Thinking about thinking

So far today I have been wrest­ling with get­ting my dis­ser­ta­tion expe­ri­ment bund­led into a stan­da­lone app via py2app. Inc­lu­ding opencv seems to be trou­ble­some. Until I find a solu­tion to the opencv pro­blem, I’m going to com­ment it out and focus on spee­ding up the real-time res­ponse force feed­back on the dis­play machine.

Spent most of the day wor­king on coding my first dis­ser­ta­tion expe­ri­ment, which will use an xbox gamepad’s ana­log trig­gers to index res­ponse force. I came up with a cool way to pro­vide visual feed­back of res­ponse force as well as res­ponse type, but I need to do some opti­mi­za­tion of the code to help it avoid drop­ping fra­mes. Enga­ged Jon Mulle to work on trans­la­ting this opencv C++ code to python for use in my expe­ri­ment for eye moni­to­ring via webcam.

Tal­ked at length with Matt Hilchey regar­ding the state of science and our sha­red disi­llu­sion with the scien­ti­fic publishing pro­cess. Advi­sed Heath Mathe­son on mixed effects analy­sis of mul­ti­no­mial data. Pos­ted an ans­wer on stats.stackexchange.com regar­ding analy­sis of ques­tion­naire data.

After mul­ti­ple dis­cus­sions with collea­gues, I’ve deci­ded to give the whole open lab note­book thing another try. I’m about to start up data collec­tion for my dis­ser­ta­tion as well as a comp pro­ject, so this is pro­bably an opti­mal time to start up again. I’ll try to update the page every cou­ple hours with what I’ve done since the last update.

I just pos­ted a pretty subs­tan­tial update to my work on men­tal chro­mo­metry. Pos­ted a new manusc­ript inc­lu­ding a correc­ted model and enhan­ced esti­ma­tion algorithm.

It’s been a while since my last con­tent pos­ting. Part of my fai­lure may be due to the usual sea­so­nal influen­ces, but I think I’m also fin­ding it dif­fi­cult to jus­tify pla­cing work here at a higher prio­rity than the low posi­tion at which it has appa­rently fallen. I guess this is one of the dif­fi­cul­ties in moti­va­ting scien­tists to engage in open note­books; the bene­fits of such an enter­prise are long-term and abs­tract com­pa­red to the day-to-day neces­si­ties of doing science. Yet I remain com­mit­ted to the idea that the trade-off ulti­ma­tely sides in favour of the invest­ment in an open note­book, and I will endea­vour to find ways to remind myself of this when my own moti­va­tion wanes. (I’ve obviously been expo­sed to the lite­ra­ture demons­tra­ting that public reso­lu­tions are the most effec­tive means to beha­viou­ral modification)

Another aspect of open note­books that has trou­bled me of late is the pros­pect of get­ting scoo­ped. When ini­tia­ting this explo­ra­tion, I knew that this was the pri­mary con­cern of most scien­tists when the idea of an open note­book is pro­po­sed, and I alla­yed my own fears by con­vin­cing myself that: (1) open note­books them­sel­ves pro­vide proof of prio­rity of a sort; (2) anyone see­king to mine my pos­tings for scoo­pa­ble mate­rial would likely remain a step behind me, since they’d only get access to my work only up to my latest pos­ting; and (3) being scoo­ped may be an accep­ta­ble cost-of-business if the bene­fits of open note­books become more pre­va­lent. Of course, (1) is only really rele­vant for future his­to­rians of science, since within the annals of science itself peer-reviewed publi­ca­tion is (and should?) be the stan­dard for prio­rity. (2) is also somewhat dubious in that it neces­si­ta­tes inten­tio­nally dela­yed pos­ting of mate­rial, which somewhat vio­la­tes the spi­rit of open­ness. So lately I am left with (3), which again neces­si­ta­tes ele­va­ting an long-term/abstract gain against imme­diate con­cerns of cost.

On the other hand, I recently expe­rien­ced a great exam­ple of the effi­ciency of open­ness while atten­ding the 50^th annual mee­ting of the Psycho­no­mics Society last week in Bos­ton. The pos­ter ses­sions were a great oppor­tu­nity to inte­ract with other scien­tists and share ideas about each other’s work. I’ve hope­fully drum­med up a cou­ple colla­bo­ra­tions out of the expe­rience. I did, howe­ver, feel a pang of regret imme­dia­tely after poin­ting out a flaw in one pos­ter and sug­ges­ting to the author an easy solu­tion. I think it was the fact that the solu­tion was easy, so easy I could have done it myself, that made the inte­rac­tion feel like a mis­sed oppor­tu­nity. On the other hand, while it would have been easy for me to imple­ment the solu­tion, it would cer­tainly have been even easier for the poster’s author to imple­ment it, since he obviously had most of the expe­ri­ment already coded. Hope­fully by fos­te­ring an atmosphere where scien­tists value the long-term effi­ciency of open­ness over imme­diate poten­tial gains through clo­sed com­pe­ti­tion, we’ll speed the pro­gress of science.

In a simi­larly vein, hope­fully down the road I don’t look back on sen­ti­ments like this as exem­plif­ying the naï­veté of youth!

Had­ley Wickham, author of many of my favo­rite R tools (ggplot2, plyr, reshape, cran­tas­tic), has a ggplot2 book now on sale at Ama­zon. I wrote a short review.