We have finally created a public code repository for the ADMIRAL project. It is at http://code.google.com/p/admiral-jiscmrd/.
We had put off establishing a code repository for ADMIRAL until we had a clearer view of the requirements:
- Should it be public, or private? If we are using the repository for user-related information, then a public repository is not appropriate. In any case, changes to Shuffl performed as part of the ADMIRAL project will be maintained within the Shuffl project (http://code.google.com/p/shuffl/).
- What kind of version management is required? We had in mind to use Mercurial, for reasons mentioned elsewhere (http://imageweb.zoo.ox.ac.uk/wiki/index.php/Mercurial_repository_publication), but did not want to commit in case any specific reasons to use a different versioning system were discovered.
Over the past couple of weeks, we have been investing some effort into configuring Samba, Apache and WebDAV to work with Kerberos authentication (http://imageweb.zoo.ox.ac.uk/wiki/index.php/Zakynthos_Configuration). It seems that what we are doing with Kerberos is (a) pushing at the boundaries of what is commonly deployed and documented, and (b) something that a number of people have asked about doing, so we felt it was time to put some of the things we are learning into public view.
We've chosen a Google Code project ("admiral-jiscmrd", as "admiral" was already taken) and have decided to go with the original plan of using Mercurial version management.
The initial content of the public code repository is a set of scripts and configuration files we are developing to automate the assembly of a virtual machine image for file sharing and web access with access control linked to a Kerberos SSO authentication infrastructure (in our case, Oxford University's).