Just saw this on programming.reddit.com - its from the ReadWriteWeb site:
New York Times API coming
' The New York Times newspaper is working on an API that aims to make the entire newspaper "programmable." '
' In addition to the API, New York Times CTO Marc Frons told mediabistro.com that internal developers at the paper will use the platform to organize structured data on the site. Following that, the paper plans to offer developer keys to the API allowing programmers to more easily mash up the paper's structured content -- reviews, event listings, recipes, etc. "The plan is definitely to open [the code] up," Frons said. "How far we don't know." '
Sounds interesting - this could open up many new possibilities. such as mashups, analysis, etc. Also seems to be in line with the trend of more and more things becoming open nowadays, spurred by the Internet, of course.
On a side note, but related, I've been of the opinion, for a while now - as I said here (see the comments on that post), that almost all software should be written to be programmable, right from the start. That is, the software, even if is is a large subsystem or a full application, should be written to be callable (by other code, as a subroutine).
Then write a caller routine to call that subsystem, full app, or smaller module, whatever it is. Though there are bound to be counter-arguments to this (mainly, that it'll take more time), there are clear benefits - the additional code will usually only take a little time, and the advantages to be gained are:
- the code can much more easily be reused for other purposes (and anyone who's being programming for more than a short time knows that most software gets used for other purposes, and lives longer than you expect (if its any good at all, that is)
- the code can more easily be called from a test harness for testing purposes
- Vasudev Ram - Dancing Bison Enterprises