'lerfu' glosses

[C_Burke%SEMPERFI@MWMGATE1.MITRE.ORG]

>Lojbab writes (although I have not seen the original question):

>> A lerfu is meant to refer to what we in English call "letters" symbols on
>>a page or other medium, but generalized to include numbers and other
>>symbolic 'marks'.  The general term for any sign or symbol is "sinxa" and
>>is not limited to visual effects or to simple ones.
>>
>Would 'characters' be a better gloss?

My two cents:
++++++++++>
'glyph' might work even better, if only the 'mark' aspect is of interest;
'character' has semantic implications, if I followed a recent discussion on
the SGML TEI newsgroup correctly.
>+++++++++++

Chris Handley comments:
++++++++++>
I like 'glyph' BTW - I think it would solve many problems.
>+++++++++++

Lojbab remarks:
++++++++++>
 The gloss for lerfu is "letterals" which is a Loglan/Lojban jargin (sic) term
from
the JCB era.  The long definition is something like x1 is a letter/digit/symbol
in alphabet/character set x2 meaning x3.
>+++++++++++

To which Colin Fine replies:
++++++++++>
I've always thought it a bit daft glossing a lojban word as a bit of
non-English jargon, especially one not known outside the Loglan world.
>+++++++++++

When I had a chance to grok this thread in fullness, I can see problems
with all of the glosses.  'lerfu' relates a specific glyph x1, from a named
or indicated collection of glyphs x2, with a semantic interpretation x3.
The x2 place is ambiguous; does it refer to an 'alphabet' (Roman, Cyrillic,
Kanji, "Ogham") or to a specific encoding scheme (ASCII, EBCDIC, UniCode)?

How many of these interpretations are correct?:
"lerfu('A', ASCII, <letter A, a vowel>)" -- omitting ASCII encoding
"lerfu('A', ASCII, 0x41)" -- omitting other interpretation
"lerfu(0x41, ASCII, <letter A>)" -- but then x1 is encoding, not glyph

You could blur x1 to mean the glyph, the character, or (possibly) the encoding,
and build tanru/lujvo if you need more specificity.  This should work in many
cases;
most people don't worry about the distinction between "short horizontal bar
centered
in the character cell", "hyphen", "minus", "Em-Dash", and "En-Dash".  If you
take this very context-sensitive approach, the other sumti have broad meanings
to be determined from context as well.

Using 'glyph' as the gloss implies that x1 MUST refer to a graphic mark,
and makes the x3 place an extension (glyphs have no inherent meaning).
'Character' has numerous other meanings in English which can cause
confusion, but does (sloppily) incorporate the idea of binding a graphic
symbol to an encoding.  'Letteral' is cute, possibly a portmanteau from
'letter' and 'literal', but it's obscure jargon and makes a bad gloss
unless the word has been explicitly defined (and used often!) before
it's encountered in the gismu list.

Doing my part to fill your inbaskets, I am...
Carl Burke