[no subject]

[Joseph Osako <scholr@xxxxxxxx>]

Some of you on the main XANADU list may recall that I had expressed
interest in applying hypertext, and in particular Parallel Typefaces, to
Knuth's concept of Literate Programming (for a definition of LP, see
http://www.cs.cmu.edu/~vaschelp/Programming/Literate/literate.html).
Apparently, others have had similar thoughts. I recently
came across this :

Reverse Literate Programming
http://www.ssw.uni-linz.ac.at/Research/Projects/RevLitProg/
and for the paper on the subject, 
http://www.ssw.uni-linz.ac.at/Research/Papers/Kna96.html


While the emphasis is somewhat different, the basic idea is clearly the
same. What strikes me is that it still remains very unidirectional, in the
way that the WWW is, in that has only one dimension to any given link. The
real power of this, IMHO, would come from multiple dimensions, especially
if they presented constraints on each other - for example, you could have a
piece of code, which then links in various directions to the next
section(s) of code (possible several if the code is used in more than one
place), the documentation accompanying it, a section of a flowchart, and a
section of a UML definitions - and any changes made to the code section
would be reflected in those places it links to (or at least would give
reminders to make the needed updates). 

The relevance of this to ZigZag, at least for myself, is that I am
presently planning to pour through the existing ZigZag code and redocument
it in FunnelWeb (a conventional LP formatting tool). I hope to eventually
move it into ZigZag itself, though the current version IIUC would not be
suitable for such use yet. It seems an appropriate use of such a
sophisticated data structuring tool to keep track of its own structure. I'd
been concerned about the feasibilty of this idea, but this seems to confirm
my initial enthusiam. Does anyone else think this is a suitable use of this
technology, or am I wasting my time?

As a side note, it is interesting to see that, like most hypertext systems
such as HTML and XML, most LP tools are based on in-band markup...
apparently, ease of implementation is once again seen as more important
than robustness. For that matter, TeX, Knuth's typesetting tool and the
target format of choice for most LP documentation, also is based on in-band
markeup.


J Osako
Programmer Analyst, Operating Systems Designer, Notational Engineer
http://www.slip.net/~scholr/resume.html