Mitchell N. Charity

Some local pages:

Education

Software Engineering

Misc

A few other "current" projects:

Education:

• F=ma presented as interactive order-of-magnitude scales - I still don't understand why an oom quantitative understanding of equations isn't taught. Despite a half century of bemoaning the lack. This is a DHTML/javascript hack to drum up interest. Interestingly, I could be surprised by even this most familiar of equations (a cell, accelerated by a magic zero-mass constant-force hand-cranked transmission, takes how long to go relativistic?).
• Cosmic Signpost: Like you have at the crossroads in European villages. But with more varied arrows. Including a transparent globe variant of pointing down (a "this _site_ up" globe - another concept which needs a webpage). And an Earth's rotation corrected pole, with planet, stellar, and galactic arrows (spinward, coreward, etc). I'm hoping to do this for a spring '02 '03 astronomy day at the boston museum of science, and spin off a web page. But that's been true for at least two three years now.
• Dot-a-day history: Having a distinct dot for every day gives a tangibility to history. Mousing over days triggers DHTML pop-ups, sidebars, overlays, etc, allowing one to browse history. Javascript. Open API.
• Fermi Question game webpage - an interactive webpage to teach FQ problem solving.
• "What is science?" - teaching from pathology: The nature of science is often taught from gem-like definitions of limited utility, or from unmotivated artifacts ("do logs"). A complementary perspective is that individual and group thought has a number of common failure modes, and science is, at least in part, a set of approaches, culture and institutions, designed to mitigate them. If one could create introductory material which systematically induced such pathologies (eg, message corruption in relayed verbal communication ("telephone" or "gossip game")), and demonstrate the efficacy of science's assorted mitigation strategies, one might leave students motivated and capable of applying them in their own lives.
  (Another, somewhat related need, is for some `observer of science' skills. `How do you recognize the health of a subfield?', etc.)
• "You can't teach what you don't understand" - Are science education difficulties a symptom of a deeper problem? The physics education research literature contains each level of teacher bemoaning the shallow and broken understanding observed in their students, and the observation that most instructors are unaware of this failure. K-12 physics understanding is bemoaned by their teachers, the teachers by their instructors, and undergrads, grads, and postdocs by their professors. What I have not seen is the observation that this difficulty is pervasive, nor an exploration of its implications.
  I fuzzily recall a story that Feynman once observed the Brazilian(?) education system was producing postdocs who seemed to understand physics, but who were unable to do research, the quality of their (rote) understanding being insufficient. US postdocs can do research, but their professors remark on pervasive problems with their quality of understanding (mostly "integrationist" defects). With the observed failures of science education, and the common sense observation that one's ability to teach something is tightly coupled with one's understanding of it, I suggest a connection.
  I'd like to line up the ducks, do a narrow-topic worked example, and forcefully deliver this meme.
• This SITE Up: Earth globes oriented so the local site is on top. So connection between globe and planet is merely one of scale and offset. Which, at least for some people, makes the relationship much more tangible. And the Sun's shadow on globe matches that on planet. :)
• Earth puddle: The Earth is a puddle. A gravitationally confined blob of liquid. It just happens that puddles in space are balls.

• Using C++ libraries from C - automatic glue creation: Parse C++ headers and generate extern "C" glue functions. So you can use C++ libraries and code from C. Currently uses gccxml.
• A programmatic code-creation environment: Instead of the common direct-manipulation approach to generating code (editors), I'd like a deeply programmatic one which permits aggressive and graceful delegation to the development environment. Every few years I take a stab at it. With the last one, I could actually lazily stitch together a semantic net describing a C program, shake, and out would pop a distribution tar-ball. Time to try again. The obstacle is, as always, broken languages making the simple, complex.
• Common Embedding/Extending C API: Various languages have C API's to allow extending them using C, and some to allow embedding them in C programs. The API's are all different. I am so very tired of language balkanization. Limited sampling suggests most extensions (of Perl, Python, Ruby, OCaml) use a common subset of functionality. So I'd like to package it up. I don't expect anyone but me to actually want to use it, but oh well. (Anyone disagree...?:)
• Inline::OCaml: Include OCaml code in perl source.
• Wrap Light-Weight C++ (lwc) as a library: LWC is a C++-subset to C translator/compiler. I'd like to try using it with libtcc (an in memory C compiler), to do diskless C++ runtime compilation and linking. Apparently wrapping LWC will require intercepting its "program termination will free things" calls to malloc/free. Little work has been done on this "project" (so I'm not sure it really belongs on this list).
• For the aggressive multilingual programmer: Some webpages of links, recipes, and commentaries...
  There are times for the disciplined, professional craftsmanship of software engineering. But there are also times (weekends:) where, in outraged frustration, one wishes to desperately pretend our field has reached the point where one can actually get work done. A time when one can mixin whatever language is optimal to a subtask. Can use most of the software our balkanized community creates. C, C++, Lua, Perl, Python, Ruby, Tcl, JavaScript, OCaml, Haskell, Prolog, Mercury, Erlang, PostScript, Java, CL, APL, dotNET. Can use closures in C (eg, ffcall), generics in C++ (Cmm), etc, etc. A time of source code analysis, transformation, and generation tools. When one can offload the pointless crud of programming. A time when code flows like water, with the viscosity of wind. ;) We are getting there. If ever so sloooowwwwwlllyyy. It only took what, a human generation, to combine Objective-C and C++? Sigh. And, well, you get the point. The pieces are coming together. They are fragments with sharp edges. But a lot of the fragments now exist. They often don't work or mesh well enough to use in production. But enough to play... and to pretend.

• plpython, aka Language::Python::InPerl: A python to perl source code transliterator. You write python directly in your perl source file, and it gets translated at runtime into perl. Handles the difficult things, including iterators, exceptions, and threads (sortof). There didn't seem to be any interest, so I didn't finish it. I never ran pybench.
• A language feature proselytization page. It is a source of some joy that I will never again have to have the "garbage collection is a good idea" argument. And there continues to be a warm and fuzzy across-the-board progress. But still, some desirable features are not yet widely appreciated, even among language weenies. The sooner they are, the sooner their absence can stop complexifying my life. :) Though, this project has been on my background todo list for a decade, so don't hold your breath. (Probably doesn't belong on this "current" project list, but I periodically wist on it.)
• Beyond SourceForge's viewCVS - "viewCode": It would be nice to be able to rapidly browse large volumes of open-source source code, aided by a zoo of code understanding and analysis tools. This could both extract additional value from open-source, and by providing an a high-exposure stage for such tools, thus perhaps helping to get the tools field unstuck.
• Raw C code parser: I've a parser for unpreprocessed C code. A mutant but readable yacc-ish grammar is munged by a perl script into Prolog which generates an AST in prolog or perl or ... (XML, etc). Contains common preprocessor conventions, with the intent that users will add their own weird ones as needed. Intended for bulk code analysis and transformation. The project is unfinished. Little semantic processing has been done. I'm looking for a driving application to motivate finishing the tool. A "refactoring browser" backend is one possibility. Maintaining kernel structure guards is another. Ideas?

Notes:

Doables:
 Add more current projects.
 Characterize other activities?  Self?
 Mention "rat in pane", a gui comment.
  It seems obvious, but google/dogpile don't list it.

History:
 2004-Feb-19  Added tcl/C/C++ multifile.
 2004-Feb-17  Added "Combined ToC".
 2004-Feb-17  A bit of organizing and dusting.
 2003-Nov-30  Added "Getting to know the Joule", direct link to tr list.
 2003-Nov-17  Added, and then removed, a page on a now-fixed specrend bug.
 2003-Sep-01  Added plpyrb source hack.
 2003-Aug-11  Added FFI patch.
 2003-Jul-08  Added Perlman's thesis.
 2003-Mar-21  Cloned afghan_map to create iraq_map.  Quick hack.
 2003-Feb-04  Re-added `Exp.Not. Meta'.  Seems to have been dropped over the years.
 2003-Jan-16  Added `Year in Scripting..'.
 2002-Sep-30  Added `Inline::SWIG'.  Changed email address.  Trimmed active list.
 2002-Sep-23  Added `Inline/wxPerl' post.
 2002-Sep-07  Added `rulers'.  Bumped FermiQ's to Active, given recent work.
 2002-Jul-23  Added `Make A 10 Centimeter Cube'.  `Exp Meta' off active list.
 2002-Jun-23  Added `interactive F=ma'.  Downgraded `language proselytization'.
              Missed spring astronomy day.
 2002-Apr-11  Pulled over old TIAC homepage as site goes away.
 2002-Mar-12  A light edit.  Gasp.  "Under construction" moved to end.
              After almost five years.  Yipes.
              Fiddled with so-called "current projects" list.  Sigh.
 2002-Mar-05  Added Inline page.
 2001-Dec-08  Reflected title change of 1740 page.
 2001-Nov-16  Added planet globes.
 2001-Nov-05  Added a sketch of some current projects.
 2001-Oct-21  Added `Freq Alloc Chrt'.
 2001-Oct-05  Added `Afghan map'.
 2001-Aug-14  Added `Angle Cards' and `chart'.
 2001-Jul-19  Added Russell quote.
 2001-Jul-09  Added `A BtYacc patch page'.
 2001-Jun-04  Added `What color is a blackbody?' and `are stars?'.
 2001-Feb-16  Added `How Big Are Things?'.
 2000-May-31  Added `yardstrips'.
 2000-May-27  Added `Treasurer's Acct'.
 1999.Mar.23  Added `Some C Language miscellanea'.
 1999.Jan.26  Added `C Language - Grammars'.
 1998.Oct.15  Added `Monty Hall Problem'.
              Upsized Musings.  Removed times from History.
 1998.Sep.22  Added `maze of twisty passages'.
 1998.Jul.08  Reshuffled links.
 1998.Jun.06  Added Cosmic View.
 1998.Mar.15  Relocated some http://www.tiac.net/users/mcharity/ pages to here.
 1998.Feb.25  Added China is down over there....
 1997.Dec.23  Added MisDesign.
 1997.Aug.18  Added link to an old homepage.
 1997.Jul.12  Added link to An Exponential Notation Meta Page.
 1997.May.29  Scooter just picked up.  Ah well.  So added note.

Time slips by:
 2004-Feb-17 - It seems people are now reading this page (not just spiders).
 So it's been dusted a bit... after almost 7 years.  And the following removed. :)

     Under construction.
     AltaVista's "scooter" just picked up this page (29 May 97), so some comment is in order.
     I do not have a current personal home page.
     This is the closest approximation. I plan to fleshed it out at some point.
     Hasn't happened yet (1998.Jul.08).  Still (2001-Feb-16).