Monthly Archive: May 2007

Global warming denial at NASA

I heard NASA head Michael Griffin interviewed on NPR this morning. He was asked about global warming and after a bit of prodding said that while he knew it was happening and human caused he didn’t see it as a problem. He said that the current global temp might not be ‘optimal’ for humans and thought some warming might be a good thing. Classic second stage global warming denial.

This is a zombie argument–wrong and long proven to be wrong but it keeps popping up. A good site for rebuttal of global warming denial zombie arguments is at Gristmill. Griffin’s zombie is handled here.

Psst, want to hear an anecdote?

Every few months I run across an anecdote about how horrible Canada’s or UK’s national health system is (but never about any of the other European health systems, hmm). Anecdotes may cause you to consider an argument, but you need to turn to data to draw conclusions.

US/Canada health system comparisons:
Big study, clearly written:
And a 2007 Commonwealth Fund study:
Med. student review of the literature:

BTW, the Canada comparison is the one in which the US looks best, several of the European national health systems are *much better* and *much cheaper*. US citizens with top-of-the-line employer provided health plans get care as good as in the best European systems (just more expensive).

Conclusion: US has the worst and by far the most expensive health system among Western nations.

2008 Republican Presidential candidates

2008 Republican Presidential candidates
Evolution question from the first Republican presidential debate, torture question from the second debate. The others believe in evolution (at least to some degree) and did not answer the torture question at the second debate.

VentureBeat understands online journalism

BTW, the previous article comes by way of a link to an article in VentureBeat. Contentwise the article is nothing special, a thinly written industry conference talk/press release. But the author Mark Coker and VentureBeat understand online journalism. The article has what every web news story ought to have but few do–relevant and appropriate links. The article links to a detailed article on the technology, the company, and the conference. This is great!

Unfortunately most news organizations haven’t got this figured out yet. They will report on (summarize) say a report by the FDA without linking to the report. Many/most of these are now on the web–the reporter read it there in many cases–but the article doesn’t link to it. Or to the FDA press release.

Science news never links to the journal article. Or to the less technical journal News and Views summary, or to the non-technical University press release.

It a real opportunity for a news organization. If I knew the McClatchy (formerly Knight-Ridder) news service did this I would seek out their articles over AP’s or the New York Times. I haven’t seen any grab for this brass ring.

Better reading tech

A company has developed a new reading technology called Live Ink (paper here). The idea is to improve the way text is displayed to improve reading comprehension. The way the brain perceives a page of text is as small region at a time. Words from several lines get picked up at the same time. This confuses the parts of the brain that comprehend words and sentences. Live Ink proposed to spread text out to avoid this mental confusion to improve reading comprehension.

There’s a kernal of a good idea there–electronic text allows reformatting to increase comprehension. But their solution sucks. It takes up too much space and it only partially helps with the problem of line confusion. The syntactic breakup may be helpful. Here’s an image of it:
Live Ink example

Computer formatting is a good idea. Here is my idea of how to do it. Fuzz out the lines before and after the current line. This could be done one of two ways. If eye movement tracking is available, follow the eye and only make the line being looked at visible with the others fuzzed out. Without eye tracking make this a timed system (like traditional light bar reading trainers). Start it on a section of text and make one line at a time is visible moving through the text. Controls would be needed to pause and go back a few lines, and flip it off.

By fuzz out the rest of the page I mean alter it so the brain doesn’t think it is text and try to decode it. Whether making it out of focus is enough or if another kind of noise needs to be added would need to be determined by trials. It would be important to leave enough information that the presence of text blocks and formatting is visible. These provide visual landmarks needed for good reading.

Technically, this could be easy–a web browser extension or perhaps as simple as some Javascript added to a page with overlay images fuzzing the text and moving as needed.