Re: Acceptable alembic glyph variants

From: Asmus Freytag (asmusf@ix.netcom.com)
Date: Mon Dec 31 2007 - 11:09:35 CST


On 12/30/2007 6:28 PM, James Kass wrote:
> ... For now I ended up just drawing a glyph which
> resembles a piece of laboratory glassware. (Or, maybe my glyph
> looks more like a kiwi bird.)
The laboratory glassware that resembles a kiwi bird would define the
'canonical' glyph for this character, in my view. With this kind of symbol
there's a fine line between treating them as symbolic representation of
an abstract concept and a mere fixed picture.

An example might make this clear. (Fair warning, this example got
rather lengthy ;-) Here goes: You can use a (stylized) image of
time piece to stand as a symbol for the concept of "time". In principle
it does not matter what sort of time piece you decide to use.

However, while a cuckoo clock, a wristwatch, an hour glass, the
face on a clock tower or a grandfather clock are all time pieces,
most authors would probably not agree that all of these are possible
free _glyph_ variations of each other. Once an author has picked
one of these representations, the substitution of another one merely
based on the underlying nature of what is to represented might be
far from acceptable.

I can see two reasons for this. One is that any concrete representation
of an abstract concept carries certain connotations that may not
be fitting in a given context. Just consider that a cuckoo clock
while undeniably being a time piece is also quite ridiculous...

The other reason is that the same concrete symbol is often used in
a number of related meanings. For example "time", "delay" (i.e.
time is passing), etc. While there is considerable overlap, there
are differences. An hourglass is more often used to indicate a
wait, than a wall clock, while the latter is often used to indicate
time values (e.g. departure time, UI widget to set time, etc).

As a result, what a character encoding encodes cannot be the
underlying semantic, but must be the iconic representation itself.
It is then left to the user to decide the range of meaning(s) in which
this character can be employed.

For the font designer, the task is then to intuit the permissible
variation of the iconic representation. A first cut, would be that
the variation remain an example of the same concrete
representation. In other words, if an hourglass is encoded,
the permissible range of glyph variation is not that of 'time
piece' but instead, the new variants must still look
like hourglasses.

Further, if the encoded character had the representation of a
somewhat realistic hourglass, it would not do to substitute a
highly stylized abstract, or geometrical shape, even if there
is an example for which the common, descriptive name happens
to be "hourglass".

These particular examples are not hypothetical, but are examples
of actual issues encountered during the development of Unicode.

U+231B HOURGLASS

is encoded in distinction to

U+231A WATCH

which encodes a different iconic representation. But it's also
encoded in distinction to the two abstract symbols

U+29D6 WHITE HOURGLASS
U+29D7 BLACK HOURGLASS

where the correlation to hourglass is merely one of convention
(vertical bowtie being the alternative) and the semantics may
be unrelated to the concept of "time" or "time period" altogether.
Instead, as is common enough for abstract symbols, this one
has acquired a semantic distinction between the black and the
white form.

On the contrary, U+231B could arguably have been realized
as a (detailed) shadow image, with or without an indication of a
frame, or even in a kind of dithered grayscale or pseudo 3-D.
Any of those (concrete) realizations would likely have fit
the glyphic variation of 231B, that is to say, that of an iconic
hourglass.

(An aside before we get back to the alembic: these underlying
principles for the encoding of iconic representations were far
from clear at the outset, and there was considerable debate
at the time what 231B had been intended for. Over time,
the need to encode both an icon and a set of abstract shapes
became established and helped to clarify the necessary
distinctions as I have outlined them here.. Btw, a wall-clock
icon, is still missing from the standard, because it is undeniably
in use, and by these principles can't be unified with either
231A or 231B).

Now to the alembic: while it is intended to be used as
an iconic representation of (modern) chemistry, I would
argue from the above, that, for example substituting a test
tube (or other types of distinctive glassware) would be
a mistake. I've only ever seen the version with the black
outline, but for a very decorative font, I could be persuaded
that an outline form (hinting at transparency) could be
a permissible glyphic variation.

The same goes for the question whether it needs a little
tripod underneath or not. To me, that would be comparable
to whether one shows a framed or unframed hourglass.
In this case, it appears a matter of style, rather than a
matter of having substituted another, different, icon.

In the end, I would choose whatever is most recognizable
if printed in small point size in running text - which is
where this symbol has been used.

A./



This archive was generated by hypermail 2.1.5 : Mon Dec 31 2007 - 11:12:07 CST