[R] [SPAM] Re: The "--slave" option

Richard O'Keefe r@oknz @end|ng |rom gm@||@com
Fri Sep 20 11:07:57 CEST 2019

Not being a jerk is a good thing.
Unthinking political correctness is not the same thing at all.
The point has already been made that the relationship between
a "master" process or cylinder and a "slave" one is intrinsically
a dominance relation where the "master" tells the "slave" what to
do.  No amount of mucking around with audible or written words
will affect the *meaning*.  Even boss/worker is uncomfortably
close to master/slave and is going to trigger anyone who is
triggered by words rather than actual oppression (such as having
someone hostile attempt to control your speech, and intrinsically
oppressive act which presumes that the would-be controller has
some sort of *right* to dominate the potential controllee).

If we ever hear of someone using R who is or was a slave, we'll
ask for *their* opinion on the matter.

It has also been noted that euphemisms erode surprisingly quickly.
In my daughters' generation, the euphemism "toilet" for the jakes
is being replaced by "bathroom" -- with the result that I never
know which room they're going too, we have to use the ridiculous
"bathroom bathroom" to indicate the place where you wash yourself --
and I'm sure another term for the necessary house will be along
soon.  In fact, we're already starting to say "wharepaku" (Foh-
Re-Pah-Koo) for clarity.  Dear knows what that will be replaced
by.  So shunning the word "slave" is not only ineffective, it
won't even be ineffective for long; whatever replaces it will
itself be the target of cries of "being a jerk".

The really annoying thing about this is that it does nothing whatsoever
to improve the actual condition of any living person.

As for NULL, NA, and NaN, you really cannot blame R for NaN.
S existed before IEEE arithmetic, and it's not *that* hard to
keep NULL and NA apart.  length(NULL) is 0.  There is literally
nothing there.  It's what c() gives you.  length(NA) is 1. There is
a place for something but it was missing or undefined.
IEEE arithmetic introduced NaN, which means "there should be something
here but your calculation went wrong, the *mathematical* result might
be defined but this machine cannot compute it."
Frankly, IEEE arithmetic made a lot of things simpler, but a lot of
other things more complicated.  You should see what it did to
floating-point comparison in C.

As for the syntax of S, it comes from the same organisation as C.
Having used GLIM, GENSTSTAT, SPSS, and BASIS (don't ask), S was a
revelation that statistics environments did not have to be arcane.

As for the documentation of R, it's pretty much the best of any open
source programming/statistics tool I've used, and most impressively,
has a huge library of packages whose authors have also produced
comparatively good documentation, but the standards of free software.

Run time error messages?  Yes, they could be improved.  Quite a bit.
You've probably heard about the famous Multics error message.  One
day, on startup at a Multics site, the machine wrote
   Hodie natus est radici frater.
You probably haven't heard about the Burroughs B6700 ESPOL compiler.
One of its error messages was
It's an open source project.  If you think an error message could be
improved, you can patch your copy and send the patch to the maintainer.
The following sentence was written by a biostatistician:
"Help files are frequently more than a little obscure."
That was written about a commercial package, not R.

SPSS costs NZD 156/month for one user, or about NZD 1871/year.
SAS prices are scary.  I don't know what the licence terms for the
free-as-in-beer "University Edition" are; since I'm no longer at a
university I suspect I wouldn't qualify.
I rather liked GLIM, but it's dead, and the syntax was idiosyncratic.
I also rather liked GENSTAT, but when you look for the price and are
invited to "ask for a quote", my "I-can't-afford-this" alarm goes off.
Plus the pages I viewed are for Windows only.

Thing is, for the price of one year for one user on one machine,
you could fund quite a bit of error message improvement for R.

As for slavery and the Bible, what's translated "slavery" in the OT is
not the "chattel slavery" that was practiced in the 19th century.
Slavery was banned in England by William the Conqueror, of all
people, and the first "official" ban on slavery by any religion that I
ever heard of was when the Vatican ruled that the native
inhabitants of the Americas could not be made slaves.  Many of
the Abolitionists derived their opposition to chattel slavery from
the Bible.  The practice of African slavery was, as a matter of
history, learned from another religion entirely, whose prophet
bought and sold black slaves himself.

Serfdom is close to slavery.  ("Serf" comes from "servus".)
describes the way salt-workers and colliers were bound to the
land in Scotland, so that 'And thus it came about that the nineteenth century
had dawned before it could be said in truth of Scotland, in the words
of Cowper:—
There are no slaves at home : then why abroad?"'
These were native Scots.

Enslavement is the vile extreme of the desire to control other people.
Let those who oppose slavery oppose that desire whatever form it takes.
(Want to read about man's inhumanity to man?  Read "The Gulag Archipelago.")

Can we perhaps return to helping each other with the use of R?

More information about the R-help mailing list