Context Navigation

TiddlyWiki Server Side

Background

There are four distinct ways in which TiddlyWiki can interact with a server:

the server can generate a TiddlyWiki file with baked-in content. Usually, this means marking up empty.html so that it can be treated as a template into which tiddler text is spliced from, say, a database
using the UploadPlugin? (or the new WebDAVSavingPlugin for TW 2.2) TiddlyWiki can save changes to a server instead of to a local file
the server can participate in TiddlyWiki's infrastructure for importing, exporting and synchronising tiddlers between occasionally or manually connected stores
the server can take over key TiddlyWiki operations to give the user the semantics of a traditional shared wiki

This discussion is chiefly concerned with the fourth category, but they overlap somewhat - these servers would still use the first technique to splice their machinery into the TiddlyWiki file in the first place.

Goals

To have a common server API that means you could link any client (from a full serverside TW plugin, to a tiny TW plugin to save Minesweeper scores to a common server, to a TiddlySnip-type browser extension) to any server (db-backed with versioning services like ccT and ZW, a read-only plugin repository, etc etc).

TW shouldn't really require it's own custom server implementations, but rather should use industry standard protocols to exploit existing servers. For example, Amazon's S3 and similar services expose commodotised data storage that is suitable for TiddlyWiki as a file-level storage medium. Similarly, Zope provides a commodotised content management system that TiddlyWiki can use as a tiddler-level storage medium.