Friday, August 21, 2009

Healthcare reform and the scientific method.

One thing about any scientific measurement is that the signal always is convoluted with noise. One tries to hone in on the signal by verifying the data with repeated measurements. The more measurements you make, the better you can characterize the noise and have more confidence in the signal.

This echoes the rule of thumb from my news reporting days at my university newspaper - always confirm information with a second source. Better yet, find a second source *AND* a third source.

It seems like there is a lot of noise out there in the health care / health insurance reform debate. Some of this is normal "noise" associated with debating complex issues. Health insurance reform *IS* complex and should be debated. But more and more in recent days, it seems to me like a lot of this noise has been injected into the debate for politically-motivated reasons.

A lot of it is plainly dishonest and not true, and although I'm sure these people truly believe in their causes, the arguments they make are not honest. Please, the ends simply do not justify the means.

Everyone is entitled to their own opinions and this should be respected. What we need is honest debate. So please make sure these opinions are informed ones.

Read the reform legislation and follow the money. See how your senator / congressman have voted in the past. Seek out the truth and see how it affects you.

And please make your own judgments. Our democracy depends on you.

These sites might help:

Mendeley: social networks for researchers.

It seems that social networks are being incorporated into everything now and academic research seems to be no exception. I recently tried out the new Mendeley Research Networks beta and I like it quite a lot.

It is a journal papers repository, citation manager and academic research social network all rolled into one. There is a desktop client that you can install on your computer which is integrated to a social network and cloud-based services that you can access through any web browser.

The client program and the web-based services are free, and the company promises to offer new advanced features in the future for a fee.

What about Zotero?

The first time I tried Mendeley was quite a while ago, and although the idea of sharing reading lists and finding out what other people in your field are reading was appealing, I deemed the early beta of the desktop client too slow and buggy to be useful. Instead I opted to use the fantastic open-source Zotero as my main citation manager.

However, Zotero is not without its shortcomings.

Although Zotero is quite powerful in capturing citations as well as just about anything else on the internet you can see in your browser, there always was a certain lack of polish in the user interface that made day-to-day use something of a chore, especially for retrieving information.

When I wrote my PhD transfer thesis using Zotero and the accompanying Microsoft Word plug-in for inserting 300+ references, it got the job done but it was slow and painful.

Each reference insert or edit would take tens of seconds to finish, and let's just say that during this writing period the word Zotero often would be preceded by another, not so flattering utterance that rhymed with "luck."

Mendeley - but is it any good?

Fast forward to today - the latest beta of Mendeley has improved its performance  tremendously, although crashes still happen quite often - at least once an hour over last couple of days, by my estimation. Compared to the rock solid 1.0.10 release of Zotero, this is a step backwards for me.

However, the interface is such a pleasure to use. Being able to not just sort but filter by author or publication is so great and helps me to find articles more easily.

Also, simple things like alternating the background color between each record helps visually to identify where each reference begins and ends. Little gold stars for favorites and green dots for unread papers that you can toggle on or off really add to the usability of this product.

Of course, there always were workarounds that you could implement in Zotero to do this - I had "unread" and "important" tags in Zotero for doing the same thing, but these little interface bonuses are so much better.

Mendeley also has a feature to let the program organize papers / PDF files for you that simply is manna from heaven. It works in a way similar to the "let iTunes manage your library" option, renaming and organizing files as you change the reference details. Simply brilliant!

The current beta does, however, lose track of PDFs once in a while and sometimes even deletes them, which definitely is not cool. Also, sometimes it refuses to match up PDF's with already entered citations. But as the program still is in beta, quirks like this hopefully will be ironed out before the 1.0 release.

The killer feature though, in my opinion, is the inline PDF reader that allows you to highlight and annotate PDFs while tagging, taking notes and referring to referenced articles.

Previously, I did this in an open Acrobat window and switched back and forth between that and Zotero / Firefox, but having everything in one place is fantastic. Being able to open multiple documents at once, take notes and switch between them all from within Mendeley really helps to streamline my workflow.

However, the PDF reader function does seem to cause the program to crash a lot, but hopefully these bugs will be resolved soon.

Social research.

The other big aspect of Mendeley is the integration with the website, which provides online storage and access to your papers (500Mb per account), as well as access to aggregrate reading trends, shared reading lists, social tagging and collective citation editing of research papers. They describe it as for research papers.

Although the social network aspects are cool and interesting, it remains to be seen how useful this is to me.

The online storage of papers, however, is useful, although 500MB is on the low side for me. My current solution is not to have Mendeley sync my papers, but instead to put my PDF repository in my Dropbox where I have 3GB of online storage.

With this solution, only my laptop at home has the PDF's associated with the references, while the desktop at the University has the PDF's (via Dropbox) and references (via Mendeley) unlinked.

Hopefully, they will add a feature to link PDFs in a directory structure with citations in the Mendeley database - like how iTunes can relink tracks to mp3 files.

One interesting thing... Although papers written by others that you upload are accessible only to you, your own publications that you upload by default are downloadable by others, which makes for a clever and legal way to sidestep the giant journal publishing companies that tradtionally have served as the gatekeepers to all scientific information.

Hopefully this is used to improve general access to scientific information along the lines of PLoS and is not part of a hidden plan to monetize the site.

Overall, I really like Mendeley, enough to switch over from Zotero (although I will not be importing all of my citations over anytime soon - there is no easy migration path, yet).

It is a fantastic, streamlined, free (although currently buggy) solution for managing your papers and citations. The social network aspect also holds a lot of promise.

Thursday, August 20, 2009

Things are different when slow...

I uploaded a few slow motion films to YouTube. The clips were taken with a Vision Research Miro4 high speed video camera at frame rates of 1000+ fps and played back at a more sane 30 fps.

Things look quite different when slowed down, don't they?