🚀 go-pugleaf

>Sandra Martin writes that there already exist defined keystroke
>sequences "that allow users to create any Latin-2 character with
>an ASCII-only keyboard."

No, she doesn't.  What she writes is

        I know that such sequences exist for Latin-1, but
        I don't know whether anyone has defined them for
        Latin-2.

The sequences exist for Latin-1, but they may or may not exist for
Latin-2.

>I'd love to hear the details -- can anyone tell me what these are?

The way these "compose sequences" work - at least on the systems with
which I'm familar - is that there is some key that introduces the
compose sequence.  You type that key (e.g., on the Sun Type 4 keyboard
it's labelled "compose", on the DEC keyboards it's labelled "Compose
Character") and then you type two (or possibly more) keys.

Some level of {hard|firm|soft}ware recognizes this sequence and
generates an ISO Latin #1, #2, etc. code from them.

A couple of examples from the SunOS keyboard driver (at least on the
'386i; support for Sun-3 and Sun-4 will come in a later release):

	<Compose> r o -> "registered trademark character"
	<Compose> a " -> "a with an umlaut"

>Where do such keystroke conventions come from?

I don't know.  Is there a standard (*de jure* or *de facto*) for them?

>Latin-2 characters are intended to handle Albanian, Czech, German, Hungarian,
>Polish, Rumanian, Serbocroatian (the Latin-alphabet form, of course), Slovak,
>and Slovene.  So basically East Europe.

For what languages are Latin-3 and Latin-4 intended?

>I don't have a machine-readable listing of the 96 upper-half
>characters, so I'm reluctant to do all that typing from the ISO document!

Besides, many systems don't have a way of displaying text in Latin-2, so
there's no guarantee that a machine-readable listing would work on her
machine anyway....  She might want to post a paper-mail address so that
printed copies can be sent.