I'm contemplating migrating this blog from drupal to Mephisto. The main reason is to put it on a more hackable platform. Drupal is theoretically hackable, being written in php, but the problem is that I find php a weak and ugly language, I'd rather play with Python or Ruby. I haven't been able to find a blogging platform written in Python, just a number of half finished projects (UPDATE: there is pybloxom but that uses files rather than a database which puts me off: SQL FTW). Ruby has at least two that I have found, both based on Ruby on Rails: typo and mephisto. Typo is a more popular application but from what I have read is leaning towards being a test bed for experimentation with Ajax. Mephisto seems cleaner and better designed so that is where I am aiming. Mephisto does seem a little cultish and doesn't have a large and active community. The principal developers seem very capable but are busy with something else.
Migrating the blog to Mephisto or any other platform has a number of problems which I am enumerating here as part of my planning:
- All the articles in the drupal database must be migrated. There is a framework here for migrating blogs to Mephisto and although it doesn't cover drupal (yet) this is a good start.
- The articles are all in wilki format and either need converting to textile (my preferred standard markup language) or a wilki filter needs to be written. Converting to textile is feasible and using a standard would make it easier to migrate again in future. However the conversion process would need to be perfect, developing a wilki filter has the advantage that I can fix presentation problems I find later, much easier than finding a problem and having to trawl the database to somehow identify and fix the broken textile.
- wilki uses the same syntax to link to other posts by node id and to link to tags by name. These links need migrating appropriately.
- Drupal offers many rss feeds based on categorys, tags, different paths, rss, atom etc. My access logs show a variety of these are being used but I don't know if any of these are actually being read by humans, scraped by bots or whether they are just ancient Rss aggregator subscriptions. I need a way to map all these old feeds onto appropriate feeds in the new system.
This isn't a simple matter. And I really need to be sure in myself that I find the right target before going through this hassle.
Peter's Moblog
