"Publishing Open Data - Do you really need an API?"

[Jury Konga]

I found this article to be of interest http://www.peterkrantz.com/2012/publishing-open-data-api-design/
I suspect this group may have some opinions. Looking forward to
seeing feedback.

Cheers Jury

[kevin curry]

Good link, Jury. Mark Headd & I were just talking about it. I think the headline/grabline does the argument an injustice and the point is a bit convoluted but I agree with the essence. It's not as if APIs aren't needed. They are. And yes, gov should dump data first before they take the time to build the wrong API. But the problem cited in the first implementation is a design issue, not an API vs. no API issue. Whether built by gov or not, the author says the second design is better. Had gov built the second design would we be having this discussion? I suppose the author would recommend putting out the data before taking the time to design the wrong API and I would tend to agree. Of course, that doesn't come without backlash either. Developers expect an API. End users expect a working product. So when we convinced Hampton Roads Transit to dump its raw log files for real time bus location, people still complained. We went from nothing to something and it still wasn't good enough. :^) The first thing I learned in my CS schooling is the one thing I always come back to: there are no silver bullets.

Mark reminded me (us) to check out this passage from the Open Data Handbook:

http://opendatahandbook.org/en/how-to-open-up-data/make-data-available.html

--
Kevin Curry
Virginia Beach, VA
http://twitter.com/kmcurry