Gopher File Names                                             03/05/24
----------------------------------------------------------------------
I  really like  the way  entries are  listed in  d1337's chronological
index[1]. Specifically, I  love that they've put a category  or tag of
sorts in  brackets before  the actual title.  Here's an  example entry
from that index, as it is displayed in lynx and in VF-1 respectively:

  (FILE) 2024-02-27: [assorted] Gopher on a Commodore 64

  [1] 2024-02-27: [assorted] Gopher on a Commodore 64

The listing is programatically generated,  it seems to me, rather than
just a list of files in a  folder. I'm not judging, but noting that as
I use a gopher file that  primarily lists files. The source of d1337's
listing looks like this (imagine it on one line if you would):

  02024-02-27: [assorted] Gopher on a Commodore 64\
  <tab>/users/d1337/assorted/c64-001.txt\
  <tab>sdf.org\
  <tab>70

My gophermap phlog listings are generated with these lines:

  =ls -1p | grep '^0_' | awk '{printf "0%s\t%s\n", $1, $1}'
  =ls -1p | grep '/' | awk '{printf "1%s\t%s\n", substr($1, 1, length($1)-1), $1}'
  =ls -1pr | grep -v -E 'gophermap|/|^0_' | awk '{printf "0%s\t%s\n", $1, $1}'

And so, they rely  on the file names themselves to  make sense for the
gopher viewer.

Getting back  to what  I was appreciating  about d1337's  listing: the
bracketed category/tag. It  adds something to the title,  a quick sort
of  organization with  very little  overhead. Take  a look  at d1337's
listing in  a browser to  see what I mean.  You can visually  pick out
things of interest very rapidly. It's  almost like having an icon, but
not  quite  as powerful  (not  to  offend the  plain-text-is-beautiful
crowd,  of which  I am  a  part, but  icons  are older  than ASCII  by
thousands of years at least, and are beautiful too.)

My current file name format is something like this:

  2023-10-31_anonymousGopherUpdate.txt

Even though spaces and other characters  have been a part of Unix file
names from  the start, my DOS  background makes them feel  icky to me,
and so  I avoid them.  People have probably already  written competing
novel-length  paper  on the  subject,  and  I  won't bother  with  the
argument, it's just what I  prefer. But those categories/tags... those
could fit. And, I might even be  fine with those brackets, as they are
a sort of minimal graphical tool.

I don't type my filenames manually, even though they are basic. I have
a bash script that  does a few other things for me  when I'm writing a
phlog post,  and so I just  rolled the filename generation  into that.
The particular line in question is this one:

  text_file=$(date +%F)_$1.txt

Not glamorous, but it  works. And as you can quickly  see, adding in a
"[category]" would be  trivial! It could be passed as  $1 and the file
name could be $2.

The only issue is... most of my posts are meandering thoughts.  Still, 
this post could be [technology],  which others could be [retrotech] or 
[society]  or  [politics] or whatever  else.  The trick is, I suppose, 
overcoming two  issues: First,  the temptation  to let  the categories
define my writing habits; I have  no desire to write about things just
to make  my category balance look  pretty, but I am  *tempted* by that
sort of  thing. The second  issue would be  the temptation to  have so
many categories that it's worthless to have categories at all.

In short,  if I could keep  the pendulum in the  middle somewhere, the
bracket categories  could be quite  useful and pleasant. I  think I'll
try it out. My format will probably look something like:

  2023-10-31_[gohper]_anonymousGopherUpdate.txt

And  I'll probably  retroactively add  categories/tags to  my existing 
posts (since I love breaking old links, apparently).

[1] gopher://sdf.org:70/1/users/d1337/chronological_index