Extensive use of abbreviations and acronyms was primarily a convenience for writers, when writing was done by hand and then by typewriter, there is less justification for it now when most writing is done by computer.

I don't agree. My impression is that the popularity of abbreviations and acronyms is being driven by the rise in popularity of text messaging, which is usually done from phones with tiny, unusable keyboards; instant messaging, which is done in real time; and social networking site based communication, which often has hard limits on message length (e.g. Twitter).

I disagree. I seem to love acronyms, mainly aesthetically for no reason at all other than a probably rare quirk of my brain. So upvoted for reminding me others might not feel the same way.

They can be a small, unnecessary barrier for newcomers to LW, LessWrong.

Extensive use of abbreviations and acronyms was primarily a convenience for writers, when writing was done by hand and then by typewriter, there is less justification for it now when most writing is done by computer.

This is a claim I don't understand. Most people with computers can't type much faster than they can write.

Incidentally, there are much worse examples than RMS. For example, FLT is Fermat's Last Theorem, Fermat's Little Theorem, and Faster Than Light Travel. Note that the first two of these are in the same area of study and only have a one word difference.

All of that said, I don't think this is really a big deal. Humans do context recognition really well. A lot of language is much more ambiguous than it would seem at first glance. It is very rare that acronyms create actual confusion.

Why not adopt the convention used in many types of writing? The first time the term is used in a text, it is written in full and its abbreviation or acronym is put after it in brackets. After that the short form is used.

I posted this in the comments at ESR's blog:

Fixed that for you.

there is less justification for it now when most writing is done by computer.

Ah, but most reading is still done by humans, and humans are good at recognizing small identifiers and mapping them to large concepts. That's what words are, and acronyms are really halfway-words (with a few, like laser, being promoted to real words mostly by virtue of their pronounceability and resemblance to other words).

Unfamiliar acronyms are difficult to parse, just like unfamiliar words. But familiar acronyms enhance readability. This is especially true when dealing with things that have clunky True Names, like many organizations (everyone refers to International Business Machines and the National Aeronautics and Space Administration by their acronyms), or in technical fields where many complicated concepts have to be referred to in quick succession.

In my line of work, I'm not averse to typing out "member function template of a class or class template" when necessary, because sometimes that level of specificity is necessary and I don't have a better name (nobody attempts to make it into an acronym). But I frequently refer to the STL (Standard Template Library), CRT (C Run-Time [Library]), RVO (Return Value Optimization), NRVO (Named Return Value Optimization), ADL (Argument-Dependent Lookup), RAII (Resource Acquisition Is Initialization), ODR (One Definition Rule), and the big dog, SFINAE (Substitution Failure Is Not An Error; always hilariously pronounced, sometimes "sss fee nay"). I could write those things out in full (and I do when introducing people to the hall of mysteries), but it would actually make things much harder to read. My intended audience knows what these things are, and isn't going to confuse them with St. Louis, the Chinese Remainder Theorem, etc.

IAWYC, but ISTM that it's not that much easier to type out the whole phrase when using computers, unless you've got a lot of custom macros going. And in the LCPW, context should be enough for acronyms, though naturally YMMV.

Essentially, YSITTBIDWTCIYSTEIWEWITTAW.

ETA: OK, that really doesn't work here, but I couldn't resist.

This post FTW. Runs away

It is pretty interesting that everyone addressed my secondary points while ignoring the more important:

There are too few alternatives, so even though there may not be much confusion in context, the first thing that springs to mind is not likely to be correct, which brings you to a screeching halt while you think about it – like a confusing mis-spelling of a word.

As I wrote, the problem isn't confusion in context; it is that most short acronyms will interfere with your reception of the information while you parse it - unless you are deeply into the narrow context being discussed. Also there will be less difficulty if there are no similar acronyms from fields you are more familiar with.

And there is no evidence that texting has anything to do with it, it seems to be more driven by the "baffle them with bullshit" mode of bureaucracy, since it has been mostly spreading outwards from gov't since the 1970s.

Easy measure: make all acronyms at least four letters long. I pick the number four because three-letter acronyms usually have many associated Wikipedia articles, while four-letter acronyms usually have at most one.

Amusingly, the obvious (in my mind) abbreviation for "four-letter acronym" conflicts with a certain free lossless audio codec.

I am really tired of three letter acronyms. There are too few alternatives,

That's why there are ETLAs: http://en.wikipedia.org/wiki/Extended_three_letter_acronym

EPR: Electron paramagnetic resonance spectroscopy or Einstein–Podolsky–Rosen paradox ?

We had one of these confusions at the Mountain View meetup this week. NLP: Natural Language Processing or Neuro-Linguistic Programming?

Nonetheless, humans are overall pretty good at dealing with linguistic ambiguities — enough so that we usually find them more groan-inducing (as puns) than actually confusing.