I will add the caveat that typesetting accuracy isn't my primary goal although it is
a secondary one. My primary goal is easily diffable sources between versions to then
facilitate version analysis. I intend to take this same approach to plenty of other
documents we have lying around in scanned form to try and get a better machine-readable
body of research material available, including stuff like the Documents for UNIX PWB 1.0
and Release 4.1 sets and the CB-UNIX sources we have bumping around the archive. Part of
my reasoning on putting these in git archives as I go is so that if someone does want to
come do some editing, corrections, etc. they're free to fork or raise a PR on my
repositories.
- Matt G.
------- Original Message -------
On Wednesday, March 8th, 2023 at 12:32 PM, josh <joshnatis0(a)gmail.com> wrote:
Hi Angelo,
G. Branden Robinson, (CCed on this email) attempted a somewhat similar mission,
re-typesetting the paper "Typesetting Mathematics by Kernighan and Cherry"
with
groff. If you look through the email thread detailing the result [1], you can
see notes about aesthetic regressions from the original troff document, and
Branden's attempts to fix them. Hope you don't mind me summoning you, Branden
:).
[1]
https://lists.gnu.org/archive/html/groff/2022-07/msg00000.html
Additionally (though unrelated to roff), the Computerphile youtube channel has
a video [2] you may find interesting titled "Recreating Dennis Ritchie's PhD
Thesis", in which they discuss how they went about making a faithful recreation
of dmr's unsubmitted PhD thesis.
[2]
https://www.youtube.com/watch?v=82TxNejKsng
Hope this generates some interesting discussion? :)
Josh
On Wed, Mar 8, 2023 at 11:30 AM Angelo Papenhoff aap(a)papnet.eu wrote:
> Something I'd like is to recreate the original troff output exactly. I
> used troff from plan 9 (so same lineage as original troff) but something
> causes the output to not look exactly like the original. I don't
> remember what it is exactly but you can easily check by comparing my
> pdfs with the scan. Line lengths, page length, something like that.
>
> I don't know if this is just a troff setting or if troff had changed
> enough to cause this difference. Unfortunately the original troff is
> lost so no way to compare. v7 (or PWB?) is the earliest version of troff
> that's still around. And even then one would need CAT emulation, which I
> haven't bothered with yet.
>
> Cheers,
> Angelo
>
> On 08/03/23, segaloco wrote:
>
> > Ouch....well I'm glad I shared then, I had no idea someone had already
done this....well good to know, I guess I can move on to the CB and MERT manuals then.
> >
> > - Matt G.
> >
> > ------- Original Message -------
> > On Wednesday, March 8th, 2023 at 2:02 AM, Angelo Papenhoff aap(a)papnet.eu
wrote:
> >
> > > I've done this a couple of years ago:
> > >
http://squoze.net/UNIX
> > >
> > > Cheers,
> > > Angelo
> > >
> > > On 08/03/23, segaloco via TUHS wrote:
> > >
> > > > So I decided to keep the momentum and have just finished the first
pass of a Fifth Edition manual restoration based on the same process I used for 3B20 4.1:
> > > >
> > > >
https://gitlab.com/segaloco/v5man
> > > >
> > > > There were a few pages missing from the extant PDF scan, at least as
far as pages that were in both V4 and V6 sources, so those are handled by seeing how V5
source of the few programs compares to V6. I'll note which pages required this in a
second pass.
> > > >
> > > > I've set my sights on V1 and V2 next, using V3's extant
roff sources as a starting point, so more to come.
> > > >
> > > > - Matt G.