interesting stuff

Adverts that follow you

Now one of these days I will finally get my ‘moving flat’ epic published, in the mean time you can read my how to buy a property guide. It really is worth every penny.

Part of this process involves buying a sofa which doesn’t suck as much as my current sofa.

Of course, I’m doing this the proper way of procrastinating and constantly looking at websites and not deciding anything (look out for my exciting new book of the same name, the ultimate manager’s guide).

One of the sofa’s which was luckly enough to make the final rounds (i.e. got to perform in front of Simon Cowell and co), was this one, called – cutely – Oscar. Simple lines, modern look, sofa bed. From Furniture Village, which seems to be DFS’ more mature cousin, though this perhaps doesn’t say too much.

All fine and good.

Then the other day I visited Engadget.com, not a site I normally visit but a link had caught my eye on Twitter.



Engadget screenshot, with interesting add

Notice that ad? A bit like travelling to furthest Siberia, walking in to the dodgyist bar and finding your nan there, distinctly out of place. Click on the image for a lager version.

At the top of an American website about gadgets is an advert for a UK middle-of-the-road furniture store, advertising the exact sofa I had been considering for some time.

This was no coincidence.

Nor had they been using Alien technology to read my mind. I had visited the furniture site using the same laptop (and presumably same browser), a cookie and advertising system was at play here.

In fact my mind was made up when this evening I saw this.

This was on Time Magazine’s website (again a US publication), on a Photo Gallery about Afghan Women (see this for background, wonderful world).

It’s that cheeky little sofa again, this time popping up next to a repressive regime. You Guys!

Finally, notice on the red border of the ad, bottom right there is a little bulge, clicking on it…

http://www.struq.com/consumer-opt-out/ “Totally personalised display ads”.

Does this freak me out. Probably should do, but at the moment it borders on fun, like most people my tastes and wants are diverse enough to create stupid juxtapositions (Serious News and Girls Aloud, Global Warming and fast cars). It becomes an issue when it goes beyond, ‘person x has looked at product y from company z so show advert to it’,and becomes one entity building a database of everything you view and do online.

What a scary vision. Think I’ll stick with what I know and just use Google and Facebook.

Twitter clients

From about an hour after signing up to Twitter until very recently I used Twirl on both PC and Mac as my Twitter client. I was happy with it, and still am, but had noticed people using other clients and wanted to see if I was missing anything.

I round up my findings here:

Picture 1.png

Three twitter clients

(more…)

University league tables combined data

Last year I collected the University League tables published from various sources and combined them in to one spreadsheet.

I’ve been updating this for this year, i.e. tables published in early/mid 2009 aimed at those starting in 2010.

You can see the UK Combined University league table data Google spreadsheet here and re-order as you please. I’ve updated the three ‘UK only’ lists, and will update the international lists in the near future.

Notes:

  • This is a bit of fun, for my own interest. Don’t take it too seriously.
  • There are plenty of good guides to UK Higher Education including The Guardian, The Times and plenty more. Use them, not this site, if you are thinking of studying in the UK!
  • A quick glance will reveal that I have never studied statistics. In particular my made up scoring system is laughable.

You’ll find last years data in a seperate sheet, accessible via a tab at the bottom of the document. Are you able to produce anything interesting with this data?

PubSubHubbub instant RSS and Atom

I have just come across PubSubHubbub via Joss Winn’s Learning Lab blog at the University of Lincoln.

It’s a way for RSS/Atom feed consumers (feed readers etc) to be instantly updated and notified when an RSS is updated.

In a nutshell, the RSS publisher notifies a specific hub when it has a new item. The hub then notifies – instantly – any subscribers who have requested the hub to contact them when there is an update.

This is all automatic and requires no special setup by users. Once the feed producer has set up PubSubHubbub, and specified a particular, the RSS feed has an extra entry in the feed itself telling subscribing client systems that they can use a specific hub for this feed. Clients which do not understand this line will just ignore and carry on as normal.Those that are compatible with PubSubHubbub can then contact the hub and ask to be notified when there are updates.

It has been developed by Google, and they’ve implemented it in to various Google services such as Google Reader and Blogger. This should help give it momentum (which is also crucial for this sorts of things). In a video on Joss’ post (linked to above) the developers demonstrate posting an article and showing Google Reader instantly update the article count for that feed (in fact, before the blog software has even finished loading the page after the user has hit ‘publish’). It reminds be of the speed of Friendfeed, I will often see by friendfeed stream webpage update with my latest tweet before I see it sent from twirl.

I’ve installed a PubSubHubbub WordPress plugin for this blog. Let’s hope it takes off

UPDATE: I’ve just looked at the source of my feed ( http://www.nostuff.org/words/feed/ ) and saw the following line:

<atom:link rel="hub" href="http://pubsubhubbub.appspot.com"/>

Amazon AWS EC2 and vufind

Today I saw a tweet from juliancheal, which mentioned he was setting up his virtual server on slicehost. I hadn’t heard of this company but their offering looks interesting. This got me thinking about cloud hosting and I decided it was time to actually try out Amazon’s AWS EC2.This allows you to run a virtual server (or multiple servers) in the Amazon cloud, servers can be created and destroyed by a click of a button.

First thing is to get a server ‘image’ to run in the cloud. Thankfully many have already been created. I went for a recent ubuntu server by Eric Hammond. This is basically a ubuntu server vanilla install, but with a few tweaks to work nicely as a EC2 virtual server. Perfect!

Signing up is quick and easy, it just uses your Amazon (the shop) credentials. Once created, you are taken back to the main control panel where you can see your new instance, including details like the all important public DNS name.  Just save a private key to your computer and use it to ssh in to your new server.

e.g.: ssh -i key1.pem root@ec2-174-129-145-xx.compute.amazonaws.com

(you may need to chmod 400 the key file, but all this is documented)

Once in, well it’s a new server, what do you want to do with it?

I installed a LAMP stack (very easy in ubuntu: apt-get update and then tasksel install lamp-server). I initally couldn’t connect to apache (but could from the server itself using ‘telent localhost 80′). I presumed it was a ubuntu firewall issue, but it turned out you also control these things from the AWS control panel. The solution was to go to ‘security groups’ and modify the group I had created when setting things up and adding HTTP to ‘Allowed Connections’. This couldn’t of been easier. And then success, I could point my browser at the DNS name of the host and saw my test index page from the web server.

Amazon aws control panel, modify to allow http connections

So now what? I pondered this out loud via Twitter, and got this reply:

vufind-twitter

Excellent idea!

Good news: vufind has some good – and simple – documentation for installing on ubuntu;

http://vufind.org/wiki/installation_ubuntu

Following the instructions (and editing them as well, they specified an earlier release and lacked a couple of steps if you weren’t also reading the more general install instructions) I quickly had a vufind installation up and running. Took around 20-25 minutes in all.

Now to add some catalogue data to the installation. I grabbed a MARC file with some journal records from one of our servers at work and copied it across as a test (I copied it just by using a scp command logged in to my ec2 server). After running the import script I had the following:

vufind results.If the server is still running when you read this then you can access it here:

http://ec2-174-129-145-75.compute-1.amazonaws.com/vufind/

EC2 is charged by the hour, and while cheap, I can’t afford to leave it running for ever. :)

So, a successful evening. Mainly due to the ease of both Amazon EC2 and Vufind.

A final note that if you are interested in EC2 you may want to look at some notes made by Joss Winn as part of the jiscpress project: http://code.google.com/p/jiscpress/wiki/AmazonWebServices

Both Ec2 and vufind are worth further investigation.

1 person likes this post.