Turns out the product goes back over 10 years. I inherited it in 2001, when I started in my current job at Cray. There was a time when Cray was owned by SGI, and during that time the documentation server was not known as CrayDoc — it was Dynaweb, a third party product. But starting in 2002, with the shipment of CrayDoc 1.0 (though I guess that versioning is erroneous, now that I know the history), CrayDoc is again a home-grown product. Versions 1 and 2 were 100% Perl, though now we use the SWISH-E search engine, which is written in C.
My work on CrayDoc has been a real education in CGI programming, HTML, databases, and code design. When the presentation is done, I’ll post it here for posterity.
UPDATE 11/22/2004: posted here PDF