[bug #65558] [UPGRADE] improve paragraph formatting within groff's line-at-a-time processing

[Dave]

URL:
  <https://savannah.gnu.org/bugs/?65558>

                 Summary: [UPGRADE] improve paragraph formatting within
groff's line-at-a-time processing
                   Group: GNU roff
               Submitter: barx
               Submitted: Thu 04 Apr 2024 11:52:22 PM CDT
                Category: Core
                Severity: 1 - Wish
              Item Group: Feature change
                  Status: None
                 Privacy: Public
             Assigned to: None
             Open/Closed: Open
         Discussion Lock: Any
         Planned Release: None


    _______________________________________________________

Follow-up Comments:


-------------------------------------------------------
Date: Thu 04 Apr 2024 11:52:22 PM CDT By: Dave <barx>
Bug #40716 seeks to bring TeX's Knuth-Plass algorithm to groff.  This is in
progress but is a monumental, and thus slow, task.

The 2014 Groff Mission Statement
(http://www.gnu.org/software/groff/groff-mission-statement.html) offers K-P as
"the most likely candidate" for improving groff's paragraph formatting.  But
discussion that took place around the time this mission statement was being
drafted reveals other contenders.

As an alternative that's simpler to implement, since it doesn't require
changing groff's underlying line-by-line approach, Peter posted this
pseudocode: http://lists.gnu.org/r/groff/2014-03/msg00322.html

(The thread continues in several pieces not linked to each other:
* http://lists.gnu.org/r/groff/2014-04/msg00021.html
* http://lists.gnu.org/r/groff/2014-05/msg00000.html
* http://lists.gnu.org/r/groff/2014-05/msg00010.html
* http://lists.gnu.org/r/groff/2014-05/msg00041.html
* http://lists.gnu.org/r/groff/2014-06/msg00089.html
)

Years later, Peter wrote (http://lists.gnu.org/r/groff/2020-07/msg00101.html):
"I worked on systems that used the formatting strategy I proposed and can
confirm that it significantly reduced the amount of intervention required to
achieve good grey on a line-by-line basis."

Consensus seems to be that K-P provides the best known automated results, but
given how extensive a change that is, in the interim if someone is willing to
turn Peter's pseudocode into working groff code, this could be a substantially
lower hurdle providing quite good results.







    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?65558>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/