Re: [RDA-L] Preferred access points for Expressions

[Riley, Jenn]

> Yes, and it is also one of the reasons why we remain so marginalized. There 
> are also other ideas, such as an
> "exchange format" that is used only for transferral. This is the fundamental 
> idea of OAI-PMH for example. In
> this scenario, each database does whatever it wants locally, but when 
> interoperating with other databases,
> they must follow OAI-PMH. I've done some work in this area. Google apparently 
> abandoned its work with OAI-PMH
> in favor of XML Topic Maps, which I do not know at all.

Google uses XML *Site* Maps , not XML *Topic* Maps 
. BIG difference.

Jenn


Jenn Riley
Metadata Librarian
Digital Library Program
Indiana University - Bloomington
Wells Library W501
(812) 856-5759
www.dlib.indiana.edu

Inquiring Librarian blog: www.inquiringlibrarian.blogspot.com

Re: [RDA-L] Preferred access points for Expressions

[Karen Coyle]

Bernhard Eversberg wrote:

The new worldview according to RDA is here:

http://www.rdaonline.org/ERDiagramRDA_24June2008.pdf

That's an entity-relationship diagram. (Can anyone sketch a relational
database design based on it? Would that be practicable? Would it scale?)


I don't think it's intended to support RDBMS. The ERD is realized in a
system based on XML. I also don't think that the ERD for RDA is any more
complex than a complete one for MARC would be -- with the possible
exception that the RDA diagram includes FRBR relationships. MARC defines
over 1700 textual data elements (if you count each subfield as an
element) and over 200 different fixed field  data elements.  If you
diagrammed it all out, you could paper the ceiling as well. :-)

kc

--
---
Karen Coyle / Digital Library Consultant
kco...@kcoyle.net http://www.kcoyle.net
ph.: 510-540-7596   skype: kcoylenet
fx.: 510-848-3913
mo.: 510-435-8234

Re: [RDA-L] Preferred access points for Expressions

[Karen Coyle]

Bernhard Eversberg wrote:



While that makes sense, the more important part of my question was "What
else have we got?" The necessary migration to something new can only
begin on a scale worth mentioning once there is a robust, extensible,
and  well-tested schema that can accomodate all the important elements
and support all the vital functions. Then, nothing convinces more than a
running model that easily demonstrates the advantages. Lacking that, I'm
afraid, MARC
remains the default.


We don't have to wait for a schema that supports all vital functions,
however. We have to learn to do iterative development. One of the
problems with our standards process is that we set out to define all
possible needs before we will consider making it public. Look at RDA
itself, years of discussion by a small group of people, massive texts
that don't give any insight into the reasoning behind the decisions.
It's almost a joke that they then ask for public comments.

We need something more like the open source process, where development
is shared, open, and incremental. You test as you go, so you find out
what does and doesn't work before you build on those mistakes.

This is one of the reasons why we went to the effort to put the RDA
elements and vocabularies in a public registry -- so some
experimentation could begin. It also looks like the RDA online product
(which, I realize, many will not have access to) will be able to output
a draft record in XML based on the registered data elements. Those
records can be made publicly available. This gives us an opportunity to
begin to build schemas and test applications based on that output. With
support for application profiles, as long as everyone works from the
same group of data elements (which are identified with URIs) we don't
all need to use exactly the same record format for our data to be
compatible.

What I'd love to be able to do (and with the agreement of Library of
Congress, which holds the control over the MARC format that we use) is
to add into the registry those MARC elements that aren't included in RDA
so that we can add in those elements as needed. Most of those elements
will be controlled lists, and some of those won't be necessary if we use
identifiers for values throughout the RDA record. But it would give us a
more complete picture and further the testing.

kc


--
---
Karen Coyle / Digital Library Consultant
kco...@kcoyle.net http://www.kcoyle.net
ph.: 510-540-7596   skype: kcoylenet
fx.: 510-848-3913
mo.: 510-435-8234

Re: [RDA-L] Preferred access points for Expressions

[Weinheimer Jim]

Bernhard Eversberg wrote:

> So, it's not even apples vs. oranges but much bigger nonsense. Of
> course, if we can't get that across, we may indeed have lost.

I agree 100%. Of course the problem is that people will always make comparisons 
no matter what we say or do. A genuine danger (or it could also be seen as a 
real opportunity) is with this Google-publishers agreement. If it is accepted, 
and some form of it will be eventually, there will be *tremendous* pressure on 
libraries to join in so that all the users can see all of the scanned books for 
free. Google will make piles of money on it. From an audacious beginning: 
simply scanning all the books they could, completely ignoring all the 
screaming, Google comes out very positively.

But once it's implemented, more and more people doing "research" will start in 
Google Books, as many are right now. People will become more and more used to 
digital books as e-readers become better. Why will people feel they need to pay 
librarians to create and maintain library catalogs?

One way is to show that we are indeed, capable of changing and moving ahead. To 
demonstrate hat there are genuine problems with Google results (this is a very 
difficult concept to get across to people). And that a higher level of 
organization is needed. This has been a problem for librarians and catalogers 
from the earliest days. Remember the debates of Panizzi to his opponents.

Librarians are not alone in wanting to fix the problems of information 
recall--lots of information specialists at least feel the problem, even if they 
don't understand it like we do. We'll see what happens. I just hope, as I am 
sure you do, that the powers-that-be realize that the future lies in genuine 
and deep cooperation.

Jim Weinheimer

Re: [RDA-L] Preferred access points for Expressions

[Bernhard Eversberg]

Weinheimer Jim wrote:

All people know is that they can find "stuff" more easily
in Google than in a library catalog. So, the moment we compare Google
products to the library catalog, we've already lost.


Sure, but we cannot agree to make such comparisons in the first place.
It is silly for more than one very big reason:
1. The stuff we are indexing is vastly different
   (lacking, for instance, the syndetic structure of the links, not
to speak of the full text of documents)
2. We cannot employ their methods of statistical evaluations of text,
   and not on that scale anyway, that they use to refine results
3. Neither do we have massive amounts of user input and klick
   evaluations to draw useful conclusions from (The German BibTip
   project does achieve some effect from that approach, but not on
   a comparable scale)
4. Our infrastructure is pathetically smaller
5. That includes our manpower (sorry, personpower. No, human resources?)

So, it's not even apples vs. oranges but much bigger nonsense. Of
course, if we can't get that across, we may indeed have lost.


... Yet the one area we could immediately
take the lead in is with providing URIs for concepts. Nobody else has
the forms and references like we have. All we need are to create the
URIs and make them available to everyone. ...


You mean, "to make them universally accessible and useful".
If only the powers in control of those resources understood that. I can
see no signs of any of this coming about. Rather the opposite: Not even
RDA will be made universally accessible, not to speak of useful.

B.Eversberg

Re: [RDA-L] Preferred access points for Expressions

[Weinheimer Jim]

Bernhard Eversberg wrote:

> Now the Google approach to making information findable is an _entirely_
> different one. For their general search engine, they rely not on
> metadata at all but on statistical and algorithmic evaluation of text as
> it is, and in huge quantities, setting huge arrays of text crunchers to
> the task and evaluating giant amounts of user input as well in
> innovative ways. None of these components is available to us, at least
> not in the quantities that it takes, and not to speak of the
> infrastructure. (They operate the largest server network on this
> planet.)
> OTOH, what works for HTML and PDF documents doesn't work so well for
> their scanned books. One reason is that the book scans lack the
> syndetic structure (the links) that are a "natural" component of
> the web files and that contributes a great deal to the success of
> the search engine. So, they've come around to use libary metadata.
> But they're far from making the most productive use of it.

While I agree with you that Google results don't work very well (for example, 
the Google Book search results are positively horrible since often, even if I 
have the exact title, author, etc. I still can't find the books that I know are 
in there), lots of people like it anyway since they don't understand a lot of 
the problems. (Just as logical to me is: if the Google Book Search is so bad 
because I know it is, how can I know that the Google web search is any better? 
Maybe I just don't know what I'm missing) All people know is that they can find 
"stuff" more easily in Google than in a library catalog. So, the moment we 
compare Google products to the library catalog, we've already lost.

But even Google appears to see some of the problems. See: 
http://googlewebmastercentral.blogspot.com/2008/04/retiring-support-for-oai-pmh-in.html
 for the announcement that Google was stopping support for OAI-PMH in favor of 
XML Sitemaps to get some structured metadata. (I remembered "Topicmaps" 
incorrectly) At least it seems you can create your own namespaces, but it's yet 
another new thing to learn.

I feel that these are some of the formats we will have no choice but to 
provide. It won't be our decision. Yet the one area we could immediately take 
the lead in is with providing URIs for concepts. Nobody else has the forms and 
references like we have. All we need are to create the URIs and make them 
available to everyone. Then let people play with them and see what happens.

Jim Weinheimer

Re: [RDA-L] Preferred access points for Expressions

[Bernhard Eversberg]

Weinheimer Jim wrote:

... If that, who's taking up the challenge?


Unfortunately, I believe other organizations are, such as Google.


Now the Google approach to making information findable is an _entirely_
different one. For their general search engine, they rely not on
metadata at all but on statistical and algorithmic evaluation of text as
it is, and in huge quantities, setting huge arrays of text crunchers to
the task and evaluating giant amounts of user input as well in
innovative ways. None of these components is available to us, at least
not in the quantities that it takes, and not to speak of the
infrastructure. (They operate the largest server network on this
planet.)
OTOH, what works for HTML and PDF documents doesn't work so well for
their scanned books. One reason is that the book scans lack the
syndetic structure (the links) that are a "natural" component of
the web files and that contributes a great deal to the success of
the search engine. So, they've come around to use libary metadata.
But they're far from making the most productive use of it.



I don't think the world will wait for us.


Of course not. Even LibaryThing, despite its name, invented everything
they have by themselves!

All of this doesn't bode well for RDA unless someone can come up with
a thrilling demonstration of its intrinsic powers. Something really
compelling, like Microsoft's "Photosynth" (which got quite some
publicity yesterday...)


B.Eversberg

Re: [RDA-L] Preferred access points for Expressions

[Weinheimer Jim]

Bernhard Eversberg wrote:

> The necessary migration to something new can only
> begin on a scale worth mentioning once there is a robust, extensible,
> and  well-tested schema that can accomodate all the important elements
> and support all the vital functions. Then, nothing convinces more than a
> running model that easily demonstrates the advantages. Lacking that, I'm
> afraid, MARC remains the default.

Yes, and it is also one of the reasons why we remain so marginalized. There are 
also other ideas, such as an "exchange format" that is used only for 
transferral This is the fundamental idea of OAI-PMH for example. In this 
scenario, each database does whatever it wants locally, but when interoperating 
with other databases, they must follow OAI-PMH. I've done some work in this 
area. Google apparently abandoned its work with OAI-PMH in favor of XML Topic 
Maps, which I do not know at all.

> IOW, "Yes, we can!", but the vehicle in which to sail down the newly
> conceived avenues still awaits its construction.
> Or is it rather "Don't ask what RDA can do for you, ask what you can do
> for better metadata!"? If that, who's taking up the challenge?

Unfortunately, I believe other organizations are, such as Google. Google 
practices are not so great, but perhaps our field needs that kind of a leader. 
The new information management ideas are quite different from librarian ideas. 
They would rather get something going now and fix it later, an idea that 
frightens someone like me to death. But otherwise everyone is faced with 
endless waits while we seek "perfection."

I don't think the world will wait for us.

Jim Weinheimer

Re: [RDA-L] Preferred access points for Expressions

[Bernhard Eversberg]

Weinheimer Jim schrieb:


I believe that XML formats of MARC are far more flexible than you appear
to believe--certainly far more flexible than any ISO2709 head-breaking
format. I wouldn't have opted in my article for MARCXML, probably a
variant MODS, Dublin Core, or even made up a unique XML coding simply
for purposes of examples. *Nobody* besides librarians understands
anything about MARC, certainly not in the ISO2709 or even in the XML
version.


While that makes sense, the more important part of my question was "What
else have we got?" The necessary migration to something new can only
begin on a scale worth mentioning once there is a robust, extensible,
and  well-tested schema that can accomodate all the important elements
and support all the vital functions. Then, nothing convinces more than a
running model that easily demonstrates the advantages. Lacking that, I'm
afraid, MARC
remains the default.
IOW, "Yes, we can!", but the vehicle in which to sail down the newly
conceived avenues still awaits its construction.
Or is it rather "Don't ask what RDA can do for you, ask what you can do
for better metadata!"? If that, who's taking up the challenge?

B.Eversberg

Re: [RDA-L] Preferred access points for Expressions

[Weinheimer Jim]

Bernhard Eversberg wrote:

> We have to keep in mind that XML as such is not on the same level as
> MARC. It is a punctuation standard and as such can only replace ISO2709,
> whereas MARC is a grammar and as such can be replaced, in the XML
> context, only by a Schema. So I suppose that's what you mean. Leaves
> us with the question "Which schema?" MARCXML? That's nothing but MARC
> in a much more unwieldy costume. What else have we got?

I believe that XML formats of MARC are far more flexible than you appear to 
believe--certainly far more flexible than any ISO2709 head-breaking format. I 
wouldn't have opted in my article for MARCXML, probably a variant MODS, Dublin 
Core, or even made up a unique XML coding simply for purposes of examples. 
*Nobody* besides librarians understands anything about MARC, certainly not in 
the ISO2709 or even in the XML version.

> One practical question: You suggest we get rid of the "preferred
> title" and have just titles, as many as needed?

In my suggestion, all titles would be treated equally. The function of the 
traditional string of "preferred title" of grouping would be handled by the URI 
which no one would see (probably). Each user could set their own "preferred 
title" so for example, if an Italian were in the US or Germany, he or she could 
set Italian forms of names. Again, this would necessitate changes in the 
current structures of our files, i.e. adding a language subfield for each form 
of name. Also, at least in AACR2/LCRI practice, if a reference conflicts with a 
heading, you are to break the conflict, but if a reference conflicts with a 
reference, you do not break the conflict. In a system such as I am proposing, 
you may have to break those conflicts as well. I am sure there would be other 
changes, too.

> > In such a system, the procedures would have to change significantly,
> > although not completely. I thought that the main change would be in the
> > "worldview" of the cataloger.
> >
> The new worldview according to RDA is here:
> 
> http://www.rdaonline.org/ERDiagramRDA_24June2008.pdf
> 
> That's an entity-relationship diagram. (Can anyone sketch a relational
> database design based on it? Would that be practicable? Would it scale?)
> Print it out on 3" by 4" stationery and wallpaper your room with the
> 15 sheets you get.

What a great diagram! That will send everybody running for sure!

But to be fair, what I meant by my statement that the "worldview' of the 
cataloger would change is the inevitable fact that records produced by German 
agencies, French agencies, Italian, Russian and others will all go into the 
same pot someday. Therefore, the cataloger's worldview should include all 
records in all rules. Not only AACR2/RDA, but in all rules. Why?

Because people will want to use--and will use such a tool. The informational 
universe of our users is changing to include much more than ever before and if 
we want to continue to be relevant, we must make tools that will serve their 
needs--not just ours. Expecting everybody to use a single form of name would be 
just as unrealistic as expecting everybody to learn Esperanto. It simply won't 
happen. The world will not change to suit our purposes--we are the ones who 
must change and use the full power today's technology affords.

This is why I think that "preferred title" is based on 19th century methods and 
ultimately will prove itself not to be sustainable.

Jim Weinheimer

Re: [RDA-L] Preferred access points for Expressions

[Bernhard Eversberg]

Weinheimer Jim wrote:


A colleague and I published an article on this but--I blame myself on
this--we gave the examples in MARC21 format when we should have provided
examples in XML. (We realized it could be done in MARC format, which was
amazing to me!) In any case, I think that if we had given everything in
an XML format, people would have understood it better.


We have to keep in mind that XML as such is not on the same level as
MARC. It is a punctuation standard and as such can only replace ISO2709,
whereas MARC is a grammar and as such can be replaced, in the XML
context, only by a Schema. So I suppose that's what you mean. Leaves
us with the question "Which schema?" MARCXML? That's nothing but MARC
in a much more unwieldy costume. What else have we got?

One practical question: You suggest we get rid of the "preferred title"
and have just titles, as many as needed?


In such a system, the procedures would have to change significantly,
although not completely. I thought that the main change would be in the
"worldview" of the cataloger.


The new worldview according to RDA is here:

http://www.rdaonline.org/ERDiagramRDA_24June2008.pdf

That's an entity-relationship diagram. (Can anyone sketch a relational
database design based on it? Would that be practicable? Would it scale?)
Print it out on 3" by 4" stationery and wallpaper your room with the 15
sheets you get.

B.

Re: [RDA-L] Preferred access points for Expressions

[Adam L. Schiff]

Hal,

I'm not sure that I understand what your concern is.  In terms of current
authority practice, we would first decide on the name for the work.  Then
if what we have is a particular expression of that work that needs to be
identified, we use what we've done for the work and add to it in some way
or another to identify the expression.  I don't see how this in any way
tries to anticipate all the possible elements that might be needed to
identify an expression at the time the work record is created.

What am I not understanding of your question?

Adam

^^
Adam L. Schiff
Principal Cataloger
University of Washington Libraries
Box 352900
Seattle, WA 98195-2900
(206) 543-8409
(206) 685-8782 fax
asch...@u.washington.edu
http://faculty.washington.edu/~aschiff
~~

On Wed, 21 Jan 2009, Hal Cain wrote:


Adam L. Schiff wrote:


A qualifier can be added to the name of the work to indicate a particular
expression of the work.  If desired, I think an access point for an
unnarrated expression of Peter and the wolf could be created following the
instructions in RDA 6.13.  RDA wouldn't prescribe what term to record for
the "Other distinguishing characteristic of the expression", but perhaps
"Unnarrated" would be the most appropriate term to use.


Um.  Are we to suppose, then, that the name of a work (however the basic
elements are constructed -- creator name + title, or title alone) should be
elaborated with all possible differentiating elements when first created?

While I'm on record as saying that I believe current AACR2/LCRI/NACO
practices discourage cataloguers from proceeding to meet in advance
reasonably forseeable conflicts, I'm aghast at the prospect of making
explicit, in advance, criteria of difference in terms of extended features
which are *not* included in the expression in hand/on screen/in hearing.  I
doubt that's what Adam means.  But I'm struck by the negative characteristic
"unnarrated" (just as much as, in another frame of reference, I'm struck by
name constructs such as "Pseudo-Augustine").

Negative characteristics may be implied: one supposes that an expression
record for J.R.R. Tolkien's own maps for The Lord of the Rings would be
without a distinguishing qualifier, while Pauline Baynes's reworked versions
of the maps would, if not attributed primarily to her, be named with some
form of distinction from Tolkien's originals.

It hardly seems reasonable to proceed as if nothing will be changed or
amended after it has been created.  Must everything be made explicit, or can
some things be added afterwards, or be left implied?

If the former (everything to be made explicit at the outset), then RDA must
fail because unworkable.

Hal Cain
Dalton McCaughey Library
Parkville, Victoria, Australia
h...@dml.vic.edu.au

Re: [RDA-L] Preferred access points for Expressions

[Hal Cain]

Adam L. Schiff wrote:


A qualifier can be added to the name of the work to indicate a particular
expression of the work.  If desired, I think an access point for an
unnarrated expression of Peter and the wolf could be created following the
instructions in RDA 6.13.  RDA wouldn't prescribe what term to record for
the "Other distinguishing characteristic of the expression", but perhaps
"Unnarrated" would be the most appropriate term to use.


Um.  Are we to suppose, then, that the name of a work (however the basic
elements are constructed -- creator name + title, or title alone) should
be elaborated with all possible differentiating elements when first created?

While I'm on record as saying that I believe current AACR2/LCRI/NACO
practices discourage cataloguers from proceeding to meet in advance
reasonably forseeable conflicts, I'm aghast at the prospect of making
explicit, in advance, criteria of difference in terms of extended
features which are *not* included in the expression in hand/on screen/in
hearing.  I doubt that's what Adam means.  But I'm struck by the
negative characteristic "unnarrated" (just as much as, in another frame
of reference, I'm struck by name constructs such as "Pseudo-Augustine").

Negative characteristics may be implied: one supposes that an expression
record for J.R.R. Tolkien's own maps for The Lord of the Rings would be
without a distinguishing qualifier, while Pauline Baynes's reworked
versions of the maps would, if not attributed primarily to her, be named
with some form of distinction from Tolkien's originals.

It hardly seems reasonable to proceed as if nothing will be changed or
amended after it has been created.  Must everything be made explicit, or
can some things be added afterwards, or be left implied?

If the former (everything to be made explicit at the outset), then RDA
must fail because unworkable.

Hal Cain
Dalton McCaughey Library
Parkville, Victoria, Australia
h...@dml.vic.edu.au

Re: [RDA-L] Preferred access points for Expressions

[Adam L. Schiff]

On Tue, 20 Jan 2009, Gene Fieg wrote:


While not reading all of the posts about expressions, it seems to me that at
least in the field of music, there could be some special circumstances.

For instance, The Nutracker is a ballet; it is intended to be a visual and
aural experience.  But in many cases, all we have on CDs is the music.  Do
we have a way of noting the expression, in a formal way (not just in a note)
that the manifestation/item is only of the music and not the dancing.


The music for the ballet is named by combining the preferred access point
for the composer with the preferred title of the musical work. (RDA
6.28.1.5).

RDA doesn't specifically address naming ballets themselves other than
perhaps through some examples.  Our current practice is to name the ballet
by title alone and the examples in RDA illustrate this.  A qualifier would
be added if necessary to break conflicts with other works with the same
title.  If one person created the music, choreography, costumes, etc. for
a ballet, then under RDA that person could be considered the creator and
the work could be named using the creator's name plus title, but I imagine
that this situation is fairly rare, as it also is for motion pictures.


How about Peter and the wolf?  I use to hear this piece narrated.  Now I am
hearing of lot performances on CD that are not narrated.  Again, is there a
formal way (not just a note) to indicate whether a particular
manifestation/item is narrated or not.


A qualifier can be added to the name of the work to indicate a particular
expression of the work.  If desired, I think an access point for an
unnarrated expression of Peter and the wolf could be created following the
instructions in RDA 6.13.  RDA wouldn't prescribe what term to record for
the "Other distinguishing characteristic of the expression", but perhaps
"Unnarrated" would be the most appropriate term to use.


Just pondering the ways of the universe of RDA (Roundabout Description
Access)

Gene Fieg
Cataloger
Claremont School of Theology
gf...@cst.edu



**
* Adam L. Schiff *
* Principal Cataloger*
* University of Washington Libraries *
* Box 352900 *
* Seattle, WA 98195-2900 *
* (206) 543-8409 *
* (206) 685-8782 fax *
* asch...@u.washington.edu   *
**

Re: [RDA-L] Preferred access points for Expressions

[Casey A Mullin]

Gene has touched on a salient issue here. Musical resources arguably
exhibit the most warrant for recording Expression-level attributes (or,
potentially, creating separate Expression records). While the examples
he cites do not currently correspond to codified components of access
points, there are cases where such components exist. For example, "Vocal
score" represents a printed Expression of a piece of music where the
accompaniment (orchestral or otherwise) is condensed into one system of
staves, meant for performance on a keyboard. This statement is housed in
its own subfield in MARC, which makes it fairly easy to identify as an
Expression-level attribute (as opposed to the rest of the string, which,
for the most part, represents Work attributes).

As for his Peter and the Wolf example, this is somewhat akin to the
"arr." statement in access points. An "arrangement" is defined as an
alteration of medium of performance, etc. Unfortunately, the boundaries
of this codified component do not cover /omissions/ of narrated parts,
per se. A case could be made that "arr." would be applicable here, but
it is a fuzzy one.

The Music Library Association Working Group on Work Records for Music
released a report last summer (see
http://www.musiclibraryassoc.org/BCC/BCC-Historical/BCC2008/BCC2008WGWRM1.pdf)
detailing what could and should appear in a /Work/ record. I think
similar work could and should be done regarding /Expression/ records,
both within and without the domain of music.

Whatever the route we take to get there, these are the kinds of
use-case-scenario discussions that must take place.

Casey M.

Gene Fieg wrote:

While not reading all of the posts about expressions, it seems to me that at
least in the field of music, there could be some special circumstances.

For instance, The Nutracker is a ballet; it is intended to be a visual and
aural experience.  But in many cases, all we have on CDs is the music.  Do
we have a way of noting the expression, in a formal way (not just in a note)
that the manifestation/item is only of the music and not the dancing.

How about Peter and the wolf?  I use to hear this piece narrated.  Now I am
hearing of lot performances on CD that are not narrated.  Again, is there a
formal way (not just a note) to indicate whether a particular
manifestation/item is narrated or not.

Just pondering the ways of the universe of RDA (Roundabout Description
Access)

Gene Fieg
Cataloger
Claremont School of Theology
gf...@cst.edu



--
Casey A. Mullin
Discovery Metadata Librarian
Metadata Development Unit
Stanford University Libraries
650-736-0849
cmul...@stanford.edu
http://www.caseymullin.com

Re: [RDA-L] Preferred access points for Expressions

[Gene Fieg]

While not reading all of the posts about expressions, it seems to me that at
least in the field of music, there could be some special circumstances.

For instance, The Nutracker is a ballet; it is intended to be a visual and
aural experience.  But in many cases, all we have on CDs is the music.  Do
we have a way of noting the expression, in a formal way (not just in a note)
that the manifestation/item is only of the music and not the dancing.

How about Peter and the wolf?  I use to hear this piece narrated.  Now I am
hearing of lot performances on CD that are not narrated.  Again, is there a
formal way (not just a note) to indicate whether a particular
manifestation/item is narrated or not.

Just pondering the ways of the universe of RDA (Roundabout Description
Access)

Gene Fieg
Cataloger
Claremont School of Theology
gf...@cst.edu

Re: [RDA-L] Preferred access points for Expressions

[Weinheimer Jim]

Jonathan Rochkind wrote:
> See, I don't think the techniques we're talking about here are really
> specific to rdbms or entity-relational databases.
..
> This is a pretty important fact of information systems that has direct
> impact on how we record metadata.  Those designing standards for
> recording of metadata need to understand it. If those doing that
> designing are catalogers are not programmers (as they probably should
> be), then those catalogers need to understand at least a bit about
> information systems. Because to live in information systems is the
> destiny of the metadata created.

Very good points. I think it is clear that in such a system, some practices and 
even information would have to change in some ways. What do you think they 
would be?

When I worked at FAO of the UN, where we used non-AACR2, non-MARC, and I was 
trying to imagine how we could fit into a VIAF/Onesac type of system. For only 
one example, FAO needs additional access for FAO corporate names, specifically, 
they need more specific access for FAO offices in each of the local offices: 
Bangkok, Santiago, Accra, and so on. In the NAF there is the single heading 
"Food and Agriculture Organization of the United Nations" but FAO needed to 
have separate headings for each sub-agency. Many times this was impossible for 
a non-FAO cataloger to know since there were little points to look for.

My idea was to include a "relations" field, where the relationship of the 
headings for bodies/entities would be delineated. These relationships could 
take all types of forms, e.g. 1:1, 1:3, 3:5, °probably the same,° or even 
unclear relationships. Naturally, there could be added language form (necessary 
for FAO but not for AACR2) and even time frame (for corporate bodies). Finally, 
cataloging/encoding rules and even specific "database where heading is used" 
could be added.

A colleague and I published an article on this but--I blame myself on this--we 
gave the examples in MARC21 format when we should have provided examples in 
XML. (We realized it could be done in MARC format, which was amazing to me!) In 
any case, I think that if we had given everything in an XML format, people 
would have understood it better.

In such a system, the procedures would have to change significantly, although 
not completely. I thought that the main change would be in the "worldview" of 
the cataloger.

I wonder what else would have to be done?

Jim Weinheimer

Re: [RDA-L] Preferred access points for Expressions

[Ed Jones]

My understanding is that VIAF is alive and well.  OCLC is using it as
one of the supports of their WorldCat Identities service, and plans to
extend it from personal names to geographic names.  I suspect that if
the VIAF gains wider acceptance there will be increased pressure from
other communities to reform the way we create authority records, which
remains anchored in the needs of card catalogs.

Thom Hickey did a presentation on VIAF at ELAG 2008, and the PowerPoint
can be downloaded at
http://library.wur.nl/WebQuery/file/formulier/profielelaglt_i00117278_00
1.ppt

Ed Jones
National University 

-Original Message-
From: Resource Description and Access / Resource Description and Access
[mailto:rd...@infoserv.nlc-bnc.ca] On Behalf Of Bernhard Eversberg
Sent: Tuesday, January 20, 2009 7:19 AM
To: RDA-L@INFOSERV.NLC-BNC.CA
Subject: Re: [RDA-L] Preferred access points for Expressions

Weinheimer Jim wrote:
>
> I have followed the VIAF for some time and applaud the general
> direction. This is the sort of project that should be given high
> priority since true exchange of this type of information can lead to
> genuine cooperation and a real savings in time and money for users as
> well as for libraries. It would also be one of the most important
> advances toward the Semantic Web, which could raise our profile
> significantly. There was another project called "Onesac" in Denmark
that
> I consulted with briefly. If was all in RDF(!!!), had authority
records
> from all over Europe and was extremely advanced. It seems to have
died,
> however. http://www.portia.dk/websites/onesac.htm
>
Interesting though this looks, one wonders what became of it. Project
ran until 2005 - but is there a final report? At least one would like
to read what can be learnt from it.

VIAF has an impressive web presence at OCLC, but even there, it is
difficult to find anything about the state of the project, and esp.,
about impending new stages or problems they might have run into.

>
> The example you point out should be eminently fixable although I don't
> know how it would work now: finding references of references. Using
URIs
> can be done in a whole variety of ways. Using URIs is not that much
> different from how relational databases work today ...

What exactly do you want to say here?
Do you really mean relational databases? I see this term frequently
used erroneously instead of "entity-relationship" databases. The word
"relational" in RDBS does precisely not say that the database cares
about relations between objects or entities. The term was created by
mathematicians who developed the first models. For them, a "relation"
was just a mathematical term taken from set theory and meaning a subset
of a table.


B.E.

Re: [RDA-L] Preferred access points for Expressions

[Jonathan Rochkind]

See, I don't think the techniques we're talking about here are really
specific to rdbms or entity-relational databases.

The idea of using a unique identifier to build a relationship between
two records is a pretty standard way to relate two records in _any_ kind
of information system.  It has one implementation in an
entity-relational rdbms, yes.  But it's not really unique to that.  In
the web world, using a URI of some kind as this unique identifier would
be the typical way to do it; a URI isn't the most natural way to do this
in a rdbms, but it's the same basic idea.

If you have a unique identifier for an entity (or for a record
describing an entity), you can use that unique identifier to record a
relationship between another entity/record and that one in an
information system. If you don't... then you can't record that
relationship in an efficient and effective way.  It's not unique to
rdbms technology.

This is a pretty important fact of information systems that has direct
impact on how we record metadata.  Those designing standards for
recording of metadata need to understand it. If those doing that
designing are catalogers are not programmers (as they probably should
be), then those catalogers need to understand at least a bit about
information systems. Because to live in information systems is the
destiny of the metadata created.

Jonathan

Weinheimer Jim wrote:

Bernhard Eversberg wrote:



 What exactly do you want to say here?
 Do you really mean relational databases? I see this term frequently
 used erroneously instead of "entity-relationship" databases. The word
 "relational" in RDBS does precisely not say that the database cares
 about relations between objects or entities. The term was created by
 mathematicians who developed the first models. For them, a "relation"
 was just a mathematical term taken from set theory and meaning a subset
 of a table.



Apologies for the shorthand. What I meant was the use of primary and foreign 
keys in databases. RDF is an extreme example of this way of operating, but in 
any case what I meant was the use of a primary/foreign key or a URI instead of 
text strings. It is my personal view that a lot of this is highly technical and 
should not be designed or decided upon by librarians or catalogers, although we 
should have a lot of input and be the primary testers. Our areas of expertise 
are different from those of an RDF or RDBS expert.

Will any of these projects happen or finish anytime soon? Of course not, but 
look how long it took to get ISBD. In many ways, the information community is 
screaming for a project that they can hook URIs into. I suspect that 
instituting such a system would take less time than ISBD so long as people 
remain flexible.

Jim Weinheimer







--
Jonathan Rochkind
Digital Services Software Engineer
The Sheridan Libraries
Johns Hopkins University
410.516.8886
rochkind (at) jhu.edu

Re: [RDA-L] Preferred access points for Expressions

[Weinheimer Jim]

Bernhard Eversberg wrote:

> What exactly do you want to say here?
> Do you really mean relational databases? I see this term frequently
> used erroneously instead of "entity-relationship" databases. The word
> "relational" in RDBS does precisely not say that the database cares
> about relations between objects or entities. The term was created by
> mathematicians who developed the first models. For them, a "relation"
> was just a mathematical term taken from set theory and meaning a subset
> of a table.

Apologies for the shorthand. What I meant was the use of primary and foreign 
keys in databases. RDF is an extreme example of this way of operating, but in 
any case what I meant was the use of a primary/foreign key or a URI instead of 
text strings. It is my personal view that a lot of this is highly technical and 
should not be designed or decided upon by librarians or catalogers, although we 
should have a lot of input and be the primary testers. Our areas of expertise 
are different from those of an RDF or RDBS expert.

Will any of these projects happen or finish anytime soon? Of course not, but 
look how long it took to get ISBD. In many ways, the information community is 
screaming for a project that they can hook URIs into. I suspect that 
instituting such a system would take less time than ISBD so long as people 
remain flexible.

Jim Weinheimer

Re: [RDA-L] Preferred access points for Expressions

[Bernhard Eversberg]

Weinheimer Jim wrote:


I have followed the VIAF for some time and applaud the general
direction. This is the sort of project that should be given high
priority since true exchange of this type of information can lead to
genuine cooperation and a real savings in time and money for users as
well as for libraries. It would also be one of the most important
advances toward the Semantic Web, which could raise our profile
significantly. There was another project called "Onesac" in Denmark that
I consulted with briefly. If was all in RDF(!!!), had authority records
from all over Europe and was extremely advanced. It seems to have died,
however. http://www.portia.dk/websites/onesac.htm


Interesting though this looks, one wonders what became of it. Project
ran until 2005 - but is there a final report? At least one would like
to read what can be learnt from it.

VIAF has an impressive web presence at OCLC, but even there, it is
difficult to find anything about the state of the project, and esp.,
about impending new stages or problems they might have run into.



The example you point out should be eminently fixable although I don't
know how it would work now: finding references of references. Using URIs
can be done in a whole variety of ways. Using URIs is not that much
different from how relational databases work today ...


What exactly do you want to say here?
Do you really mean relational databases? I see this term frequently
used erroneously instead of "entity-relationship" databases. The word
"relational" in RDBS does precisely not say that the database cares
about relations between objects or entities. The term was created by
mathematicians who developed the first models. For them, a "relation"
was just a mathematical term taken from set theory and meaning a subset
of a table.


B.E.

Re: [RDA-L] Preferred access points for Expressions

[Weinheimer Jim]

Bernhard Eversberg wrote:

> Here's where the VIAF idea comes in. It was conceived _because_ not
> everybody wanted to use English forms. And it may be the best
> starting point currently in existence to support your vision!
> With VIAF in place, a user may enter any form of name, and as long as
> VIAF knows that form, it will silently replace that name with its
> IdNumber or URI, whatever, and send that one instead of the name
> to the (VIAF-enabled!) catalog(s) in question, not bothering the
> user with this maneuver.
> BUT: This works as long as the name entered leads to one and only one
> authority record in VIAF. There are two other situations:
> 
> 1. There are several candidates for that name
> (When truncation is used, this will happen more often)
> Which of potentially very many name forms of the
> several records
> should be displayed? All of them, with no one
> emphasized?
> 
> 2. VIAF doesn't know the name
> In this case, the best thing to do would be to open an
> alphabetic
> index in the vicinity of the name in question and let
> the user browse
> and pick. This may lead on to a case of situation 1.

I have followed the VIAF for some time and applaud the general direction. This 
is the sort of project that should be given high priority since true exchange 
of this type of information can lead to genuine cooperation and a real savings 
in time and money for users as well as for libraries. It would also be one of 
the most important advances toward the Semantic Web, which could raise our 
profile significantly. There was another project called "Onesac" in Denmark 
that I consulted with briefly. If was all in RDF(!!!), had authority records 
from all over Europe and was extremely advanced. It seems to have died, 
however. http://www.portiadk/websites/onesac.htm

The cases you point out are probably just the tip of the iceberg when it comes 
to problems in implementing something like the VIAF.These are are a couple of 
possibilities, but other people will have all different kinds of ideas.

> I wonder, though, in what way VIAF may be of help in boolean searches
> where one of the terms is a name. For you will want to enable users
> to enter "sawyer clemens" and the system to find Twain's "Tom
> Sawyer".
> Me seems neither RDA nor VIAF address that kind of situation. Maybe we
> should have an interpolated search in some kind of works authority file,
> and from the result, use a (or several) work URIs to do the actual
> search. But as always, it's easy to build castles in the air, trouble
> starts only when you try to move in.
> 
> VIAF, as mentioned before, would have to be extended to include work
> authority records. Which do currently not exist, but should come into
> being with Scenario 1 of RDA.

The example you point out should be eminently fixable although I don't know how 
it would work now: finding references of references. Using URIs can be done in 
a whole variety of ways. Using URIs is not that much different from how 
relational databases work today; just on a grander scale. The records referred 
to in the URIs could reside on the web, be downloaded automatically to a local 
system, be updated automatically, or who knows what? The technology exists 
right now.

> And if I understand you right, you also advocate a general deregulation of 
>citation practices?

I think it's fairly deregulated now. My reference experience has shown to me 
that people almost always get citations wrong and that it is a huge waste of my 
time to assume that a specific citation is correct. Also, when I do an 
information literacy session, the no. 1 most popular and exciting thing I can 
show people is... automatic citations! I show them how to get them on WorldCat 
and in each electronic database. Faculty and students both love them!

This shows me that citation practices will probably change from the painful 
practice they are now to something much more exact and automatic. Our present 
citation practices are, as so often, based on 19th century technology and 
practices. So, I guess I am an advocate of much more regulation in citation 
practices! :-)

Jim Weinheimer

Re: [RDA-L] Preferred access points for Expressions

[Bernhard Eversberg]

Weinheimer Jim wrote:


But here is exactly where everything begins to disintegrate: which will
be the preferred form in the universe of the World Wide Web? Will
everyone be expected to use the English form? (I doubt that very
much)  The German? The Czech?

Here's where the VIAF idea comes in. It was conceived _because_ not
everybody wanted to use English forms. And it may be the best
starting point currently in existence to support your vision!
With VIAF in place, a user may enter any form of name, and as long as
VIAF knows that form, it will silently replace that name with its
IdNumber or URI, whatever, and send that one instead of the name
to the (VIAF-enabled!) catalog(s) in question, not bothering the
user with this maneuver.
BUT: This works as long as the name entered leads to one and only one
authority record in VIAF. There are two other situations:

1. There are several candidates for that name
   (When truncation is used, this will happen more often)
   Which of potentially very many name forms of the several records
   should be displayed? All of them, with no one emphasized?

2. VIAF doesn't know the name
   In this case, the best thing to do would be to open an alphabetic
   index in the vicinity of the name in question and let the user browse
   and pick. This may lead on to a case of situation 1.

I wonder, though, in what way VIAF may be of help in boolean searches
where one of the terms is a name. For you will want to enable users
to enter "sawyer clemens" and the system to find Twain's "Tom Sawyer".
Me seems neither RDA nor VIAF address that kind of situation. Maybe we
should have an interpolated search in some kind of works authority file,
and from the result, use a (or several) work URIs to do the actual
search. But as always, it's easy to build castles in the air, trouble
starts only when you try to move in.

VIAF, as mentioned before, would have to be extended to include work
authority records. Which do currently not exist, but should come into
being with Scenario 1 of RDA.

And if I understand you right, you also advocate a general deregulation
of citation practices?


B.Eversberg

Re: [RDA-L] Preferred access points for Expressions

[Kelleher, Martin]

Just to check it out, just tried Google book search, with a random name - John
Gardiner - and the results didn't seem organised any better than if I'd put them
in Google itself... So the answer is, Google Booksearch probably doesn't do any
job of bringing together what belongs together at all.

But I think most of us knew that was going to happen, didn't we?

Martin Kelleher
Electronic Resources/Bibliographic Services Librarian
University of Liverpool

-Original Message-
From: Resource Description and Access / Resource Description and Access
[mailto:rd...@infoserv.nlc-bnc.ca] On Behalf Of Bernhard Eversberg
Sent: 19 January 2009 14:01
To: RDA-L@INFOSERV.NLC-BNC.CA
Subject: Re: [RDA-L] Preferred access points for Expressions

Weinheimer Jim wrote:
>
> I think we're in agreement, but the main point I want to make is not
> to confuse "An entity needs a name!" (with which I agree) with "An
> entity needs a [single] name!' Today, this is no longer necessary and
> all of the variant names can be found, and displayed, in all kinds of ways.
This is also what the VIAF idea wants to achieve or support.

>
> It's also important to realize that this is nothing new. Thomas Hyde's
> catalog of the Bodleian library from the 1600s appeared to work in a
> similar manner. Although I can't find a copy of his catalog online,
> his headings were remarkable in that they included all of the variant forms.
> I remember the heading for Peter Abelard was something like:
>
> Abaelardus, Petrus, seu, Abelard, Peter,  Abeilard, Pierre, Abelardo,
> Pietro, [...].
>
Looking at the multitude of headings existing even for someone like Immanual
Kant, that seems plain impracticable to me.

>
> I can see something very similar with URIs. The "gathering point" will
> be the machine-readable URI, and the display of the heading[s] would
> be based on various factors. No. 1 would be based on the user's
> search, but the others could be based on IP address, user preferences,
> or who knows what else. Of course, the machine could be set to display
> only one or two lines and if this is not enough to display all the
> variants, then "[more...]" can be displayed.
>
Yes, but we cannot expect VIAF URIs to be available for everything anytime soon,
and not for works anyhow.

> This would demand some changes in our policies and procedures however.
> One example would be that each heading should have a language
> component, and there would be other changes as well. But it is
> important to realize that today, all forms can be "equal" and there is
> no need for "preferred form" anymore.
>
Although it can still be a big help. How does, for instance, Google Booksearch
do its job of bringing together what belongs together? It has got nothing but
textual strings to go by. Therefore, it will miss many references out there that
use idiosyncratic forms of names and titles. I think we need more tools for
interoperability than pie-in-the-sky URIs which are still very far from being
ubiquitous and not likely to be used much in citations and quotations at all.
I'm getting the impression, with all due respect, that you have yet to overcome
a certain main entry phobia that was rampant some years ago and that was eager
to throw the baby out with the bathwater. It is still, let me repeat, very
helpful and thus a Good Thing to have a clear and consistent name for as many
entities as possible. As for RDA as it stands now, it would otherwise have to be
rewritten in a major way.


B.Eversberg

Re: [RDA-L] Preferred access points for Expressions

[Weinheimer Jim]

Bernhard Eversberg wrote:

> Although it can still be a big help. How does, for instance, Google
> Booksearch do its job of bringing together what belongs together? It
> has got nothing but textual strings to go by. Therefore, it will miss
> many references out there that use idiosyncratic forms of names and
> titles. I think we need more tools for interoperability than
> pie-in-the-sky URIs which are still very far from being ubiquitous
> and not likely to be used much in citations and quotations at all.
> I'm getting the impression, with all due respect, that you have yet to
> overcome a certain main entry phobia that was rampant some years ago
> and that was eager to throw the baby out with the bathwater. It is
> still, let me repeat, very helpful and thus a Good Thing to have a clear
> and consistent name for as many entities as possible. As for RDA as it
> stands now, it would otherwise have to be rewritten in a major way.

But here is exactly where everything begins to disintegrate: which will be the 
preferred form in the universe of the World Wide Web? Will everyone be expected 
to use the English form? (I doubt that very much)  The German? The Czech?

If we rely on preferred forms, then interoperability will be limited only to 
those who share those forms. This is the situation as it is now. But I want my 
users to use your records for discovery; I want to be able to use your records 
for copy. I also don't think that people will ever search separate databases to 
discover the myriad versions of the "preferred forms" for Leo Tolstoy, not to 
mention subjects that are used in different databases. Why? Because they don't 
search separate authority databases today--why will they search something that 
is even more complex?

For me, I think it's great that we do have the "pie-in-the-sky" URIs available 
as a possible real solution. Can they be implemented tomorrow? Certainly not, 
it would take years of development if not longer, but some places are really 
trying with the Semantic Web, and at least it does allow for the promise of 
real interoperability. I haven't seen any other genuine proposals out there, 
although I may be missing something, but in this economic climate, we 
absolutely must work together.

How would citations work in such a system? I don't know, but from my experience 
with reference work, it certainly isn't done very well today. Figuring out 
citations often needs some pretty amazing acts of imagination!

One thing I am sure of: someday, perhaps sooner or later, all of the 
bibliographic records will be dumped together somewhere and there will use 
automated methods for finding duplicates and so on. We should all keep in mind 
that Google is working very hard to mine the "hidden web" which includes us, 
and I'm sure they will eventually succeed. Where will our preferred forms be 
then? What purpose will they serve?

It would be nice to have a clear and consistent name (although very few people 
use a clear and consistent name concept today). In the world of print, it was 
done with text strings and organization of records and cards, but today there 
are other options which may even be simpler. If all we are aiming for however, 
is to come up with unique text strings, AACR2 does that right now.

And by the way, Bernhard, you did catch me--I do have a "main entry phobia" ;-) 
but to be more precise, it's a "single main entry phobia," since the purpose of 
the single main entry died with the end of the card catalog. Some changes to 
the MARC format (preferably the XMLMARC version) would allow for the strange 
idea of "multiple main entries," or in DC terms: creators vs. contributors. 
(But that's another topic)

Jim Weinheimer

Re: [RDA-L] Preferred access points for Expressions

[Bernhard Eversberg]

Weinheimer Jim wrote:


I think we're in agreement, but the main point I want to make is not to
confuse "An entity needs a name!" (with which I agree) with "An entity
needs a [single] name!' Today, this is no longer necessary and all of
the variant names can be found, and displayed, in all kinds of ways.

This is also what the VIAF idea wants to achieve or support.



It's also important to realize that this is nothing new. Thomas Hyde's
catalog of the Bodleian library from the 1600s appeared to work in a
similar manner. Although I can't find a copy of his catalog online, his
headings were remarkable in that they included all of the variant forms.
I remember the heading for Peter Abelard was something like:

Abaelardus, Petrus, seu, Abelard, Peter,  Abeilard, Pierre, Abelardo,
Pietro, [...].


Looking at the multitude of headings existing even for someone like
Immanual Kant, that seems plain impracticable to me.



I can see something very similar with URIs. The "gathering point" will
be the machine-readable URI, and the display of the heading[s] would be
based on various factors. No. 1 would be based on the user's search, but
the others could be based on IP address, user preferences, or who knows
what else. Of course, the machine could be set to display only one or
two lines and if this is not enough to display all the variants, then
"[more...]" can be displayed.


Yes, but we cannot expect VIAF URIs to be available for everything
anytime soon, and not for works anyhow.


This would demand some changes in our policies and procedures however.
One example would be that each heading should have a language component,
and there would be other changes as well. But it is important to realize
that today, all forms can be "equal" and there is no need for "preferred
form" anymore.


Although it can still be a big help. How does, for instance, Google
Booksearch do its job of bringing together what belongs together? It
has got nothing but textual strings to go by. Therefore, it will miss
many references out there that use idiosyncratic forms of names and
titles. I think we need more tools for interoperability than
pie-in-the-sky URIs which are still very far from being ubiquitous
and not likely to be used much in citations and quotations at all.
I'm getting the impression, with all due respect, that you have yet to
overcome a certain main entry phobia that was rampant some years ago
and that was eager to throw the baby out with the bathwater. It is
still, let me repeat, very helpful and thus a Good Thing to have a clear
and consistent name for as many entities as possible. As for RDA as it
stands now, it would otherwise have to be rewritten in a major way.


B.Eversberg

Re: [RDA-L] Preferred access points for Expressions

[Weinheimer Jim]

Bernhard Eversberg wrote:

> In this regard, the "preferred access point" is of course a misnomer.
> The important function here is not the access aspect, but the naming
> aspect. An entity needs a name! For wherever an entity is mentioned,
> cited, listed, referred to or related to, the question is what should be
> the name (or name-title reference) displayed to the end-user so they can
> best make sense of it? Internally, URIs can serve to do all the linking,
> and other new tricks, but externally we need consistent naming of works
> and everything else.

I think we're in agreement, but the main point I want to make is not to confuse 
"An entity needs a name!" (with which I agree) with "An entity needs a [single] 
name!' Today, this is no longer necessary and all of the variant names can be 
found, and displayed, in all kinds of ways.

It's also important to realize that this is nothing new. Thomas Hyde's catalog 
of the Bodleian library from the 1600s appeared to work in a similar manner. 
Although I can't find a copy of his catalog online, his headings were 
remarkable in that they included all of the variant forms. I remember the 
heading for Peter Abelard was something like:

Abaelardus, Petrus, seu, Abelard, Peter,  Abeilard, Pierre, Abelardo, Pietro, 
[..].

and there were references from each form. I personally found this method of 
presenting all the variant forms along with the heading to be excellent, and it 
was much clearer for the user than the modern methods (except that everything 
in his catalog was in Latin!)

I can see something very similar with URIs. The "gathering point" will be the 
machine-readable URI, and the display of the heading[s] would be based on 
various factors. No. 1 would be based on the user's search, but the others 
could be based on IP address, user preferences, or who knows what else. Of 
course, the machine could be set to display only one or two lines and if this 
is not enough to display all the variants, then "[more...]" can be displayed.

This would demand some changes in our policies and procedures however. One 
example would be that each heading should have a language component, and there 
would be other changes as well. But it is important to realize that today, all 
forms can be "equal" and there is no need for "preferred form" anymore.

This would be looking toward building something for the future, which is what 
we need to do.

Jim Weinheimer

Re: [RDA-L] Preferred access points for Expressions

[Bernhard Eversberg]

Weinheimer Jim wrote:

Bernhard Eversberg wrote:

 > 5.1.3
 > The term preferred title refers to the title or form of title chosen as
 > the basis for the preferred access point representing a work. ...
 >
 > 5.1.4
 > The term preferred access point refers to the standardized access point
 > representing an entity. The preferred access point representing a work
 > or expression is constructed using the preferred title for the work. ...
..

While this is interesting, it has all been dealt with before, and I wish
that new thinking were taking hold. This again, is transferring the
practices of the card catalog ("preferred access point") when this is no
longer holds.


In this regard, the "preferred access point" is of course a misnomer.
The important function here is not the access aspect, but the naming
aspect. An entity needs a name! For wherever an entity is mentioned,
cited, listed, referred to or related to, the question is what should be
the name (or name-title reference) displayed to the end-user so they can
best make sense of it? Internally, URIs can serve to do all the linking,
and other new tricks, but externally we need consistent naming of works
and everything else.

B.Eversberg

Re: [RDA-L] Preferred access points for Expressions

[Gene Fieg]

What did he say???

Gene Fieg
Cataloger
Claremont School of Theology
gf...@cst.edu


-Original Message-
From: Resource Description and Access / Resource Description and Access
[mailto:rd...@infoserv.nlc-bnc.ca] On Behalf Of Stephens, Owen
Sent: Friday, January 16, 2009 8:43 AM
To: RDA-L@INFOSERV.NLC-BNC.CA
Subject: Re: [RDA-L] Preferred access points for Expressions

I'm glad you said that because I'm really struggling with the purpose
of access points. What is the reason for them in the context of RDA

Owen

On 16 Jan 2009, at 15:54, "Weinheimer Jim"  wrote:

Re: [RDA-L] Preferred access points for Expressions

[Stephens, Owen]

I'm glad you said that because I'm really struggling with the purpose
of access points. What is the reason for them in the context of RDA

Owen

On 16 Jan 2009, at 15:54, "Weinheimer Jim"  wrote:

Re: [RDA-L] Preferred access points for Expressions

[Weinheimer Jim]

Bernhard Eversberg wrote:

> 5.1.3
> The term preferred title refers to the title or form of title chosen as
> the basis for the preferred access point representing a work. ...
> 
> 5.1.4
> The term preferred access point refers to the standardized access point
> representing an entity. The preferred access point representing a work
> or expression is constructed using the preferred title for the work. ...
..

While this is interesting, it has all been dealt with before, and I wish that 
new thinking were taking hold. This again, is transferring the practices of the 
card catalog ("preferred access point") when this is no longer holds. Since any 
part of any record is now an "access point" and we no longer have to search for 
Goethe under "G" but in all kinds of ways, the purpose of the traditional 
access point is now to serve as a grouping for items that are similar in 
various ways as determined by the cataloger. The label of that grouping (the 
old preferred form) can--and should--vary based on the needs of catalogs and/or 
individual researchers, and this points to some interesting new directions.

It seems that RDA is simply rehashing the old card cataloging rules. If this is 
what it's all about, I see little reason to change from AACR2. With the 
introduction of URIs for these labels however, new and interesting 
possibilities for search and display of these "preferred forms" can arise. All 
of the forms of titles could display in a format similar to the former 
lcsh.info and people could choose the one that is familiar to them. I have in 
mind something similar to:
http://www.aimatshape.net/resources/aas-ontologies/plonejavaapplet.2006-11-14.8975779371.

I think these are the directions we need to go, and to get away from the card 
conventions. We would need some new rules and practices, but what would they 
need to be to help make the sort of tool above the most useful for researchers?

Jim Weinheimer

Re: [RDA-L] Preferred access points for Expressions

[Bernhard Eversberg]

Karen Coyle wrote:

Did you actually find references to "first expression" in RDA? If so,
could you point that out. I looked for such a statement but didn't find
it.

The database indeed reveals easily that there is no occurrence of the
phrase "first expression".


It always appeared to me that all expressions are treated equally.


This is in accordane with the view that the "work" is an abstract
entity. As such, it is not something you can look at to determine its
"real" title (or any other detail). What, then, should be regarded as
the "preferred title" for the work?

Let's look at what the rules say: they do not refer to a "first
expression" but only to the "title in the original language" by which
the work has become known:

5.1.3
The term preferred title refers to the title or form of title chosen as
the basis for the preferred access point representing a work. ...

5.1.4
The term preferred access point refers to the standardized access point
representing an entity. The preferred access point representing a work
or expression is constructed using the preferred title for the work. ...

6.2.2.4
For works created after 1500, choose as the preferred title the title in
the original language by which the work has become known through use in
resources embodying the work or in reference sources.

Perhaps the rules should somehow indicate what it all boils down to when
there is just one manifestation of a single expression, i.e., all items
are exactly like the one in hand. (The burden of proof for this is not
always light, but in the majority of cases there is not much in the way
of reasonable doubt.)
Would "Scenario 1" make a "work record" mandatory in all cases?


B.Eversberg

Re: [RDA-L] Preferred access points for Expressions

[Karen Coyle]

Did you actually find references to "first expression" in RDA? If so,
could you point that out. I looked for such a statement but didn't find
it.It always appeared to me that all expressions are treated equally.

kc

Chris Todd wrote:

It's Friday afternoon here and we're grappling with the RDA draft and have come 
very unstuck over preferred access points for expressions.

We are trying to establish the form of the preferred access point for the first 
(& only) expression of a work.

We realize that the use of identifiers rather than preferred access points to 
link manifestations and expressions would avoid this issue, but we would still 
like to see if we can understand the requirements and would appreciate any 
pointers anyone can give.

For a textual work with a single author the preferred access point for the work 
following 6.27.1.2 would be the author + title  (preferred forms of each)  eg 
Johnston, Alexa. Ladies a plate.

Looking at rule 6.27.3 it seems that the preferred access point for the 
expression  would be Johnston, Alexa. Ladies a plate. Text.  Note that our 
resource is the first manifestation of the only expression of this work.

There are no examples in the draft that show this kind of construction and all 
the examples cover the situation where more than one expression already exists.
As around 70% of the titles we encounter are likely to have only 1 expression 
we are keen to know whether we have actually reached the right conclusion about 
expression access points.

We looked briefly at dvds and it appeared that in order to create a preferred access 
point for a first expression we would be adding "two-dimensional moving image" 
to the title. A bit clunky!

Then (we didn't quit while we were ahead!) we went back to rule 6.13 to look for distinguishing 
characteristics of expression and found examples using the film Blade Runner which showed the use 
of the terms "Final cut" and "Directors cut" for revised versions of the film, 
but again nothing for the original version.

 Is there something we're missing?

Chris Todd & Charlotte Stretton

National Library of New Zealand






--
---
Karen Coyle / Digital Library Consultant
kco...@kcoyle.net http://www.kcoyle.net
ph.: 510-540-7596   skype: kcoylenet
fx.: 510-848-3913
mo.: 510-435-8234

Re: [RDA-L] Preferred access points for Expressions

[Ed Jones]

My understanding of the last sentence of 6.27.1.1 is that the instructions 
given under 6.27.3 are applied only "for new expressions of an existing work" 
(not for the first expression), i.e. the instructions relate to distinguishing 
these new expressions from the first one.

Ed Jones
National University 


-Original Message-
From: Resource Description and Access / Resource Description and Access on 
behalf of Chris Todd
Sent: Thu 1/15/2009 7:14 PM
To: RDA-L@INFOSERV.NLC-BNC.CA
Subject: [RDA-L] Preferred access points for Expressions
 
It's Friday afternoon here and we're grappling with the RDA draft and have come 
very unstuck over preferred access points for expressions.

We are trying to establish the form of the preferred access point for the first 
(& only) expression of a work.

We realize that the use of identifiers rather than preferred access points to 
link manifestations and expressions would avoid this issue, but we would still 
like to see if we can understand the requirements and would appreciate any 
pointers anyone can give.

For a textual work with a single author the preferred access point for the work 
following 6.27.1.2 would be the author + title  (preferred forms of each)  eg 
Johnston, Alexa. Ladies a plate.

Looking at rule 6.27.3 it seems that the preferred access point for the 
expression  would be Johnston, Alexa. Ladies a plate. Text.  Note that our 
resource is the first manifestation of the only expression of this work.

There are no examples in the draft that show this kind of construction and all 
the examples cover the situation where more than one expression already exists.
As around 70% of the titles we encounter are likely to have only 1 expression 
we are keen to know whether we have actually reached the right conclusion about 
expression access points.

We looked briefly at dvds and it appeared that in order to create a preferred 
access point for a first expression we would be adding "two-dimensional moving 
image" to the title. A bit clunky!

Then (we didn't quit while we were ahead!) we went back to rule 6.13 to look 
for distinguishing characteristics of expression and found examples using the 
film Blade Runner which showed the use of the terms "Final cut" and "Directors 
cut" for revised versions of the film, but again nothing for the original 
version.

 Is there something we're missing?

Chris Todd & Charlotte Stretton

National Library of New Zealand