ScrollServer Initial Release

I am very pleased to announce the initial public release of ScrollServer, version 0.4.

ScrollServer is a Python based http server that provides access to a scrollkeeper database with any web browser.

ScrollServer is in very early development, but it is already useful. It has been tested on RedHat 7.1 and Debian Woody. Reports of success or failure on other systems greatly appreciated.

I'm working towards a 1.0 release which will implement basic functionality.
It must be usable in text browsers but attractive in X.
It must Serve all documentation formats supported by ScrollKeeper.
It must provide browsing by document title or topic.

In later releases searching should be powerful but easy. It's a classic trade off, but I'm going to try. It will probably mean writing my own DocBook-aware search library.

There are more ambitious goals in the long term. Eventally, ScrollServers will be able to talk to one another, and to share information with each other. One ScrollServer will `publish' a document or five hundred, and another will `subscribe' to the `publication'. This is mainly for the benefit of the Linux Documentation Project and others who provide more general Linux information.

Each subscription can be left offsite or cached locally, updating the local ScrollKeeper database. If the local ScrollServer database has a cached copy of the document, it will be used in preference to the online copy.

Eventually a reader might be able to rate a document, and have their comments sent to the publisher (i.e., the other ScrollServer) and also to the author of the document, which can be extracted from the DocBook.

Sound like an interesting project?

Known Bugs

Gnome specific uris such as gnome-help:foo don't work. I and the gnome-doc people are working on a solution. It may just mean work in the xsl stylesheets.

KDE documents haven't even been looked at yet.

HTML is cached, but there is no cache management.

David Merrill
david@lupercalia.net