The LionShare project has been progressing on quite well over the course of the last year. Because this is an academic oriented project, we had to evaluate hundreds of schemas, frameworks, API’s, and many other “standards” for use in this project. Some of these projects are logical fits for LionShare, but the resounding majority of the projects that get pushed our way are not a good fit.
One pattern I’ve seen quite a bit of since starting this project, is academic oriented standards that refuse to provide implementation code or in the case of schemas, data files. Many academics for either philosophical reasons or lack of resources/coding experience, define standards but provide no example implementation. There are even standards organizations in academia that have policies that discourage sponsored projects from providing implementation code or data files. Many of the projects that have sample files tend to keep them buried in their website, to the point where if you want to find an XSD example of an academic metadata standard, be prepared to google.
The academia model is the complete opposite of the open source model for creating standards. The open source model goes by the law of “put up or shut up.” I couldn’t imagine trying to create an API, library, or data format for use in open source software that does not provide an implementation of some sort. Lets say for a moment, that the open office file format didn’t exist and I wanted to create an open source document format standard. Do you think my standard would have a chance in hell of getting adopted by anyone, if I didn’t provide a data sample and a library?
In academia, many people are forced to adopt formats or “standards” because of politics rather than common sense. While open source software is not politic free by any means, open source developers have the freedom to choose whatever software and standards they want to use without having to worry that someone will be offended. Its just another dimension of the freedom in free software, freedom from political crap. If somebody doesn’t like it, they are free to take their code and do with it as they wish.
The academic standards community would be well served by looking at how other development communities define, adopt, and use standards. Academic standards organizations should require real working implementations for all their offerings. Not test code or placeholders, real ready-to-use code and/or data structures.