Guidelines / policy relating to ancillary web services (e.g. Google Custom Search)
Due diligence could include (with search as an example):
0. Understand all the associated issues (e.g. privacy issues, lack of transparency in search prioritisation, dependence on one company for special features, risk of monopolies suppressing innovation, ...)
1. Do a thorough assessment of existing libre software options.
Start with Lucene and its sub-projects:
Find others - e.g. mentioned on WikiEducator, Wikipedia, etc. (via your favourite search engine :-).
2. Consider more creative options which might lend themselves well to education such as the Semantic MediaWiki extension.
3. Compromises may be possible. For example, allow users to set a preferred search service in their preferences. The default should be a libre option such as MWSearch. It may be possible to develop browser extensions (e.g. for Firefox) which search specific repositories.
4. Decouple the best educational search services from WikiEducator (not a new idea - e.g. http://www.oercommons.org). Establish an OERF project to establish or enhance best of breed federated education-oriented search across multiple OER repositories. Encourage third parties to develop educational search engines which also cover multiple sites. User may specify the search to include (e.g.) only WikiEducator [and/or some other group of repositories such as Wikiversity, Connexions, ...]. License should be something like: GNU Affero General Public License.
Hi Kim,
Thanks for these inputs -- much appreciated.
This solid well-founded advice.
- The current Mediawiki search facility is based on Lucene technology, and from what we can see, Lucene is an active and growing community. In alignment with our values -- WikiEducator should not replace an Free Software solution in favour of a non free service. For example, we shouldn't replace the default MW search for a non-free external service.
- We have been talking about the Semantic Mediawiki extension -- it is an impressive technology and I'm keen to investigate further. At the same time, we're being prudent at this stage preferring not to overcommit until we have reasonable assurances of adequate funding to support technologies like SMW as well as code development work to contribute back to the community.
- Agreed -- the default is Libre and our commitment to free cultural works (most notably open and editable file formats for our core OER business is non negotiable.)
- Agreed on the decoupling of services -- search is not our forte but we can focus on creative solutions on our end to provide metadata in ways which can be harvested more effectively through services like oercommons.org (but I do worry about their default license which by our interpretation is a non-free license :-(.), search,creativecommons.org and others.
