• Georg Wrede (54/54) Sep 27 2006 Litterate Programming was all the rage a few years ago.

Georg Wrede <georg.wrede nospam.org> writes:

Litterate Programming was all the rage a few years ago.

In article

http://www.scottcollins.net/articles/Loops.html

Scott flaunts his agility with languages and the results of using 
Litterate Programming tools.


What came to my mind was a few discussions with Walter some years ago, 
that rose from the fact that the DMD compiler can use HTML as input. I 
suspect sensations of Literate Programming had influenced this decision.

However, it seems that this feature has been largely unexploited by the 
community, which is a shame. It becomes understandable if one considers 
this HTML-decision to be 'just an expression of the subconscious 
aspiring towards goals perceived as worthy or useful', without the 
rigorous cognitive analysis and evaluation that such a major decision 
deserves, and therefore as half-baked as the thoughts in this text.(1)

(In English: This is gonna be a good thing, just don't know yet or have 
the time to sort out the details.)

The mundane reason for exactly why HTML-source code for D is extinct is 
probably not that it's a bad idea, rather, that it's possible that 
people have not found any way to incorporate HTML-d files into their 
development toolchain. (How to edit a file that needs both HTML and D 
highlighting, how to figure out the line numbers for erros, how to 
combine DDOC comments with regular comments with regular text in the file?)

(One side track of our discussion was about whether it is good or bad to 
throw the HTML through a preprocessor, which could potentially enhance 
both the HTML and the .d code, for example by unraveling some ENUM vs. 
case details (IIRC)).

---

In another discussion I was entertaining the notion of a program code 
editor where the "program source" never actually existed, neither in a 
source file nor as a string in memory.

Having said the above and all the previous, I must say it felt eery to 
read about the Leo Editor.

http://webpages.charter.net/edreamleo/front.html

Leo implements many of the ideas that were floating in the air at the 
time of our discussions.

---

Literate programming seems to be (not an alternative of, but an 
enhancement to) something we probably should start using, irrespectively 
of whether we adhere to some other Programming Paradigm or whatever.

My question is, wouldn't it be much more efficient to implement 
something Leo like in D than in Python. (Ok I admit: if we exclude the 
in-line directives and enhancements that are possible thanks to the mere 
fact that Python is an interpreted language.)

Something akin (and possibly more powerful) could be achieved if this 
Leo-successor implemented a way of writing ad-hoc D routines that would 
automatically get included (somehow DDL, rdmd, and/or plugin like) if 
written in-line.

---

(1) I have thrown out merely my feelings here, and thus the ideas here 
are, at best, equally "half-baked".

Still, somehow it would be an itriguing thought if one day a Leo-like 
editor would be an integral part of any DMD distribution, including all 
the infrastructure and possibilities that came with it. "Why settle for 
a regular IDE for an exceptional language?"