The Long Now Foundation has produced a Rosetta disk containing 13,000 pages of information regarding 1,500 human languages. The text is engraved, not encoded. The text starts out large enough to read with the naked eye and becomes continuously smaller, strongly suggesting one should examine the disk under magnification to read further.

Long Now is trying to preserve documentation for thousands of years, but I just want to know how to preserve documents even for a few months or years. They want to hold on to knowledge as civilizations come and go. I’m just trying to hold on to knowledge as personnel come and go.

Mundane document preservation is a very difficult problem. Preserving the Declaration of Independence is easy; preserving meeting notes is hard. Preserving the Declaration is a technical problem. If you keep it in a glass case filled with nitrogen, keep the lights low, and make sure Nicolas Cage doesn’t steal it, you’re OK. Millions of people know that the document exists, and they know where to look for it. And besides the original paper copy, the text is available electronically in countless locations.

How do I preserve the document that describes why my internal software application uses the parameters it does? Make notes in the source code? Good idea, but most of the people who want to know about the parameters are not software developers. What about version control systems or content management systems? Great idea: put everything associated with a project in one place. But wherever you put the information, someone has to remember that it exists and know where to look for it.

Greg Wilson has a new blog post Legal Frameworks for Reproducible Research.

The Orfeo Toolbox is an open source library of image processing algorithms developed by the  French Space Agency (CNES).

The tagline for the Orfeo Toolbox project is “Orfeo Toolbox is not a black box.”

I just found out about  coNCePTuaL, a project that promotes reproducible research in the context of performance measurements for high-speed computer networks.  (The capitalized letters in the name stand for Network Correctness and Performance Testing Language.)

The project is located at  http://conceptual.sourceforge.net/ and is described in the paper Reproducible network benchmarks with coNCePTuaL by Scott Pakin of Los Alamos National Laboratory.

Some of the highlights of coNCePTuaL:

• Performance tests (timed network-communication patterns) are described in a precise but English-like “executable pseudocode” designed for basic readability, even by someone not familiar with coNCePTuaL.
• Output files produced by coNCePTuaL-based performance tests include not only the measurements but the code describing the  test itself and a detailed description of the experimental platform on which the code ran.  This enables a third party to see exactly what was run, how it was run, and what the results were, all in one file.
• coNCePTuaL can automatically produce space-time diagrams of the communication pattern for additional clarity of presentation.