I think there's a time and place for open-source, and a time and place for proprietary solutions which support open specifications. NYC.gov has a lot of complexity and varying degrees of reliability requirements which can't be supported by just one platform/architecture. So I didn't vote, as from my point of view, both answers are simultaneously valid and invalid.
How about "Should the main NYC.gov website run to the maximum extent feasible on open source technologies?"
Sure the NYC.gov site can be run on open source. There are a few elements that cannot be run on open source such as the NYPD, but as far as i know, most other legislative government sites (made for the people) use drupal. The New York State Senate, The White House, etc…..however there are no open source .mil sites.
seems like a good rule of thumb
Coming from a group that focuses on analytics, we have made modest use of R and Python for statistical analyses and web analytics. In the past, our group has primarily used SAS and SPSS (not open). For our use cases R and Python could definitely play a much larger role than they do now, but until we have a critical mass of users (like we have with SAS and SPSS) we cannot move the more substantive aspects of our workstream to R and Python. I.e. we need something that is going to be maintainable despite turnover. There are other dimensions and tradeoffs involved with moving to R and Python, but I will keep it short.
If folks are currently using or interested in using R or Python for analytics, please reach out. I think this sort of interagency collaboration may be one way to address the critical mass issue.