proposals to build a third runway at Heathrow the government slipped out news that MPs’ expenses claims will be exempted from the Freedom of Information act. What are the national security implications of publishing MPs’ expenses claims?

Interestingly, this exemption was apparently lobbied for by both Labour and Conservative committees. A rare bit of cooperation and consensus among the infantile squabbling of parliamentary business.

I think this illustrates the utter contempt that MPs hold the electorate and taxpayers of this country in. They don’t belive they’re elected to serve their constituents, rather that they’ve latched on to a rather splendid gravy train and they don’t want it derailed.