this week i

visited isaiah's school.  sunday dinner at ruth's house with custom armoire, we ate beans and okra and eggplant and polenta and watermelon and

do you like pineapple?
i love pineapple
great, today is your day
 



 
rest atop one of seven hills, a potential energy hoard.  butterfly benjamin button'd become . . . a caterpillar?  why're we wired to record the days?


find suggestion agreement, unclear on grammar syntactician.  humble captcha reminds despite our prosperity & wit, we've yet to replicate ourselves

organize discussions of professional expertise to make friends, zero failures so far.  christine (from arua) a few hours after our meal: am still satisfied
 
 
dream what mozart'd have done, electric bass & kit of drum. 'sadder huckabee in military kit and/or lula learned nought from the biden hubris family


yelled scram at a domino's pizza abroad then realized the futility of the moment. guh guh ghost in forty nine syllables, present tense death fascinates

pair an ear and spare a quid
for a séance or a kid
'fore four eyes flashed two lifes rid
coiled basket cobra hid
till i frisbee-tossed the lid
bede & me goodbyes bid
damndest thing i ever did

 

submitted a ltte

 

Official Statisticians Should Abandon Proprietary Software

The cost-cutting drive spearheaded by Josh Gruenbaum ("Procurement chief vows to continue supplier price squeeze", Report, July 22) would win plaudits from nearly all segments of society if accompanied by a review of open source technologies. Agencies like the U.S. Census Bureau and the broader Federal Statistical System deserve neverending praise for their global standard-setting commitment to publicly available, respondent-level datasets (often termed microdata). Major research initiatives like the Current Population Survey can be downloaded with ease, on an annual basis, dating back to the JFK era. Unfortunately, whether due to entrenched interests or perceived changeover costs, government choice of analysis software has not yet made commensurate strides.

During the final quarter of the twentieth century, only closed-source tools like SAS (previously known as the "Statistical Analysis System") were capable of digesting what was then considered big data and, lacking alternatives, quantitative offices within the federal workforce grew reliant on this expensive, subscription-based syntax. But SAS and its proprietary peers continue to dominate government-funded research efforts even as top-tier universities enter their third decade training graduates in computing languages built for this side of the millenium - I first encountered the free & open-licensed R project at JHSPH in 2007 and have not looked back.

I would encourage anyone skeptical of the maturity of the open source community to skim the "Official Statistics CRAN Task View" page and find there detailed explanations of advanced technologies freely available to all. To follow Mr. Gruenbaum's principle, I challenge closed-source enterprises to explain to a 15-year-old (or anyone else) why they deserve taxpayer dollars where superior tools can be had for nothing. I commend my many colleagues committed to reproducible research and transparent methods, and I ask federal agencies to similarly aim toward this forward-looking paradigm.

Anthony Damico
Washington D.C.


8/7