[DCRM-L] FW: [RDA-L] [UNICODE-M] Roman Numerals

[Deborah J. Leslie]

I am forwarding this message as it appeared on RDA-L, which itself was
forwarded MARC, and that forwarded apparently from UNICODE-M. I have no
idea what the ramifications for rare materials cataloging might be if
this is implemented. Perhaps John Attig or someone else better versed in
unicode can comment. --DJL

-----Original Message-----
From: Resource Description and Access / Resource Description and Access
[mailto:RDA-L at INFOSERV.NLC-BNC.CA] On Behalf Of Adam L. Schiff
Sent: Thursday, 14 December, 2006 21:53
To: RDA-L at INFOSERV.NLC-BNC.CA
Subject: Re: [RDA-L] [UNICODE-M] Roman Numerals (fwd)


This appeared on the MARC list, but since it included a recommendation
for
RDA, I thought I'd forward it here.   --Adam Schiff

---------- Forwarded message ----------
Date: Thu, 14 Dec 2006 16:42:24 -0500
From: "[ISO-8859-1] Joan Aliprand" <jaliprand at POBOX.COM>
Reply-To: UNICODE-MARC Discussion List <UNICODE-MARC at loc.gov>
To: UNICODE-MARC at LISTSERV.LOC.GOV
Subject: Re: [UNICODE-M] Roman Numerals

>From a practical point of view, it is easier to input these Roman
numbers
using letters from the regular keyboard than to invoke an alternate
input
method to obtain the Roman numerals in the Number Forms block.
(Furthermore,
most people will not know that the Roman numerals in the Number Forms
block
exist.)

>I see no advantage to having two ways to encode roman
>numerals though and would suggest that the character range cited above
be
>excluded from MARC records.

I think the issue of transcription of Roman numerals relates to
cataloging
rules rather than imposing limitations on the MARC 21 repertoire.
Vendors do
not always implement what MARC 21 specifies.

AACR2 includes rules on the transcription of Roman numerals, or
substitution
of Western-style Arabic numbers. Perhaps these rules could be expanded
in
RDA to specify (a) transcription of a Roman numeral as one or more
alphabetic letters rather than as a single combination form, and (b) to
provide explicit alphabetic letter substitutions for Roman numbers that
cannot be transcribed directly; for example: U+2180 ROMAN NUMERAL ONE
THOUSAND CD.

> Having two ways to encode some text strings will, I suspect,
>complicate authority control matching for names with diacritics but
display
>happens more often.

The issue of string comparison is a critical aspect of authority
control,
and so library processes in general. Those responsible need a thorough
understanding of Unicode, particularly canonical and compatibility
equivalences, normalization, casing, and character folding.

-- Joan Aliprand