I’ve created a pretty amazing SMW intranet installation for our engineering department. Without spending too much time describing it, at the high level it
- Promotes reusability though automatic cross-referencing and advertising of content
- Uses forms and templates to support symmetric looking page types (like project pages, technical articles, status reports, user profile pages) while still including free text for customization of information
- Includes parametric search based on several families of tags that are attached to almost everything
It’s been evolving and running for over a year, has a few thousand meaningful articles, and has been well-received by our engineering community.
We also have a foswiki installation. It started as twiki more than 7 years ago and was migrated to foswiki about a year ago. The foswiki install is on an IT-managed server and is the official corporate wiki.
In an effort to streamline operations, the decision has been made to scrap our SMW server. For the most part, the decision makers used the point that SMW is a wiki, foswiki is a wiki, foswiki is already adopted, we’ll use foswiki. The infrastructure I added was essentially deemed insignificant in the arguments (non-engineers made this assessment, I couldn’t convince them otherwise after a year of trying, nor could other engineers). However, they did say that my designs could be handed to our foswiki team and adapted to it.
I’m writing to the developer list to ask for insights into why foswiki can or cannot support such a system. I know it supports forms and templates, but I have my doubts about the rich query support I’m doing, given that foswiki uses text files for storage and its own core for access, while Mediawiki uses a database and lets the mysql process do a lot of work for it. What I don’t have is real evidence. Does anyone have any empirical or other statements for why such a transfer can’t take place?