Re: [ANE-2] typing cuneiform
- Yes, that's the one, thanks, Dean. I haven't seen you in ages!
Maybe your argument will be valid years from now when all those front-end
gadgets are in place, but as a rationale for abandoning the century-plus-old
order of signs in the standard lists, it doesn't hold water.
I once came across a magnificent web page by Borger blasting the decision, and
now that I've had to type a couple of charts of signs, I see he's totally right.
I also remember huge fights on this question at an AOS or two that you were
involved in! But it seems that whenever there are varying views on how to encode
ancient scripts in Unicode, the least useful is the one that wins. Coptic is
Peter T. Daniels grammatim@...
>[Non-text portions of this message have been removed]
>From: Dean Snyder <dean.snyder@...>
>To: ANE-2 list <ANEfirstname.lastname@example.org>
>Sent: Tue, July 6, 2010 4:30:56 PM
>Subject: Re: [ANE-2] typing cuneiform
>Peter T. Daniels wrote at 11:20 AM on Friday, July 2, 2010:
>>Does anyone know where Unicode has hidden the sign AZ (Labat no. 131)?
>>has several signs called "NINDA2 TIMES " this and that, but it doesn't
>>be in that group.
>I'm not at home with my books but I believe this is the sign you are
>1228D CUNEIFORM SIGN PIRIG TIMES ZA
>>(Does anyone here wish to defend the Unicode listing of cuneiform signs
>>order of the transliteration of a Sumerian reading, instead of the
>>order by shape?)
>This is what the Unicode Standard says about the order of cuneiform signs:
>"Ordering. The characters are encoded in the Unicode Standard in Latin
>by primary sign name. Complex signs based on the primary sign are
>to graphic principles; in some cases, these correspond to the native
>Earlier drafts of the cuneiform proposal included this paragraph:
>Various alternatives have been suggested including using the traditional
>numbering, which itself
>derives from a formal ordering by first millennium character shapes;
>devising an ordering based on
>form which is better matched to the third millennium glyphs which appear
>in the character tables; or
>simply ordering the characters alphabetically according to their
>transliteration. A variant of the latter
>alternative has been chosen in this document: it is alphabetical by
>primary sign name with complex
>signs based on the primary sign organized according to graphic
>principles; in some cases, these
>correspond to the native analyses."
>My personal take on cuneiform sign ordering in Unicode follows.
>Because cuneiform signs changed shape over the four millennia of their
>use, all shape-based schemas for ordering cuneiform signs are
>necessarily of limited scope, bound to specific time frames, locales,
>and even individual hands. For a computer encoding of Sumero-Akkadian
>that must encompass all periods and locales of usage, ordering the signs
>by their transliteration values, which were more stable and more widely
>applied than the shapes were, seemed the right thing to do. This has the
>added advantage that all complex signs formed on a given base sign will
>be grouped together in the standard.
>But the naming and ordering of the signs, to which you refer, is really
>only a convention useful for Unicode-internal purposes; it is not
>intended for end user use. Entering cuneiform text on computers will
>only be useful for end users when we actually have cuneiform text input
>methods as part of our computer operating systems. And these input
>methods will provide different views into the master inventory of signs
>(the inventory encoded in Unicode). These views will present signs
>organized by period, locale, shape, phonemic values in different
>traditions, etc. For example, you will be able to request a view of :
>* only those signs that are attested in Neo-Babylonian documents
>* the evolving shape of the NA sign across all periods and locales of
>* only those signs with dentals as part of their phonemic sequence
>These views, these filters applied to the master inventory of signs,
>will be how we interact with cuneiform on a daily basis; we will not
>interact with the Unicode names.
>Dean A. Snyder
>Center for Inherited Disease Research (CIDR)
>Johns Hopkins School of Medicine
>Bayview Research Campus
>333 Cassell Dr, Triad Bldg, Suite 2000
>Baltimore, MD 21224
>cell:717 668-3048 office:410-550-4629
- Peter T. Daniels wrote at 10:35 PM on Tuesday, July 6, 2010:
>Maybe your argument will be valid years from now when all those front-endMy guess is it will be 3 to 4 years before we have cuneiform text input
>gadgets are in place, but as a rationale for abandoning the century-plus-old
>order of signs in the standard lists, it doesn't hold water.
methods as part of Mac and Windows operating systems, with Linuxes
>I once came across a magnificent web page by Borger blasting thedecision, and
>now that I've had to type a couple of charts of signs, I see he'sAre you referring to Borger's introduction to the "List of Neo-Assyrian
Cuneiform Signs: A practical and critical guide to the Unicode blocks" ,
linked to from here <http://www.sumerisches-glossar.de/downloads.htm>?
>I also remember huge fights on this question at an AOS or two that you wereCharacter encoding order is one of the least significant aspects of
>involved in! But it seems that whenever there are varying views on how
>to encode ancient scripts in Unicode, the least useful is the one that wins.
>Coptic is another example.
encoding a script, if for no other reason than the fact that oftentimes
collation sequences simply do not match encoding sequences. Witness,
e.g., the differing alphabetic ordering for the same umlauted vowels in
Swedish and German. A-umlaut, or "LATIN CAPITAL LETTER A WITH DIAERESIS"
as Unicode calls it, sorts alphabetically in German as "AE", i.e., near
the beginning of the alphabet; but the same character sorts near the end
of the alphabet in Swedish. Which alphabetic ordering should Unicode
adopt for its code point sequences, i.e., which tradition should be
anointed as canonical? What Unicode actually does (for various reasons,
and ignoring the decomposed forms) is to completely bypass the issue and
place this character in another block of characters separate from the
basic Latin block, in the "Latin-1 Supplement" block, at code point 196.
("A" is at code point 65.)
Conflicting and overlapping character ordering issues on computers are
dealt with by applying algorithms and/or tables to generate appropriate
collation sequences. Thus, if you set your computer operating system to
use Swedish localization methods it will do the right thing and you will
get lists of words sorted in Swedish order, not German order. But
underneath all text is still being processed and displayed using the
same, single underlying inventory of Unicode characters.
The Initiative for Cuneiform Encoding, the group responsible for the
Sumero-Akkadian cuneiform encoding in Unicode, chose to employ a
"neutral" character ordering based, within the confines of the Unicode
character naming restrictions, on the names of the primary signs,
thereby avoiding the discipline politics and technical limitations
associated with ordering the signs according to their shapes attested in
only one of several periods or locales.
We were fully aware that the names and ordering were not all that useful
for end users, and that transparent and intuitive software tools would
be needed to make productive use of the encoding.
In short, the encoding can be viewed as the foundational component of a
suite of components all of which must be in place before we can really
do any productive digital cuneiform text work.
Dean A. Snyder
Center for Inherited Disease Research (CIDR)
Johns Hopkins School of Medicine
Bayview Research Campus
333 Cassell Dr, Triad Bldg, Suite 2000
Baltimore, MD 21224
cell:717 668-3048 office:410-550-4629