[Moin-user] WikiBooks and Quiz Software: General Discussion

[R.Bauer]

Hi

Am 23.07.2011 23:40, schrieb Neal Holtz:
> Hello - I'm just looking for general comments, if anyone cares
> to give them. (I apologize in advance for the length of this msg).
> 
> I'm planning to start developing a wiki text book for a university
> level engineering course I'm teaching.  I've so far only got a page
> or two and a rough structure as a proof of concept.  In addition
> to the easily added extra value such as images, photos, problem
> solving videos (ala Khan Academy), I'ld like other media there
> as well, such as interactive self test questions, application
> software, perhaps links to things like sagemath worksheets, etc.
> 
> Oh yes: a lot of mathematics is involved.

Since Mathjax there is no problem with latex formulars anymore
I have recently packaged it for xstatic and added it to moin-2.

http://www.mathjax.org/
http://pypi.python.org/pypi/XStatic-MathJax/1.1.2
mathjax uses fonts, css und js to render latex formulars.
This works with any markup.


> 
> = One =
> 
> I suppose I should really start with WikiBooks, particularly as
> there is the start of a text on the subject there already.  It does
> provide some nice stuff out of the box, such as pdf creation
> for an entire text (or subset, I assume).  However, the things 
> I don't like are:
>   - I don't think I can get the software, and I really don't want
>     to host this on their server.
>   - MediaWiki on which it based can be installed, but I really
>     don't want to program in php if I don't have to ...  And I
>     do expect some new code will be necessary
> 
> So I guess if I want to extend a wiki system using Python, that 
> MoinMoin is the best choice.  That was my first discussion point.

We have written a small tool
https://utils.icg.kfa-juelich.de/hg/xmlrpc/file/96318c07f68f/wiki2pdf/wiki2pdf.py
(part of this repository
https://utils.icg.kfa-juelich.de/hg/xmlrpc/)

By this you can fetch an hierarchy of pages and attachments. This is
then transformed by sphinx to a book.


> 
> = Two =
> 
> Markup language: I was pulled toward reStructured text,
> again because of all the extra stuff that is there already
> built around it.  However, after doing a couple of pages,
> I really dislike it (esthetically).  I really do prefer using
> wiki markup.  I looked at creole in hopes of getting something
> that might be more portable but that doesn't seem to
> be the silver bullet either, yet.

Because of sphinx as the processor we had to use reST. For some special
pages with very complicated tables we introduced also a server side
parser which converts moin wiki page markup to a svg, png, pdf, latex
table. The table is shown as svg graphic to the user. The script lets
sphinx use pdf files if available to render the document. If some images
are missing placeholders become inserted.


> 
> I'll want things like the ability to generate text-book-like
> pdfs of a set of pages, so students can print easier.  
> 
> Should I stick with wiki markup and develop the required
> tools based on that?  (they are not high priority).


Have a look at sphinx
http://sphinx.pocoo.org/
if you write python code you anyway may decide to use it for the
programs documentation.

moin-2 has already a convertor for domtree conversion of moin wiki
markup to reST markup. So if we are ready for publishing 2.0 you can
keep another markup if you dislike reST. But there are also some
differences in the markup. Probably you can't lossless convert.



> 
> = Three =
> 
> Now - self test questions:  I'm intrigued about leveraging
> moodle for this purpose; I can easily see having questions
> for grading students within moodle - I will probably eventually 
> do so.
> 
> Moodle stores questions in a relational database, and can
> export them in several formats.  I think it might be possible
> to develop some python code to either access the moodle
> database directly, or read the exported xml (I'ld probably
> do both). Then render the questions for anonymous self
> testing for wiki book readers.  This is a significant amount
> of work, but I don't think its huge (I have done this before,
> in 1994, believe it or not ... :-[
> http://http-server.carleton.ca/~nholtz/tut/doc/doc.html
> )
> 
> Most of the hard work is in the user interface to develop
> questions, and thats already been done in moodle.
> 
> The way I see this is developing a parser for moin that
> would read either the moodle xml or moodle database
> info, and present the question.
> 
> Any other quiz engines in python?

see

http://pypi.python.org/pypi?%3Aaction=search&term=quiz&submit=search

> 
> If you got this far, you patience is remarkable.
> 
> Thanks - any kind of comments are welcome.
> 
> 
> 
cheers
Reimar