A Little Spring Scripting
-------------------------

Spring is here, and all week long the winds were stripping
cherry blossoms from the trees; a pink blizzard blowing past
my office window.

April is exam month and already the campus is notably
quieter than before, though the Library is still packed with
students hunkered over their laptops (and the occasional
book).  My administrative work has somewhat lightened up and
there are more meeting-free gaps in my calendar. All signs
pointing to the imminent arrival of summer session, when
things get really quiet, and I shift into project mode.

This week I was able to set aside a few hours to do a little
scripting. In Perl, presumably because I'm a crusty
irascible cantakerous old person full of stubborn ideas[1].
Well, it was the first scripting language I ever learned,
when it was the New Hotness back in the mid-90s.  And while
poor Perl gets a bad rap for readability, not only has it
served me well ever since, it has saved me from having to
learn other, nuttier languages where even the most trivial
spaces and tabs are Significant. ;-)

It was pleasant to do some technical work for a change,
though I couldn't help wondering whether my scripting skills
(never better than adequate) had perhaps deteriorated
somewhat from lack of use.

Nevertheless, the problem seemed simple enough.

Last fall, the Library installed some time-of-flight
people-counting sensors over our main entrance.  These post
hourly reports of total entrances and exits via https to a
PHP script that updates a database on one of our servers,
from which our Data Analyst in turn produces nice reports in
Power BI.  However, at a certain point the sensors ceased
communicating with the PHP script[2], and it took a while
before someone noticed. Like, a while as in three months,
which is a long gap in the data that's supposed to help
justify why the Library needs to occupy as much space on
campus as we do.[3]

Fortunately, the sensors keep internal logs of their
activity going back all the way to when we first installed
them.  So, all I had to do was download the sensor logs and
load the missing data into the database.  Simple, right?
Shouldn't take more than a couple of hours, I thought.

Except that that sensor logs recorded every exit and entry
at the time of the event, and what I wanted was an hourly
summary.  Including hours when no sensor events were
recorded, because the Library was closed.  And allowing for
the fact the sensors reset the counts to zero once per day,
but did not adjust the reset time for daylight savings ...

Anyhow, I won't recite further my Boring Litany of Problems,
but suffice to say my initial two-hour time estimate was a
tad optimistic. And on reflection, I think it's not so much
my scripting that has grown rusty from lack of use, but
rather a certain kind of problem-solving ability.  Writing
the 150 lines of Perl wasn't that hard.  Figuring out how to
reliably convert the logs into hourly reports was what took
the time.

I can't shake the feeling it wouldn't have taken me quite so
long when I was in my programming prime.  Back then, I could
hyper-focus on these kinds of problems in a way that
increasingly eludes me now, as ever more often my gaze is
drawn from the monitor, to watch the petals drift in the
wind.



References and Notes
--------------------

[1] http://www.wordnik.com/words/curmudgeon

[2] Not coincidentally, the very same day the server was
moved behind a load balancer.

[3] Yes, I should have set up a cron job to monitor the
database and alert me if it wasn't getting updates, but I
honestly thought someone else was paying more attention to
it than it turned out they were.  Still, my bad. 

Sat Apr 12 17:21:44 PDT 2025