Last modified: 19970630
All files on the tapes were examined individually and separate log files in SGML format were maintained for each tape. The DTD for the log files was:
<!doctype filelist [ <!element filelist - - (title,resp,date,note,files)> <!element files - - (filedesc+)> <!element filedesc - - (size,name,contents?,date?,resp?)> <!attlist filedesc salvage (done|todo|dump|unexamined) unexamined> <!element note - - (li)+> <!element li - - (em|#PCDATA)*> <!element (title|resp|size|name|contents|date|em) - - (#PCDATA)> ]>
All GIs should be self-explanatory, except perhaps resp (person responsible for latest report) and li (item in list of notes for tape).
SGML log files were converted to HTML using a custom SNOBOL4 script and Phil Budne's SNOBOL-in-C (C-MAINBOL). The home-grown fast-and-dirty SNOBOL scripts parsed the SGML files without parsing the SGML itself, and therefore required that the latter be formatted consistently. A more robust approach to this conversion would have be based on an SGML parsing engine, such as Omnimark.