From: "M. Edward (Ed) Borasky" <znmeb@...>
Date: 2006-07-05T07:46:44+09:00
Subject: Re: ed borasky and his second favourite programming language

Phil Tomson wrote:
> However, having some experience with creating DSLs for describing and
> simulating quantum circuits [QDL] and for RTL/behavioral level
> description/simulation [RHDL], Ihave the suspicion that Ruby is
> flexible enough such that we could create some domain specific
> languages with it (where the domains can be science and math).   It
> would be cool to create several little specialized languages in Ruby
> so that we could have the 'best of both worlds' - good for specialized
> domains while also being good at solving general propgramming
> problems.  One can imagine that we could create a specialized language
> using Ruby that could be every bit as useful as R for statistics, for
> example.
Useful, sure ... as fast as, as well supported by "professional
statisticians" ... that I rather doubt. Languages -- both human and
programming -- co-evolve within a community of creators and users. R
co-evolved in a community of statisticians with a highly numerical
approach. Maxima co-evolved in a community of scientists with lots of
symbolic math to do, mostly having to do with the physics of nuclear
energy. It's pretty clear to me that Rails co-evolved in a community of
web application designers and users, but I'm not sure about Ruby -- I'm
a latecomer to Ruby.


>
> Also, one can imagine that we could perhaps create something very new
> given Ruby's strengths right now (Rails and web programming in
> general): The web-enabled research paper.  When I was working on my
> master's thesis I kept thinking that it would be great if I could not
> only read the reference papers, but also access their experimental
> data seamlessly.  And perhaps not only the data , but also the
> programs used to generate that data.  What if graphs in a web-enabled
> research paper were not just graphs, but you could also 'push-into'
> them and access the underlying data?  It would make it easy to then
> compare data from other research in your own paper.  crazy idea?
> Couldn't a mashup of web programming and sci-research-languages lead
> to a new level of collaboration?
Could ... and ... (pregnant pause) has. Donald Knuth calls this
"literate programming." In the Python universe, it's well supported by
Ed Ream's "Leo" editor
(http://webpages.charter.net/edreamleo/front.html). R has literate
programming "built in" using something called "SWeave". See

http://www.bepress.com/bioconductor/paper2

for some pointers. I have quite a few links stashed away on literate
programming, mostly in R. I haven't taken the time to do a search for
literate programming in the Ruby context, but I'm sure it's there.
>
> For example: my master's thesis is up on the web in a pdf.  A couple
> of weeks ago someone doing similar research sent me some questions
> about it and asked for more info.  What if instead of a static pdf, my
> thesis was a web app?  The person who was sending the inquiry could
> have clicked on the graph in question and downloaded the data used to
> create the graph as well as the program used to create the data.  They
> could have then gone on to modify the algorithm used in the program
> somewhat to test their theory and then 'publish' the results back to
> the web app (or perhaps post their own web app).  Maybe it would work
> something like a wiki so that the research continues to grow on one
> site?   That could certainly lead to a new level of collaboration.
Axiom is another language that's heavily into literate programming and
in fact has a wiki as its main portal.

http://wiki.axiom-developer.org/FrontPage

>
> ...just brainstorming...
Amazing, isn't it? "Not even the tread of mighty armies can stop an idea
whose time has come." :)

-- 
M. Edward (Ed) Borasky

http://linuxcapacityplanning.com