As I’ve written about before, I use a combination of Vodafone 3G and the AT&T Network Client on Windows to access my corporate VPN. Recently, I’ve been seeing this error a lot when ‘dialling’ the 3G network (I connect it via the Windows dialler, rather than the Vodafone client):

This morning, I figured out what it was that was causing the problem. AT&T was open when I was dialling, and had the modem selected in the ‘Cellular’ tab (which I don’t normally use):

I hypothesise that it’s got a lock on the modem when this is selected (even if the ‘Cellular’ tab is not selected). Quitting the AT&T client appears to allow the connection to be created. You can then restart it and connect to the VPN in the normal way (using ‘existing internet connection’).

Hope this helps someone.

I’ve written before about how poor the Vodafone 3G Mobile Connect client is, and the alternative in Windows. However, it’s also worth mentioning that the card itself is far from perfect. It is very sensitive to the position of the SIM Card – having it off by even a few millimetres can mean it doesn’t connect, and it’s not always obvious what is wrong. The crucial thing is to have it poking out of the top, but only by half a millimetre or so. Don’t push it in all the way! This photo shows what it should look like:

It’s a shame Vodafone themselves don’t bother making this clear. This wasted about an hour of my time when  I first got the card.

Been to see two films in almost as few days recently. Slumdog Millionaire was entertaining; perhaps not the classic it could have been – I think it lacked a little clarity in storytelling – but still great. Frost/Nixon was more compelling, certainly for me: I was not expecting Frank Langella’s performance as Nixon to be as good as it was – in the end, it perhaps beat Anthony Hopkins’s classic performance from Nixon. Watch out for Oliver Platt’s wonderful performance-within-a-performance imitation of Nixon.

If you’re anything like me, you hate the Vodafone Mobile Connect client - it’s buggy, unstable, and requires frequent reboots/restarts/taking-the-card-out-and-putting-it-back-in-again to make it work. What I only discovered today – although I should have known from ancient dial-up modem connections years ago – was that you can just use the regular ‘Network Connections’ facility of Windows. Just locate the ‘Vodafone Mobile Connect’ entry, right-click and select ‘Connect’ (I think this will only appear once you’ve installed the Vodafone software):

Easy-peasy connectivity. You can still track how many bytes have been uploaded/downloaded in the current session by turning on ‘Show icon in notification area when connected’ from the Connection’s Properties dialog. You’ll then see an icon in the system tray with a pop-up showing usage. Sadly, this is only for the current session, not the month as a whole – haven’t yet found a decent bit of software to do that (surprisingly).

(This post is probably of most interest to my IBM colleagues – however, since the AT&T Network Client is also available for other organisations, I’ve posted this here in case it’s of use to other folks.)

I recently installed Ubuntu 64bit (amd64) on my Lenovo Thinkpad T61. Unfortunately, the Network Client is not available in a 32bit edition. So I had to perform a few tweaks to get it to work. Here’s the steps I took:

• Install the ia32-libs package and all it’s dependencies:

sudo apt-get install ia32-libs

• Install the AT&T client itself (IBM colleagues can obtain this from the OCDC website):

sudo dpkg -i --force-architecture agnclient_1.0~2.0.1.3000-3_i386.deb

• Add some symlinks:

sudo ln -s /usr/lib32/libssl.so.0.9.8 /usr/lib32/libssl.so.4
sudo ln -s /usr/lib32/libcrypto.so.0.9.8 /usr/lib32/libcrypto.so.4

• You should now be able to start the Network Client from the Internet section of the Gnome menu, or from /opt/agns/bin/agnclient. Click Connect and enter your details as normal.

After yet another grumble at jt, I’ve finally broken and will give Twitter a try (my ID is andrewferrier). I’m still struggling to see where the benefit will come from, but he made the valid point that I can’t say for sure ’til I try it – so I’ve given in and will give it a go. To be fair, I’ve come across two bits of useful info. already, so early signs are promising.

My concerns are:

• It’ll suck time. I don’t think that can be avoided – it’s just a question of whether there’s enough benefit there to make it worthwhile.
• I haven’t yet found an interface that works for me. twitter.com is too much of a stream-of-consciousness, and not rich enough to show clear threads of discussion or filter stuff out. I’m avoiding desktop clients as a matter of habit these days. Integrating the RSS feed into my Google Reader stream will just overwhelm me. This could be a deal-breaker; I’m already wondering whether Twitter actually lends itself to a inbox, read-everything model or if I should chill and let stuff slip past unread (something that sends prickles down my spine).
• Twitter’s help pages suck. Still haven’t found the number to SMS updates to in the UK (OK, I haven’t tried that hard).

Early this year I wrote about how the French were standing up to anti-smoking legislation. The Dutch are now doing the same. I wish the British didn’t roll over so easily.

Forcing pub owners (or anyone else) to enforce your preferences is wrong and a thoroughly illiberal idea. It makes the world more homogenous and less interesting. Don’t stand for it.

I don’t normally do plugs on here, but my good friend Ella has just recently completed her first novel: Blood, Sex and Coffee. It’s a mammoth achievement, weighing in at a good 700 pages, and is a campy vampire novel bringing in more themes than you’d think could be fit in such as a size – although is fundamentally a huge tract of story crossing space and time. I haven’t yet had a chance to read it – and I’m not normally a big fiction reader – but I have a 12 hour flight to Cape Town at the weekend, so maybe be burning through some of it then. I hope to post a review sometime soon.

You can download it for free or buy a paper copy for £13, either of them via Lulu.

For some time, I’ve had a Perl script that runs regularly, backing up my Google Reader subscriptions using the standard OPML format:

#!/usr/bin/perl
#
# Usage:
#  backup-google-reader-opml file-to-write-to.opml google.user.name@domain google-password
 
use strict;
use warnings;
 
use WWW::Mechanize;
 
my $mech = WWW::Mechanize->new();
 
$mech->get("http://reader.google.com")
    or die "Cannot reach Google Reader Homepage";
 
$mech->submit_form(
    form_number => 1,
    fields =>
    {
        Email => $ARGV[1],
        Passwd => $ARGV[2]
    }
)
    or die "Cannot submit form";
 
$mech->get("http://www.google.com/reader/subscriptions/export");
$mech->save_content($ARGV[0]);

However, I recently wrote another script (this time Python) that then takes this OPML, parses out all the URLs that are tagged with ‘podcast’, and outputs a serverlist file for podget (an automated console-based podcast downloader). This enables me to subscribe to a podcast in Google Reader, and have the podcast automatically added to the download list. The script looks like this:

#!/usr/bin/python
#
# Pass in the OPML file as the first command-line parameter. Will output the
# podget serverlist on stdout.
 
import re
import sys
import xml.dom.minidom
 
doc = xml.dom.minidom.parse(sys.argv[1])
 
body = doc.getElementsByTagName("body")[0]
 
p = re.compile('^\W+')
 
for outline in doc.getElementsByTagName("outline"):
    if outline.getAttribute("text") == "podcast":
        for subOutlines in outline.getElementsByTagName("outline"):
            title = subOutlines.getAttribute("title")
            title = p.sub("", title)
            print subOutlines.getAttribute("xmlUrl") + " NoCategory " + title

Feel free to use and adapt to your needs.

I’ve had my TripIt Personal Calendar Feed specified as an iCal feed into Dopplr for some time, to keep the two in sync by adding every trip in TripIt into Dopplr automatically (I’d give up on Dopplr entirely, as it seems a bit overdesigned and buggy – but it is a bit shinier than TripIt and produces nicer stats, so I’m hanging on to using it against my better judgement). But unfortunately the folks at TripIt broke this integration recently by adding exact geographic locations to the feed. This caused Dopplr to be over-zealous and make the destinations too specific (regions of Madrid and Paris, for example – rather than the whole city).

So I wrote a Yahoo Pipe that fixes up the stream – by removing the exact geographic locations and any items that describe activities and not the trip. Feel free to use it if it’s helpful – the single parameter for the pipe is the URL of your TripIt Personal Calendar Feed, and you’ll want to grab the URL of the pipe as an iCal feed to specify in your Dopplr account.