Date:         Tue, 12 Sep 2006 19:37:02 -0400
Reply-To:     UNICODE-MARC Discussion List <[log in to unmask]>
Sender:       UNICODE-MARC Discussion List <[log in to unmask]>
From:         Joan Aliprand <[log in to unmask]>
Subject:      Re: Character Repertoire Expansion TIme? (2)
Content-Type: text/plain; charset="ISO-8859-1"

Karen Coyle already responded to Jim's initial question in the thread above.
(I reproduce Karen's posting here for everyone's convenience.) Scroll down
further for some comments from me replying to Jim's postings.

On Sat, 9 Sep 2006 11:35:51 -0700, Karen Coyle <[log in to unmask]> wrote:

>Jim,
>
>In the time that I've followed MARBI, changes have gone thusly: changes
>need to be coordinated between LC (who produces the majority of the
>original MARC records in the US) and the library "utilities" (was OCLC
>and RLG, now ?). Since Canada and the UK are now co-actors in the MARC21
>space, I assume that they will be included in any planning for changes
>taking place. To get all of these systems ready for the change, if it is
>a major one, takes a few years. At some point we'll see the announcement
>of a target date for implementation.
>
>Although there are already some systems that have expanded their
>character repertoires, I don't believe that any use a translation
>technique like the one included in the recently approved proposal, so
>some systems work will need to be done. I don't know if AVIAC (Autmation
>Vendors Information Advisory Committee, NISO) is the organization where
>vendor discussion will take place, but it seems logical.
>
>kc

I agree with Karen's view that the lead will come from LC and its peer
libraries (BL, LAC, NLM? others?) plus OCLC. Note that OCLC has already
added Thai and Tamil, and the British Library is said to be working on Armenian.

The restriction that is currently in the MARC 21 Specifications dates from
the time when the OCLC and RLIN databases were not ready to receive
characters outside those in the MARC-8 character sets in contributed records. 

NDMSO controls the specification, and will remove the restriction when the
time is right. This includes consulting with the parties listed above, with
vendors, and with MARBI if appropriate. No amount of tub thumping will make
it happen sooner.

A key issue of readiness has to do with NACO contributed authority records.
Is everything in place, including strict comparison of heading data that may
include Latin script letters, diacritical marks, and/or symbols beyond the
characters of MARC 21 Basic and Extended Latin plus Superscripts and
Subscripts? Can controlled headings established under AACR2 include any
characters from scripts other than Latin?

Since vendors sell their products internationally, they are unlikely to have
imposed any restrictions on what can be entered into their databases (other
than limitations imposed by the platform software they are using).
Restrictions on characters that may be exported in MARC 21 records may not
exist. Some years ago, I saw MARC 21 records from one vendor that included
characters outside the specified limitation. (The characters were
precomposed accented forms that were used internally, rather than decomposed
character sequences that would comply with the MARC 21 limitation.)

Rather than spending time agitating for removal of a rule, it would be far
better to try to anticipate what problems that could arise when our systems
receive characters that are not in the MARC-8 character sets.

-- Joan Aliprand