I recently wrote about the newly published collection of Richard Feynman's letters. I had some time to read the book this week. The book has pretty well defined chronological slices, at least in the beginning. Almost all the letters I am reading now are to and form his wife Arlene, who was sick with TB during Feynaman's work on the Manhattan Project.

The letters reveal a pretty amazing love story, which makes this a highly unusual reading as pretty much all such powerful love stories I've heard of were made up. The story is also very tragic - I am currently reading letters from May 1945, remembering from other books that Arlene died in the late summer that year. This tragic feeling reached the apex for me when I read this from Richard's letter to Arlene:

"The doc came around special to tell me of a mold growth, streptomycin, which really seems to cure TB in guniea pigs - it has been tried on humans - feir results except it is very dangerous as it plugs up kidneys or something - and some have nearly been killed by it. He thinks they may soon lick that - and it it works it will become available rapidly..."

They licked it and streptomycin did become available rapidly, ending centuries of the TB plague, but just a couple of years too late for his wife. Somehow I am sure he thought of that.

(Curious side note: I actually indirectly owe my existence to the bugger. My grandmother was a TB specialist for the Russian army during WWII, and my grandfather... - can you guess?)

On a mostly unrelated topic. I was curious to see how the web reacted the new book. Feynman is a cult figure and I expected this to be a fairily noticable 'event'. Combining this curiosity and my obsession with del.icio.us I decided to do something about this, and came up with this script. Unfortunately the functionality I needed, getting all the posts by a given tag, is not *really* available in the delicious.py API. The function get_posts_by_tag is actually there (in the aptly named DeliciousNOTAPI API), but is scrapes the RSS feed, which does not give you an interesting enough time window for active tags. So I have to roll my own. Thanks to the truly beautiful BeautifulSoup this turned out to be reasonable. I actually think adapting the results to the delicious.py "post" format should be trivial, but I am all out of candles :).

The results can be seen here, with the last spike being very close to the book release and the attendant PR campaign.


Unfortunately due to, I assume, a missing piece of matplotlib trickery the results are useful, but leave a lot to be desired - for the life of me I could not get the X axis to stretch. I could really use some advise here.

To run you need a working copy of matplotlib with WXAgg backend and also BeautifulSoup.

Also since del.icio.us is still working on some performance issues, please use with discretion or risk getting throttled.