Iran showing fastest scientific growth of any country | Hacker News Reader

Iran showing fastest scientific growth of any country

newscientist.com

31 pointsprat13 comments

"Fastest growing" is usually an annoying statistic. It's really meaningless. Say Germany produced 100,000 scientific studies last year, and this year produced 150,000 studies, they grew at 50%. If Iran produced 1,000 studies last year and this year produced 2,000, they now have a 100% growth, capturing the "fastest growing" title.

Indeed. The more interesting fact in this paper is how China is picking up with the US in number of papers.

If I'm reading graphs right, there were about 300k publications from the US in 2009 and about 10k from Iran, so, about 3% the output of the US, or about 14% adjusted per capita.

(This naturally doesn't argue your point -- I was just curious what the real numbers were.)

Measuring the amount of papers is questionable as well. One paper with a ridiculous breakthrough (cold fusion, etc) could outweigh all the others.

Having gone through graduate school, I've seen my share of nonsense papers created just to have a topic for a thesis.

Same here, one of the reasons I quit my phd program. I saw how much of the publication volume is bs--it's really hard and rare to produce a meaningful research paper, and that rarity will only increase.

If you go to a top cs conference, chances are 80% of the papers are throw-away...literally no one will read other than the reviewers, and I'm talking about top-conferences let alone the countless lesser known ones. And to be honest, a lot of emerging countries seem to be mass producing a lot of low quality papers in CS, and it's flooding the field. One reason might be how cheap it is to fund a cs research project as opposed to an experimental physics or bio lab.

same in medicine, they collect garbage from the Internet, make a 1000 page book and become specialist doctors with it.

http://xkcd.com/605/

Papers are about as good an indicator of productive output as lines of code.

As mentioned briefly in the article, citations are probably a much better metric.

that's too much oversimplification.. moreover there is, if any, a inverse correlation between lines of code and productive output. not so with papers,

Citations can be gamed as well unfortunately.

I find it unhelpful when a single questionable statistic is used to support a complicated thesis. In these cases, it would be better to listen to experts whose opinion is based on a significantly more complicated set of criteria. In that case we still have to judge the quality of the expert, but perhaps if that seems daunting as well then we should perhaps consider ourselves unable to judge the answer at all. I find that more appealing than walking away with a certainty of opinion on the subject of this article.

Iran can only accept any associated prize if it first demonstrates how such statistics are unscientific.

Reply on news.ycombinator.com