[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Math Arrows and Harpoons



david carlisle says,
    If you say so, then I am sure that acurately reflects the current
    unicode rules but if so, then at least as far as Mathematics is
    concerned the rules are just plain wrong.

<smile>  no argument here!

    ... If you have to assign spurious
    meanings to characters in order to slip this stuff past the unicode
    consortium, then I suppose you have to, but you can not expect that the
    symbols will then be used that way in real mathematical documents.

true and expected.  just trying to do the best possible job under the
circustances.

    You accept that `A' does not mean the same thing in all natural
    languages. Unicode doesn't call it `the indefinite article', it calls it
    `latin capital A' (or some such name). Why then insist that in all
    mathematical contexts |-> should mean `maps to' ?

not insisting.  but i *am* trying to establish a reasonably short
"tex" name for any symbol not already there -- and using the names
that are already there if i know something exists; if anyone can
point out a name that exists in association with a font available
on ctan, where i've used some other name, please do so.  please
bear in mind the precedence of my sources:
 - the texbook
 - the amsfonts user's guide
 - appendix d from the ams mathfile user's guide (out of print,
   but a record of actual use in math reviews)
 - the latex manual
 - st. mary's road documentation
 - invention, taking into account experience in discussing with dek
   the naming conventions to be used for the amsfonts

david again:
    Actually though I am slightly confused as when later Hans made
    the (good, I think) point:

        Suppose one only uses naming by
        rendering: Then the original "maps to" |-    might be named
        "rightheadleftvertical", and |=    could be named
        "rightheaddoublestemleftvertical", or something.

    You said

    > here i agree.

    But this type of naming scheme would explicitly avoid the need to assign
    mathematical meanings to these arrows, which is why it would be a good
    thing, but it seems to contradict your rule above.

i agree with that.  however, these particular examples lead to names
that are unwieldy and error-prone.  i don't insist on "[M|m]aps";
maybe "[B|b]ar" at the beginning, or "bar" at the end with the initial
letter cap'ed, would be better?  consider that when i set up the first
version of the multi-source table for the stix project, i allowed the
"tex name" field to be 15 positions wide; it's now 24 wide, and one
name has had to be truncated (\leftrightarrowtriangle) as too big,
but it's also too big a problem to change the width of the column in
the table because of various programs written to operate on it.  (i
wasn't the author; it was patrick ion for some of them, and some folks
from the unicode technical committee for others; sebastian also got
into the act.  those are the ones i know about.)  consider that i have
also tried to create sgml entity names -- and those are limited to 8
or fewer positions, 6 preferred.  that's *too* short in many cases for
scrutability, never mind clarity.  the ideal, seems to me, is somewhere
in between -- and i cheerfully accept suggestions.

    Ah unless I am to interpret the above `rule' to mean that unicode will
    only accept the character if it has been used, at least once, in a
    published source, in which it had a definite meaning. That would be
    reasonable I suppose, and would be a rule designed to stop Taco
    arbitrarily rotating the arrows by fractions of a degree to fill up
    an entire plane of unicode....

my information from the utc members who have been assigned to work with
me on the stix submission is that, if a symbol in one orientation can
be sufficiently documented, and other similar symbols already appear in
unicode not only in that orientation but others, then the new symbol
should be accepted in multiple orientations even though only one has
been fully documented.  two utc members in particular are my gurus here,
and they are being very, very helpful; it is my impression that once i
have convinced them, they will not only support my presentation to the
utc, but will take up the ball and run with it if i falter.  so what i
am now engaged in doing is acquiring documentation to convince them.

    On the other subject Hans raised. on the angle of the arrow head on
    ~-> and friends. I'd agree with him that they should be the same
    direction as on straight arrows. I have often needed a variant arrow with
    a wavy stem, but never seen or needed one in which the heads point off
    in some arbitrary angle. (I'll try to find some real printed refs for you
    later)

i don't question this, and am coming to the conclusion that these are
in fact two different kinds of arrows.  if you can come up with real
references, i will be happy to add them to the submission.  (of course,
i'd love to have documentation showing *both* arrows in context, as
evidence that they have different meanings ...)
							-- bb