Latest CKAN news for the past two weeks:
- One of our Amazon EC2 servers had catastrophic hardware failure(!) taking with it serveral community CKAN sites. Nils and Friedrich restored them on new infrastructure, with the larger sites no.ckan.net br.ckan.net ca.ckan.net etc. are now running on the latest CKAN release and much faster too! The smaller ones are looking to be integrated into thedatahub.org
- Now we have upped the translations of the CKAN UI, David Read has been getting browser language detection working and the language switcher working more reliably. To try language detection, clear your current language setting by deleting your CKAN cookie and browser cache and then going to http://thedatahub.org/.
- Drupal integration continues to be a hot topic with various parties getting setup. The 'loose-coupling' method rounded on at the open Skype meeting last month continues to progress.
- The Archiver extension is being quickly developed by David Raznick. Now when you add any CSV link to thedatahub.org, a background process will download the CSV data into the Webstore. So when we now just need to expose in the web page a button to preview, download it multiple formats and access a nice API to it. We also want to add a progress bar for the download and report errors in the CSV file. Any more ideas and help, do let us know at ckan-discuss.
- We're having a Webstore meeting on Skype this Thursday, as we look at merging in features from some similar open source projects. Join in if interested or want to contribute to this open source data store. http://ckan.okfnpad.org/webstore-meetup
- New developer Ian Murray starts in the team today, with another couple of developers being interviewed in the next week. We're always on the look-out for talent: http://okfn.org/jobs/
- Ira has mocked up an improved workflow for adding a dataset and is looking for feedback - look out for it on the blog this week.
- A big discussion is going on amongst the dev team about how to express dataset groups/publishers and the permissions. It's basically it's a choice between a simple/easy model like Github, where every dataset has an owner, or something more flexible like unix file permissions. Modelling, use cases and discussion is going on here: http://wiki.ckan.org/Groups_Refactor
- Integration of PyCSW into CKAN - discussions of how to do this have been moving forward, with the PyCSW team chatting with Adria on the CKAN IRC channel (irc.freenode.net #ckan)
- A big milestone: CKAN 1.5 is looking ready to release this week and James will also have it ready as an Ubuntu package for easy install.