Saturday, November 17, 2007

Randomised trials

A simple explanation about the need for randomized trials in testing medicine, critical of the exemption homeopathic medicine seems to expect from this. Also, how some homeopaths have reacted to criticism.

Has mainstream Western society lost a belief in the objective to try and gain back the romantic?

(PS. This is not say that "big pharma" are not capable of running their own scams, but that fairly conducted randomized trials are the way to test medicinal claims, regardless of who is making the claims.)

Friday, November 9, 2007

Quote on integers

Leopold Kronecker
God created the integers, all else is the work of man.

Tuesday, September 4, 2007

Teaching by passwords

Overcoming Bias's Eliezer Yudkowsky has a painfully accurate description of how science can quickly become a game of passwords with no true understanding. This should be required reading for all school teachers.

Monday, August 20, 2007

SciVee a PLoS initiative

It will be interesting to see if PLoS's newest initiative, SciVee, will be a success. It allows researchers to upload a 10 minute video presentation about their publications (although at the moment it only allows open-access publications to be described, hopefully that will change soon).

I was pleasantly surprised to find my university has already gotten in on the act.

Friday, August 17, 2007

Presenting the design flaw...

Have you ever had the experience of going to a presentation where the presenters laptop/projector did something dumb like showing a screensaver, turning off etc. Actually let's turn that around, can you recall the last time you saw a presentation where presenter's computer DIDN'T do something dumb? I've seen (and have been the presenter on occasion) when all 3 major OSes do something stupid, like suspending, showing the screensaver, installing updates and trying to reboot or deciding that now seems like a great time to reindex the hard-drive (and don't get me started on the voodoo you see people have to go through to make their in-slide movies play consistently).

How hard can it be??? Surely, this is not a difficult problem, to have the ability for presentation software to tell the computer "I'm doing something really important right now, don't do anything else!" It's a sad comment on the current state-of-art in software engineering that this still remains and outstanding problem.

Will somebody please fix it (and it will have be one of the "other" OSes because Microsoft considers unreliability a feature)?

Monday, August 13, 2007

Ten simple rules...

Philip Bourne (and friends) has been writing a serious of 10 simple rules for young researchers in PLOS Computational Biology which I've found helpful:

Sunday, August 12, 2007

The science of gender

Over the weekend I had the chance to watch a highly interesting debate between professor's Stephen Pinker and Elizabeth Spelke. The debate, held in 2005, was part of Harvard's Mind/Brain behavior initiative.

You can watch it here (requires Real player)

Saturday, August 4, 2007

Microsoft knows best.

David Berlind caught Vista in the act of installing updates and then automatically shutting down with no ability for the user to pause the process.

This actually happened to me the other day, but on XP, the laptop in use for a presentation caught a wiff of the wireless network, grabbed itself an update off the net, and then the presentation just froze while Windows decided to install the update. No warning, no way to pause the process and regain control of the computer. We had to get another laptop before we could resume.

And to think this company manages to sell server software?

Thursday, January 25, 2007

Pectin Builder

As part of work with BJPS, I've made pectin builder - accessing through the web or using the Python scripts (tar.gz or browse svn). Accepts a simple 1D sequence string (G or E's) and builds a 3D PDB file.

These are just simple scripts that use Sweet II to do all the real work - cool stuff over in Germany.

PHP for accessing GData authentication/blogger

Well, I didn't find the documentation that great - but after a bit of mucking around and some help from the mailing list you should find here simple PHP scripts for using Google authentication and posting to blogger.

Wednesday, January 17, 2007

Quick, Dirty ODS reader

I discovered, to my great sadness, that reading an Microsoft XML spreadsheet in Python appeared to be easier the Open Document equivalent. There are plans to make a fully-functional Open Document library, but they seem to be minimal, and a few other libraries but nothing did what I want.

This is my attempt at a simple ODS parser, which attempts to read in the numerical values (as opposed to say =) which I use now, so that I can store experimental data in ODS instead of CSV files.

Much tidier.