[ documentation home ]
[ configuration ]
[ design/implementation ]
Configuration of a complete wftk system can roughly follow this plan:
As you can see, there's a lot for me to document still. Please bother me until you're satisfied.
- Interface to existing systems
Interaction with other systems is done by means of adaptor. In some cases, there may
already be adaptors for the interoperable systems you need (for instance, there are database
adaptors for ODBC on Windows and for Oracle, there is a notification adaptor for SMTP outgoing
mail, and so on), but you may need to write some adaptors. If you find yourself in that
situation, see the Interface Writer's Guide.
- Determine front-end needs
The wftk at present (July 25, 2001) has no front-end software; it is a library suitable for
embedding in other programs. (The exception to this is the command-line utility, but that
doesn't expose all the functionality of the library.) Work is in progress for a Zope frontend,
which will include a Python wrapper, and CGI programs are also easy to create for wftk.
- Create a configuration file
The configuration file needn't actually be a file. The configuration is attached to the session
by the calling program, so it can be implemented any way you need. An example config.xml file
is included in the distribution; at the moment there is no other documentation. (Coming soon.)
- Define users and groups
If you write a user adaptor to an existing directory, you don't need this step.
- Define permissions, if you will be using the action interface
- Define processes (procdefs)
See the example chair procdef for, yes, sigh, the only documentation I've written so far.
Copyright (c) 2000 Vivtek. Please see the licensing
terms for more information.