Pages

Sunday 31 July 2011

Open Surface and the value of Public API

Microsoft recently posited the idea of "open surface" as an alternative to "open source". Correctly noting that a stable public API, protocol, standards are of great value.

This is actually a position I have great sympathy with:

  • OGC open standards are considerably easier to adopt; then the pay to play ISO ones.
  • Sextante - Is propietary software better for developers? noted that when a public API is the only mechanism for communication with developers that the API / Docs / Code examples are very much improved as a result
  • This is contrasted with "Open Source Code" translating as open the source and read the code - a different value proposition to be sure!
  • API stands for "Application Programming Interface" - and is generally the information needed to use an application or library as a programmer. It documents what can be done on a function by function (or method by method) level. Traditionally this was limited to describing the parameters; inputs and outputs, and for really good docs providing an example use.
Related documentation:
  • GeoTools javadocs
  • uDig - does not publish javadocs (as they are generated as needed in Eclipse and shown as tooltips during development)
  • GeoServer javadocs - I could not find them easily as the website is primarily user focused. As with uDig docs I did not notice they were "missing" as Eclipse generates them on the fly.
I will point that the general approach here is to really lean on tools like Javadocs, or Sphinx that process API documentation that is coded along with the source code.

No comments: