Aside about jns' justify; while I could build the new version like
CXX=clang++ gmake
afterwards
LANG=ascii ./justify -i foo.txt -o j.foo.txt
no longer works for me; sorry I didn't investigate it. Using jns' justify
classic then without my former paragraph-separation strategy.

I had wanted  to be done part two of my acl2 just-like-make-game.   But  you
would never guess - I have had some difficulty  proving measure  conjectures
for   good   complexity    theorems   of  my   game's   applicative    media
functionality...!   The problem isn't hard but you know how things  that are
not hard can still be frustrating sometimes.  Have a distraction:  
There are
two tall temporary  fences obstructing  the raised, wooden  ramp down to the
riverside  I am haunting recently.  The wire fences say "construction"   (no
other evidence of construction). The entry side has a gap to squeeze through
though   the lower  side does not - if you scramble  over and cling  to  the
outside  of the railing then jump outwards with sufficient  vigor, the slope
means  it is not so far down.  If not construction,  the ramp  is absolutely
swarming with paper wasps in that way fresh, lightly salted wood often is. I
feel like wasps  get a bad rap.  My house mate hates them,  but likes  bees.
That's like saying coyotes  are evil wolves (if wolves made honey and didn't
hunt?   This metaphor  is flimsy).   When you kill off wasps  in residential
places,      you      suffer      20      times      the      volume      in
things-the-wasps-would-have-preyed-upon  (not just wooden ramps).  At  least
it's complicated.  
I read jynx's second-to-recent post Lies, Damned Lies, ..
recently.  That's me to a t as well.  There seems to be a bit of demographic
undercurrent  like us (sez me). You also meet the almosts, or the incipients
who bravely weather casual dismissal  to hear an odd verse not pulled from a
corporate  advertising jingle.   The most sly counter-narrative  I know goes
like this:   These dissenters  have noticed  a few tens of cases  where  bad
actors really are bad acting, then threw out the reasonable  actors who were
mostly minding their own business with the bad bathwater.  
Computingwise,  I
am reasoning  myself out of reasoning myself into reasoning  with case based
reasoning.   I collected a few articles and some source.  Riesbeck  seems to
have been a lisp hacker.  I will read one of his books. Some of the people I
know who research around this are alright, though I am deeply mistrustful of
ex-90s  high budget  business friendly  java software.   I have also already
opted  for Gabor's  MGL sofaras grab bags of linear algebra  and statistical
decisionmaking    (learning/deep    learning  in  this   case)   go.    More
controversially, and we will see what Riesbeck convinces me of, it annoys me
a lot when people  hackily add a notion of inheritance  to structs  and talk
about how fast they are (I'm looking at you too, golangphers)  and then do a
notion of searching upon them. I feel like this is eerily close to what ANSI
Common  Lisp's  CLOS method  combination  was already  doing in some  deeply
controversial  opinion  sense. I haven't finished  fathoming  my way through
that.  Lastly,  if case based reasoning is a heuristic  low context decision
maker,  what about- every calculus,  the worlds of research around modelling
nonlinear phase dynamics, ... why would CBR warrant special attention? I did
like Douglas  Adams' book Dirk Gently's Holistic Detective  Agency (which is
about  a software engineer (not the title character)  who works at a company
that got rich off the military  buying some CBR software he wrote).  In real
life I think this was where CBR's funding/business popularity came from too.
Military interest, not Douglas Adams.