Knowledge and information systems

What an organisation knows is rarely kept in one place. It sits in mailboxes, on drives, in people’s heads and in a filing structure only the person who created it still understands. A knowledge and information system gathers that up and makes it findable.

All open source, as with the other building blocks: Wiki.js and DokuWiki for manuals, procedure descriptions and anything that can be written as a page. XWiki and MediaWiki where structure, permissions and templates matter. Nextcloud Collectives when the knowledge should sit where the files already are.

DokuWiki needs no database and stores everything in text files, which makes backups and moving house simple. Wiki.js is the more modern interface, with full text search and permissions. XWiki handles structured fields and forms, MediaWiki carries large holdings and many authors. Which one fits is decided by the case, not by preference.

First the question of what this is about: which knowledge is regularly missing, who looks for it, and what the search fails on today. The choice follows from that, not the other way round.

Then the setup, the structure of the content, permissions and access, moving existing holdings across, and the connection to what is already running.

And afterwards the running of it: updates, backups and introducing the people who write and maintain. Consulting, prototype, delivery and upkeep come from one hand.

On your own infrastructure. What an organisation knows does not belong on someone else’s service whose terms can change, and open source leaves you the choice of where it sits.

A system without anyone to look after it. A wiki nobody has written in for six months is worse than none: it looks like an answer and gives a wrong one.

And filing every document in one place just because it is possible. A knowledge system answers questions. A record system manages cases. Those are two different jobs, and a tool meant to do both at once usually does neither well.