[UEB Maths] Two queries from Bruce and Janet's meeting in Sydney

Joe Sullivan uebmaths@nbp.org
Fri, 27 May 2005 11:32:45 -0400


--=====================_8011870==.ALT
Content-Type: text/plain; charset="us-ascii"; format=flowed

Hello Janet and everyone,

The idea that "shortcuts" can easily be taken in UEB has been discussed 
from time to time, and might even be considered one of the minor design 
principles. We have figured that, in writing for their own use (as when 
taking notes in class), people will often drop the dot 5 from the common 
math operators and not bother with explicit grade switches. All of that 
seems good and natural and consistent with what we all do when writing 
informally, no matter what writing system we use. However, I'm worried 
about trying to make such shortcuts part of the formal system, lest in the 
process we abandon some of UEB's strengths.

Using a skipped line to signal a switch into or out of CBC was part of the 
original CBC specification. It seemed like a good idea at the time (mid 
1980s), when the notion that people would be translating whole composite 
documents automatically from braille to print, though understood, was 
thought to be a rarity not worth worrying very much about. Also, any 
computerese that would be "displayed" in set-off fashion would likely be a 
fair chunk of code -- at least a line or two from a computer program -- so 
that a human reader would be very unlikely to have any trouble 
distinguishing which mode, CBC or literary, made contextual sense. Even 
though those assumptions made sense then, and perhaps still do most of the 
time, the CBC committee as of several (8?) years ago actively considered 
recommending that the explicit indicators be used within the (otherwise) 
blank lines to clarify the situation. I recall that the CBC committee had 
heard that that practice had already taken hold in Australia -- though I 
have no idea whether or not that is true.

Coming back to UEB, the obvious first casualty of using a skipped line or 
similar format device as an implicit grade switch is reliable automatic 
braille-to-print translation. There are lots of reasons that one might skip 
a line, and of course it's not always clear whether a change of page also 
incorporates a "skipped line."  Adding indentation might at least partly 
solve that problem, though again there are other reasons that one might 
indent material. Moreover, indentation subtracts from available space, with 
all that implies. In any case, in order not to cause ambiguity problems for 
braille-to-print processes, any adopted format grade switch signal would 
have to be used consistently and exclusively for that purpose, which I'm 
skeptical would be practical for general literature. If we say that it is 
to be used only for a certain class of books, e.g. math textbooks, then 
we'd have to define yet another term and task transcribers with 
prescreening books to see whether they fall into the given category.

Human readers would no doubt experience ambiguity problems much less 
frequently, but I'm not sure it would be never -- especially for very short 
fragments of text, such as an isolated "s" that could mean "so".

I don't recall anything in UEB design that absolutely forbids putting a 
grade 1 passage indicator on a line by itself; just a general sense that in 
braille it's best not to have lower or right-hand configurations isolated 
where they could be read as upper-left -- e.g. in this case, perhaps 
"bbb"  or three semicolons, neither of which is obviously likely to have 
much contextual meaning. Of course braille-to-print translation is not 
affected by this kind of "positional" ambiguity -- but still, it somehow 
doesn't sit well to put those kinds of indicators out in space. Another 
possibility that might be explored would be having a "contracted" grade1 
passage indicator. The one that comes immediately to mind is dots 146, 
which we traditionally do not use as "ing" at the beginnings of words in 
grade 2, even in words such as "Ingersoll" where the sound is reasonably 
consistent. So, although that sign has two meanings in grade 1 (radical 
close and visible space), it has no defined meaning following a space in 
grade 2. Should be consider using it as a "contracted" grade 1 passage 
indicator? It would cut down on clutter even in in-line situations, and of 
course it wouldn't have any positional ambiguity problems.

That wouldn't solve the problem with the closing grade 1 indicator, dots 
56, 3. Maybe this would be a good use for our lone unassigned single-cell 
symbol, dots 16? Just musing aloud.

Turning to the question of having shortened signs for the math functions, 
the only mechanism that we have within present UEB design is that of actual 
contractions, and of course that implies grade 2, which we have been 
assuming will normally not be in effect for the very kinds of complex 
expressions where shorter signs would be most useful. I suppose we could 
invent some symbols that would have a defined meaning regardless of grade 
-- or another way of looking at it would be to have another intermediate 
grade. This is not unknown in other languages. German, for example, has a 
small set of elementary contractions that are used practically from the 
beginning, and then a more elaborate set that is added on top of those. We 
do have plenty of two-cell symbols that could be assigned in that way, 
still leaving adequate room for future needs and even expansion of the 
contraction system (if people eventually want to do that, which wouldn't 
surprise me at all). The only thing I would want to warn against would be 
to use these symbols just for math, because then we would be getting away 
from our roots and sliding towards a separate "math code." What the heck, a 
"tan" contraction would surely be very popular in sunny Australia, while a 
"log" contraction would serve us well here in New England where winter just 
won't let go and we're still huddled around the hearth after a week-long 
nor'easter. A "sin" contraction would probably be the most used everywhere, 
of course.

Cheers,
Joe

--=====================_8011870==.ALT
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<html>
<body>
<font size=3D3>Hello Janet and everyone,<br><br>
The idea that &quot;shortcuts&quot; can easily be taken in UEB has been
discussed from time to time, and might even be considered one of the
minor design principles. We have figured that, in writing for their own
use (as when taking notes in class), people will often drop the dot 5
from the common math operators and not bother with explicit grade
switches. All of that seems good and natural and consistent with what we
all do when writing informally, no matter what writing system we use.
However, I'm worried about trying to make such shortcuts part of the
formal system, lest in the process we abandon some of UEB's
strengths.<br><br>
Using a skipped line to signal a switch into or out of CBC was part of
the original CBC specification. It seemed like a good idea at the time
(mid 1980s), when the notion that people would be translating whole
composite documents automatically from braille to print, though
understood, was thought to be a rarity not worth worrying very much
about. Also, any computerese that would be &quot;displayed&quot; in
set-off fashion would likely be a fair chunk of code -- at least a line
or two from a computer program -- so that a human reader would be very
unlikely to have any trouble distinguishing which mode, CBC or literary,
made contextual sense. Even though those assumptions made sense then, and
perhaps still do most of the time, the CBC committee as of several (8?)
years ago actively considered recommending that the explicit indicators
be used within the (otherwise) blank lines to clarify the situation. I
recall that the CBC committee had heard that that practice had already
taken hold in Australia -- though I have no idea whether or not that is
true.<br><br>
Coming back to UEB, the obvious first casualty of using a skipped line or
similar format device as an implicit grade switch is reliable automatic
braille-to-print translation. There are lots of reasons that one might
skip a line, and of course it's not always clear whether a change of page
also incorporates a &quot;skipped line.&quot;&nbsp; Adding indentation
might at least partly solve that problem, though again there are other
reasons that one might indent material. Moreover, indentation subtracts
from available space, with all that implies. In any case, in order not to
cause ambiguity problems for braille-to-print processes, any adopted
format grade switch signal would have to be used consistently and
exclusively for that purpose, which I'm skeptical would be practical for
general literature. If we say that it is to be used only for a certain
class of books, e.g. math textbooks, then we'd have to define yet another
term and task transcribers with prescreening books to see whether they
fall into the given category.<br><br>
Human readers would no doubt experience ambiguity problems much less
frequently, but I'm not sure it would be never -- especially for very
short fragments of text, such as an isolated &quot;s&quot; that could
mean &quot;so&quot;.<br><br>
I don't recall anything in UEB design that absolutely forbids putting a
grade 1 passage indicator on a line by itself; just a general sense that
in braille it's best not to have lower or right-hand configurations
isolated where they could be read as upper-left -- e.g. in this case,
perhaps &quot;bbb&quot;&nbsp; or three semicolons, neither of which is
obviously likely to have much contextual meaning. Of course
braille-to-print translation is not affected by this kind of
&quot;positional&quot; ambiguity -- but still, it somehow doesn't sit
well to put those kinds of indicators out in space. Another possibility
that might be explored would be having a &quot;contracted&quot; grade1
passage indicator. The one that comes immediately to mind is dots 146,
which we traditionally do not use as &quot;ing&quot; at the beginnings of
words in grade 2, even in words such as &quot;Ingersoll&quot; where the
sound is reasonably consistent. So, although that sign has two meanings
in grade 1 (radical close and visible space), it has no defined meaning
following a space in grade 2. Should be consider using it as a
&quot;contracted&quot; grade 1 passage indicator? It would cut down on
clutter even in in-line situations, and of course it wouldn't have any
positional ambiguity problems.<br><br>
That wouldn't solve the problem with the closing grade 1 indicator, dots
56, 3. Maybe this would be a good use for our lone unassigned single-cell
symbol, dots 16? Just musing aloud.<br><br>
Turning to the question of having shortened signs for the math functions,
the only mechanism that we have within present UEB design is that of
actual contractions, and of course that implies grade 2, which we have
been assuming will normally not be in effect for the very kinds of
complex expressions where shorter signs would be most useful. I suppose
we could invent some symbols that would have a defined meaning regardless
of grade -- or another way of looking at it would be to have another
intermediate grade. This is not unknown in other languages. German, for
example, has a small set of elementary contractions that are used
practically from the beginning, and then a more elaborate set that is
added on top of those. We do have plenty of two-cell symbols that could
be assigned in that way, still leaving adequate room for future needs and
even expansion of the contraction system (if people eventually want to do
that, which wouldn't surprise me at all). The only thing I would want to
warn against would be to use these symbols just for math, because then we
would be getting away from our roots and sliding towards a separate
&quot;math code.&quot; What the heck, a &quot;tan&quot; contraction would
surely be very popular in sunny Australia, while a &quot;log&quot;
contraction would serve us well here in New England where winter just
won't let go and we're still huddled around the hearth after a week-long
nor'easter. A &quot;sin&quot; contraction would probably be the most used
everywhere, of course.<br><br>
Cheers,<br>
Joe<br>
</font></body>
</html>

--=====================_8011870==.ALT--