Masonik: Mason and IPC-Run

Hi,

I need to give you guys a little intro to set the stage for my 
discussion. Please bear with me.

Mason appears pivotal in a new initiative a group us are launching to 
restart an Apollo-era technology of mathematical programming. I am the 
only remaining original developer, and the rest of the group 
(physicists, mainly) are former users. This is a "metaphoric" 
programming technology for end-users who are very skilled scientists but 
are "casual" programmers. Compared to "very-high-level languages" like 
Perl or Python, this is way up in the stratosphere. Metaphoric in this 
context means "non-algorithmic", but the languages are derivatives of 
Fortran. Under the hood they do automatic differentiation (a vector 
calculus arithmetic) and recursive differential coordinate 
transformations to automatically generate "holonic" algorithms. But the 
users are only slightly aware of this, and it is like magic to them (all 
hidden). The net effect is a novel property we call "super 
simplification" which is experienced in other metaphoric programming 
media like spreadsheets (i.e. many spreadsheet users don't think of it 
as programming).

The technology goes by the name "meta calculus" (AKA synthetic 
calculus). It can be thought of as a mathematical analog of SQL. Whereas 
SQL searches databases, meta calculus searches mathematical surfaces 
that are synthesized dynamically by simulation (differential equations, 
etc.). The primary meta-calculus metaphor is the Find statement, a 
direct analog to the SQL select statement. Like SQL select, Find can be 
recursively nested. This is what the differential coordinate 
transformations and "holonic" algorithms are all about.

This technology was marketed nationally in the time-sharing era of the 
1970s, in a language called PROSE. Over 170 man years of development 
have gone into it, and today there is nothing out there that comes close 
to its power. That is why we want to bring it back. We think it can 
create a renaissance in R&D computing, where labor-intensive algorithmic 
programming has become too overspecialized to awaken demand. (See FAQ 
<http://www.metacalculus.net/doc/initiative/Meta_Science_Testament_FAQ.htm> 
for more info)

Ok. The reason I am starting with you guys is that I think the 
combination of Mason and IPC::Run (under Apache and ModPerl) can produce 
a new scripting layer that can make the web a casual 
application-development medium for scientific end-user--starting with 
meta calculus.

I wrote a Mason script that employs IPC::Run to control a child process 
(a meta calculus demo program that interacts with the user). I 
successfully tested it as a Perl wrapper, and it works fine, generating 
text and png/gif graphs. But when I converted it back into a Mason 
component running in ModPerl, it does not generate any web output. I 
have not exhaustively debugged it yet, and perhaps it is something 
minor. But since I am racing beyond my expertise, I thought I had better 
talk to the gurus about this whole plan. I may have extrapolated beyond 
the current state of the capabilities by combining these two modules in 
the Apache ModPerl environment. If I have, then I want to make this an 
issue of needed functionality in the whole community, because I see it 
as a great leap forward.

What I am talking about is a functionality I am calling WebLinux 
<http://www.metacalculus.net/doc/initiative/Meta_Science_Testament_FAQ.htm>. 
I have created four projects on the Perl Apprenticship site which I want 
to populate. So I am officially calling for your help (contacts, etc.) 
in getting this inititative rolling.

Thanks in advance for your examination of these issues.

-- 
Joseph 'Bear' Thames
Meta Software Engineering
Tel: (408) 873-0658
[email protected]