Re: keyboard entry

From: Markus Kuhn (
Date: Wed Jul 28 1999 - 07:06:42 EDT

Jonathan Rosenne wrote on 1999-07-28 03:23 UTC:
> How will mathematical symbols be entered in plain text?

For example:

You press the compose key, followed by some mnemonic that will probably
be rather similar to the TeX or SGML mnemonics (TeX's choice is usually
better because it's shorter and more pragmatic, something committees
outgrowth are not very good at sometimes). However, you convert this
within the keyboard driver into Unicode already. The application only
sees the Unicode values. The mnemonics are applicable across
applications, because they are provided by the keyboard driver or some
comparable system-level entity.

There were several discussions within CEN to standardize something
called UTF-mnemonics, but they all led to nowhere so far. The problems
in these projects is that people try to look for mnemonic collections
that try to cover everything, and this is in my opinion bound to fail.
It would already be nice to have standardized compose key mnemonics for
just the few hundred most widely used non-keyboard symbols. This would
cover all the CP1252 repertoire, and a fraction of the mathematical
characters. The remaining characters either are so rarely used that you
can enter them easily via cut&paste, table selection, hex entry, or in
the case of entering entire sentences in a different script by switching
into a different keyboard mode.

I want to be able to type say

  <Compose> <-> < > = MINUS
  <Compose> <-> <-> = EN DASH
  <Compose> <_> <_> = EM DASH
  <Compose> <`> < > = LEFT SINGLE QUOTATION MARK
  <Compose> <`> <`> = LEFT DOUBLE QUOTATION MARK
  <Compose> <+> < > = DAGGER
  <Compose> <+> <+> = DOUBLE DAGGER
  <Compose> <^> <=> = ESTIMATES (corresponds to)
  <Compose> < > <v> = LOGICAL OR
  <Compose> <!> <v> = LOGICAL AND
  <Compose> <i> <n> = ELEMENT OF
  <Compose> <s> <b> = SUBSET
  <Compose> <s> <p> = SUPERSET
  <Compose> <i> <y> = INFINITY
  <Compose> <<> <=> = LESS-THAN OR EQUAL
  <Compose> <>> <=> = GREATER-THAN OR EQUAL
  <Compose> <=> <=> = IDENTICAL
  <Compose> <c> <=> = EURO SIGN (also on AltGr + e)

etc. (example assumes Sun practice of having all compose sequences two
characters long, other sequence terminators are imaginable). I am
confident that I can memorize the dozen or so characters that I need
frequently in a few minutes this way. For more special characters, I can
also memorize the hex sequence.


Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK
Email: mkuhn at,  WWW: <>

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:50 EDT