Hi All,
Whilst we should discuss the possible solutions to this, I feel I should sound a note of caution. Any solution that requires 'hiding' or 'rewriting' URLs into the Daisy namespace is complicated and will require a significant amount of my time. Time that I do not have to give to openEHR at the moment. I spent nearly 2 whole weeks exclusively working on the new openEHR web site recently, this was to the detriment of the work that I get paid to do. That work now has to take priority.
That said I still want to contribute to the discussion, to ensure any agreed solution is supportable.
The Apache rewrites we put into place to get the openehr.org site working, with the URLs hiding the Daisy namespace structure are complicated and fragile. And there are still some aspects of Daisy that do not work properly via this mechanism. Adding more sites, with the requirement of hiding the Daisy namespace is going to add to the complexity, fagileness and have an impact on maintainability.
Daisy is designed so that namspaces are determined by the top level directory in the URL. All the solutions shown in the Daisy docs and Wiki assume that this is the case, and there will be a single site that has that hidden, with all others showing the top level directory.
I can think of a possible way that openehr.cl could work, by setting up another virtual server for that site. But this would need to be setup, tested and any maintainability issues considered. We would also need to consider the scalability of this solution, if and when other country/language versions are required.
Another possibility (and one that I think is more supportable) is that we rewrite openehr.cl/ to be openehr.cl/es-cl/ as a single rewrite, but all other URLs have the '/es-cl' directory exposed. This would only require a single rewrite in Apache and would work better with the Daisy srtuctures.
It seems to me that we are constantly trying to 'fight' against the basic organisational structure that Daisy is designed to work with. That is always going to cause problems. The openEHR community needs to accept that this is part of the solution that was chosen and start working with it.
At this stage I feel I must say that I am not defending Daisy, I was not involved in the original decision and I am currently neutral about the technology.
Sergio Carmona wrote: