paper dist.words headings f77.ms .B good & many done lex.ms .I marginal done yacc.ms .I not useful getital done spin.ms .CW most .IH sam.ms .CW ok commands with ngetcw done - must edit sam.ind to get rid of dbl ref cmds pico.ms .CW ok cmds with ngetcw done ipc.ms .I ok getital done upas.ms .I/.CW only ok done security.ms .CW ok done anim.ms .CW/.UL/.I good withmore too many terms cip.ms history.ms DEFERED uucp admin.ms DEFERED uucp ms/ms none ok .BI SCHEME FAILS - DOESN'T use mpm eqn/guide IT CW good - must edit guide.ind - merge sub/sup up/down etc - out CW to s/guide/eqn ideal/user must edit s/user/ideal/ & resort tbl contains .tr~ - disaster - fixed sed ok must edit cmd dbl reference setup ok edit setup.ind for make reference pic grap UL/IT/UC useless ped sed 's/ hd / /' ped.x before making index mpx mkindex proto.x mpx netb ok embedded \f...\f no caps in terms files - diction can locate - maybe write sed script since it has line number of match .Tm words puts out pg# +words on stdout to get terms: gettop file gethd file getital/getcw/getul >file.dist getdist BI >>file dist maybe sed -f tools/fontch.sed (gets \\f.*\\f) constructs) sort file.top file.hd file.dist|uniq >file.terms terms of form words or words \t[sg] where s=specific - just entry under `name' g=general - just entry in general, not under name no \t entry in both places to get index: runind tr A-Z a-z <file.terms|sort|uniq|awk -f tools/dups.awk >file.lterms diction -n -f file.lterms -l -o file.hits file.ms >/dev/null doinsert base >base.ms prefer ... base.ms 2>file.x mkindex file.x solutions - awk prog to make sure not inserting term in .TS or .PS tr $ to ~ in hits & back again in mkindex remaining editing - index has lost caps, font changes and plurals bugs - .Tm before title is lost .Tm containing troff \stuff is converted - shouldn't be