[BioPython] Re: [Bioperl-l] wiki

Johann Visagie johann@egenetics.com
Thu, 7 Sep 2000 10:32:26 +0200


Jeffrey Chang on 2000-09-06 (Wed) at 15:39:14 -0700:
> 
> [I switched this from the bioperl-l list to biopython.  I'm probably
> breaking some etiquette, but it's more relevant here anyway.  :) ]

:->

> Yes, I'm fascinated by it as well.  However, I'm trying to figure out how
> WikiWiki might be used to support biopython documentation.  I like the
> idea that anyone can edit the pages, so that creating documentation
> becomes a community effort.

Wiki is by no means the only way to have online content that can be edited
_in situ_, so to speak.  The John Udell article that I have quoted numerous
times lists a number products that allow you to do this, so I won't repeat
all that.

All these different products allow you to do "in place" editing of web
documents in different ways.  Wiki - at least the original WikiWikiWeb
- was intended for very free-form...  "post-modern", if you will...
brainstorming.  Don't get me wrong - I *adore* wiki - I just don't think
it's necessarily the best tool for producing something as inherently
structured as documentation.

Now, to get back onto my favourite hobby horse, Zope.  (Crowd goes, "Boo!")
:-)

Since wiki now dominates mindshare for in place web editing mechanisms, one
way of describing Zope is as "wiki+++, with the added bonus that you can also
manage _active_ content in an object oriented fashion".

Every Zope URL represents an object (strictly, a method of an object - often
a default "display" method analogous to Python's string representation of
objects).  Every Zope object can be managed online, "in place" by appending
"?manage" to the URL.  (Actually, most objects have a number of management
methods.)

Most Zope applications are written by subclassing built-in Zope classes and
extending their functionality.  As an example, implementing wiki in Zope is
fairly easy.  The "ZWiki" Zope product actually contains surprisingly little
code.  Built-in zope classes already contain all the methods for displaying
themselves via the web, editing themsleves in place, etc.  The author of
ZWiki merely had to specialise some of these to emulate how things are
usually done in a wiki.  In effect, it's Zope with spcialised (simplified)
front-end.

Ditto for many other projects, e.g. I recently played with a "knowledge base"
product for Zope, which gives a user a simple interface whereby a few sorts
of objects can be added or edited, e.g. a "FAQ page" object which is a
container for "FAQ question" objects, a "HOW-TO" object, etc.

> However, we do need to have some concrete documentation that we can
> package and distribute in some format that people can print out.  Some
> people (like me) are not always working in places where they have internet
> access.  Thus, we'll keep on working on our LaTeX-based documentation.

Now there's the point - and I agree with you:  These online documenting
methodologies are all very funky, but they're not likely to replaced
structured text (be it TeX-based, SGML-based or XML-based).  One of the
elements of the putative "online collaboration server" I've mentioned in the
post to the BioPerl list would be a facility for collaboratively building
structured documentation, with the ability to export to other structured text
formats - something I see a real need for.

However, this is all still pie-in-the-sky stuff...  The closest I see anyone
coming to this is the "StructuredTextNG" project which is part of Zope, and
which may be incorporated into "WikiNG".  But we'll see.  As I said, I am
talking to some people who also want to see something similar happen, and
maybe we'll even get some momentum going yet.  :-)

-- Johann