HALVORSEN@SU-CSLI.ARPA (09/20/84)
Can I expect "The writer's work bench", a software package marketed by AT&T Technologies, which is supported for version to of Unix system V from Bell to run under Berkley 4.2 Unix without problems. If the answer is no, then what problems should I expect? Thanks, Per-Kristian Halvorsen
bierma@nprdc.ARPA (09/21/84)
We have WWB running under 4.2BSD (with one exception). It wasn't as easy asn the WWB installation manual made it out to be. The manual makes a number of refferences to loading WWB under Berkeley UNIX, most all of which are wrong. Apon calling the folks at AT&T they stated that all the refferences (There were many) were actually misprints, and if we weren't running suported System V they didn't want to talk to us (At which point I told them we wouln't pay unless we could load the tape). Fortunatly the story has a happy ending since we managed to get almost everything up and running. If anyone is trying to bring up WWB I'd be happy to share what was done here to make it work (it's not very difficult). Now, for the one thing that doesn't work. The routing "org" (for analyzing the orginization of a paper) looks hoplesly wired in to using the "mm" macros. Has anyone out there modified or rewritten the lex file to enable "org" to use analyze other macro packages. --Larry APRA: bierma@nprdc.arpa UUCP: {decvax,ucbvax,ihnp4}!sdcsvax!sdcsla!nprdc!bierma PSTN: (619) 225-2161
rad@mitre-bedford.ARPA (09/26/84)
I installed wwb on our BSD4.1 VAX in August. It runs fine, but there are a couple of things you might want to know about when you install it. First, you won't be able to 'tar' it off tape, as the installation guide tells you. At first I tried using 'dd' and editing out the files one by one, but that was a bit much. I found by experimentation that the BSD release, like System III/V, is done in cpio format. The release has some nice utilities for configuring itself to your system (directory locations, etc.). The actual install is pretty easy. One problem you might have is that a couple of routines from the string(3) library are renamed. 'rindex' is 'strrchr', and 'index' is 'strchr'. Put a "-Dstrchr=index" and "-Dstrrchar=rindex" in your makefile CFLAGS line to care of these. (Thanks to Guy Harris: [...!rlgvax!guy] for this timely tip.) One change that I made and that you may too is the substitution of the 'more' command for 'cat' in many of the shell scripts that type out help files. As for the 'org' command (bierma@nprdc's earlier response): the lex code which diddles with your nroff input expects -mm macros. Its output is piped to nroff for formatting with the mm macros. The bug section of the manual page admits that the restriction to -mm is a problem. I found that by changing the org shell to nroff the output of the lex program with the -ms macros, I still got fair results. According to the August Unix Review, the mm macros are a descendant of the ms macros. Apparently, they are close enough for org to do its work. One last point. It was not clear from the advertising literature or from the conflicting reports we received from the people at Western Electric that wwb expects and, in some cases, requires [nt]roff input. We planned to use wwb to analyze documents prepared on another word processing system (word11). It works well enough with most commands on plain text. (Org is not one of these.) However, non-nroff users should not expect full functionality or accuracy. Dick Dramstad rad@mitre-bedford.arpa ...!linus!bccvax!rad