Blooming Together_ Growing a Community Garden Worksheet.docx
Special Topics: Aggregates and Diachronic Works (Transcript)
1. ROUGH EDITED COPY
ALA-SPECIAL TOPICS
AGGREGATES AND DIACHRONIC WORKS
JULY 30, 2019
CART CAPTIONING PROVIDED BY:
ALTERNATIVE COMMUNICATION SERVICES, LLC
www.CaptionFamily.com
* * * * *
This is being provided in a rough-draft
format. Communication Access Realtime
Translation (CART) is provided in order to
facilitate communication accessibility and
may not be a totally verbatim record of the
proceedings
* * * * *
>> We're just under 10 minutes away from
today's workshop. Please feel free to introduce
yourself in the chat space on the right side of
the screen. We would love to hear who you
are -- we'll do more sound checks before we get
started. In between sound checks, there will be
only silence, thank you.
>> We’d like to encourage you to use the
chat space to say hello, introduce yourself or
reintroduce yourself as the case may be, and we
will be starting in 7 minutes, thanks.
>> Hi, everyone, this is Dan Freeman from
ALA Publishing. This is a sound check. We are
five minutes from our start time. In the
meantime please feel free to do.
>> Hi, everyone, this is Dan Freeman from
ALA Publishing. This is our final audio check.
2. We will be beginning in a little less than two
minutes. Feel free to introduce yourself in the
chat space. Hi, everyone. I'm Dan Freeman from
ALA Publishing. Welcome to the third event in
the RDA special topic series, aggregate and
diachronic works. I'm aggregate make a very
quick run through the technical info. I know
most of you are returning so you've heard this
already. I'll keep it short. The chat space is
on the lower right hand corner on the screen.
These events are best when they're interactive.
If you need help, webinar support is our host is
here. Private chat him by clicking the pull down
window underneath the chat space that says "to" .
You do not need to hold your questions until the
end. Type them in the chat space and we'll make
a note and make sure they get relayed to Ed. If
you're having audio problems, go to communicate
at the top of the screen and that will give you
the opportunity to -- usually if you're streaming
and you have a problem, dissecting and
reconnecting is -- two broadcasts open
simultaneously. Closing one should resolve that
right away. We are recording today, if you need
to hear this event again. We encourage you to
check out the ALA store for additional e-learning
opportunities. We are very glad to have Ed Jones
with us today. Ed is the associate director for
assessment and technical services at national
University in San Diego. He's been involved in
RDA pretty much since its inception so with that,
I will turn things over to Ed and he will get us
started. Welcome.
>> Thank you, Dan. Hello, everyone. Today
I'll be talking about aggregates and diachronic
works. I'll talk about aggregates first followed
by questions and answers and follow by diachronic
works followed by questions and answers. I hope
this will work better than putting them together.
A few slides will include bracketed citation
numbers in red. If you copy these down, you can
-- toolkit for more information.
While I'll be basing the descriptions in this
3. webinar on linked data concepts, the examples I
give will be using MARC11 encoding format. And
the examples will be in MARC21 as it currently
exists -- will be getting together to fine tune
-- so the first part of the webinar will be about
aggregates. I'll present the learning
objectives, background oh where the model -- and
MARC21. Bear in mind we're squeezing a lot in
here so this will be a superficial -- I hope you
won't be disappointed. By the end of this
portion of the webinar, my hope is that you will
be able to define an aggregate -- understanding
-- Library Reference Model, describe an aggregate
-- explain the cardality of -- pretty impressive
learning objectives. So what is an aggregate?
This is the definition from the RDA glossary.
I'll take you through it step by step.
First, in terms of the FRBRWEMI stack, the
aggregate itself is a manifestation. And this
manifestation embodies an aggregating work and it
also embodies multiple expressions or as RDA
definition has it, one or more expressions of one
or more works.
And these expressions realize the plan for
aggregation. If you're still wondering what is
an aggregate, don't worry. My hope is at the end
of the session you'll have a thorough
understanding of an aggregate. Let's get
started. First, a little history.
What I'm aggregate tell you is the story of
what led to the modeling of aggregates.
Aggregates were recognized as a problem, or
problematic, soon after FRBR was published in
1998. In 2005, IFLA -- OCLC to address this
problem. And six years later they issued the
final report. Unfortunately by this time, RDA
had already been published so the fruits of their
labor had to work. WGA defined an aggregate as a
manifestation embodying multiple distinct
expressions. It doesn't distinguish the
aggregating expression from the others. And the
WGA did make this distinction. The WGA further
identified three main kinds of aggregate, an
4. aggregate collection of expressions, such as a
collection of poetry or essays, an aggregate
resulting from augmentation such as a classic
novel with a scholarly introduction, and an
aggregate of parallel expression. Note that
while these are useful categories for looking at
aggregates, they are not the only ones, nor are
they mutually exclusive. They can occur in many
combinations as we'll see later.
And this brings us to the abstract model that
WGA came up with. All aggregates can be reduced
to this model. As I've said the WGA defines an
aggregate. We see the aggregate on the bottom of
the diagram. On the left we see multiple
distinct expressions embot in the manifestation.
Those stories, short poems, whatever. What's the
stuff on the right?
This is the real innovation of the model.
The WGA proposed in addition to these familiar
entities, a manifestation and expressions
embodied in it, something called an aggregating
expression, realizing an aggregating work. This
aggregating work is the work of the agent
responsible for selecting and arranging those
others expressions. It involves a plan. A plan
to select and arrange and maybe do other stuff
with those expressions on the left. Such a plan
may be relatively simple like arranging for two
novels to be published together in one volume.
It may be complex like selecting, arranging,
annotating the poems to be published in an
anthology. It may seem a bit confusing to call
it a work since it represents an activity and
framework rather than actual content but that's
how it's modeled as special kind of work with
special constraints. You might not think that
one or more of the expressions embodied in the
aggregating expression is important enough to be
described in your catalog or your discovery
service or whatever.
Whether an expression is significant or not,
that is whether you choose to describe that
particular expression is up to you. So this
5. diagram represents all possible expressions and
works, not the subset that you may actually
decide to describe.
So what does all this mean in practice?
I could use a simple aggregate as an example
but we know there's no such thing so instead I'll
use a typical aggregate. This is a volume of the
library of America. The library of America is a
diachronic work but that's another story. We'll
save that for the second half of the webinar.
For now you need to know this volume is an
aggregate. What kind of aggregate?
It contains -- 1936 and 1941. This makes it
an aggregate collection of expressions. It also
contains notes so it's also an aggregate
resulting from augmentation. And finally, Stein
Beck's widow serve as a special consultant.
How do we catalog this thing?
Well, it's an aggregate, which is a
manifestation by definition, so we start off
describing the manifestation. If you go to the
RDA guidance chapter on aggregates and look under
describing an aggregate -- clicking on this takes
us to a condition and four options. As you
probably know by now, conditions and options is
the standard structure for RDA instructions in
the new toolkit.
The condition in this case is a manifestation
is an aggregate. Bear in mind that the options
that follow are not mutually exclusive. That is
you can select more than one option if you want.
We'll go through the options one by one. You'll
be surprised how familiar they are. The first
option is to record information about the
embodied expressions in a note. You probably
don't want to do this with a collection of 300
poems but you might with say a set of three or
four novels. In this case, I looked at the table
of contents and a note of this sort seemed
doable. In a linked data environment this would
be recorded using the RDA element note on
manifestation.
And use an unstructured description because
6. that is what you do on a note with manifestation.
In a MARC21 environment, we use the 505 content
note. Mission accomplished.
In the second option, we relate the
manifestation to the aggregating expression, that
is to the plan of selecting and arranging and so
on. And we do this using the expression
manifested element. In our example, part of the
aggregate is work involved coming up with a title
for the aggregate. And they came up with the
grapes of wrath and other writings. The title
proper of the aggregate usually serves as the
preferred title of the aggregating expression as
well as the preferred title of the aggregating
work. When an aggregating work is planned to be
a compilation of works by one agent, one option
is to create an authorized access point for the
aggregating work consisting of the -- aggregated
expressions, Stein Beck in this case. Followed
by access point for work based on the preferred
title of the aggregating work.
I've included citation number in red if you
want to view the condition and option in the
toolkit.
In a linked data environment, this authorized
access point would be the value of a nomen string
related to the work but most of us are not yet
working with the linked date environment so
ignore that for now. In a MARC21 environment, we
would use a combination of values in the 1XX and
245 shields shown here. Boring, but familiar.
The third option -- wait a minute. Now, you
might choose instead toe use a conventional
collective title in this situation. And this is
in fact an option in the toolkit. In this case,
because the expressions have been aggregated
represent a variety of literary forms, short
stories, novel, non-fiction work, we use the
conventional collective title works followed by
the title selections. RDA recognizes what it
calls a distinguishing characteristic of the work
as I valid addition to an authorized access
point. We've taken advantage of that in this
7. case and used the series and the date of
publication since the library of America includes
several Stein Beck volumes that would also
qualify for the collective title works
selections.
You might also note that the short stories
embodied in this aggregate were not aggregated by
this aggregator, they came already aggregated by
an earlier aggregator, sort of pre-aggregated.
But we won't go down that rabbit hole.
The third option is to relate the
manifestation to one or more of the aggregated
expressions which would involve creating
authorized access point for one or more of the
selected expressions. You can pick and choose.
Here they're all recorded in 7X fields in MARC21.
You can also relate the manifestation to the
aggregating work at the same time as on the
previous slide, since relating the manifestation
to the various aggregated expressions and also to
the aggregating expression are not mutually
exclusive choices.
The fourth option is interesting. It shows
how we relate the aggregate manifestation to an
agent who's responsible for one or more of the
aggregated expressions when the manifestation is
not actually linked to these expressions. In
this case, we're trying to provide access to
Robert demot who wrote the notes section in the
volume. This often happens when you're
describing an agamentd aggregate. You want to
provide access. In a linked data environment,
this is tricky because what you technically
should be saying is Robert demot is the creator
of notes and the notes are an expression embodied
in the aggregate, even though they don't have a
title other than notes but we'd prefer not to
create the separate expression for notes. RDA
uses what we call a shortcut bypassing that --
aggregate manifestation. This shortcut is
expressed using the manifestation element
contributor agent expresser. In this case, the
contributor person of text element is used to
8. relate the creator of the notes Robert demot to
the manifestation. In MARC, we would record
contributor person of text as a resident
designator. You might say that's not a user
friendly decision but there's nothing to prevent
an application -- further refining the element by
declaring some local subelements like writer of
notes. RDA leads refinements up to the local
application.
So we've dealt with two of the three most
common types of aggregate, compilations of
expressions and aggregates resulting from --
What's left?
Parallel aggregates. Now, from the title
page you might not think what you have here is a
parallel aggregate, but look at the text
clarifies the situation. We have got the text of
three plays. We have them in the original Greek
edited by Jeffrey Anderson and we have them
translated to English. This is par the lobe
classical library so a diachronic work is also
one of four volumes that is issued over time and
numbered one through four so also a diachronic
work. The beginnings of diachronic works very
soon. Here we'll look at this volume as an
aggregate. This volume is actually two kinds of
aggregate at once. An aggregate expressions and
a parallel aggregate, the six different language
expressions. We can see how this plays out when
we apply the third option. For each play, we
have a Greek expression and an English expression
distinguished in the authorized access point by
adding the language and Henderson was involved in
both so we use his name to distinguish it
further.
As with the Stein Beck volume, we can also
apply the second option to provide an authorized
access point for the aggregating work -- and
conventional collective title, in this case for
the particular form of these works, plays. But
we didn't do that here because here we were
concerned only with the handling of those
parallel expressions. But the aggregating work
9. and its aggregating expression have consequences,
this is because an aggregating work can be
realized in one and only one aggregating
expression and vice verse. What's called a
one-to-one cardality. This means we have a work
expression lock for aggregates. The consequences
become clear when you ask what do we talk about
when we talk about -- soul. If you recall that
diagram from earlier, the general model for
aggregates, there was an aggregating work on the
right that represented the plan for creating the
aggregate manifestation.
The album on the left is the beet else album
-- it consists of 14 songs. The album on the
right is rubber soul as it was subsequently
released in the United States on the Capitol
label that consists of 12 songs. Some are the
same and some are missing. In terms of LRM, they
embody different aggregating works because they
involve slightly different selecting and
arranging. It gets even more complicated for it
when you consider what's happened in the years
since it first came out. New versions appeared
that have been reharvested, remixed, had bonus
tracks added, lots and lots of aggregates.
People generally think there's one rubber soul.
Wikipedia for instance, thinks there is one. How
do we deal with this situation?
Well, one approach RDA provides is something
called workgroups. And what is a workgroup?
Well, here's the glossary definition from
RDA. A workgroup is a group of two or more works
that have a common Appalachian assigned from a
vocabulary encoding scheme.
What this means is we can use a common access
point assigned from a vocabulary encoding scheme
to bring together closely related works. Let's
see how.
In RDA this common app ulation would be
represented by a nomen string. In our example
would be beat else rubber soul. In this case,
library of Congress, authorities. This probably
isn't what TL authority record was meant for but
10. it will do for this example. Once we have the
workgroup, we can use the nomen string as a
subject heading on a description of a book about
rubber soul. In the linked data environment, the
various aggregating works would be linked to
nomens that would share the string rubber soul
and identify the source vocabulary encoding
scheme library of Congress authorities. I should
-- none currently exists. We'll be touching on
workgroups when we talk about diachronic works.
So this completes the first part of this
webinar on aggregates. I've tried to make it as
unscary as possible because if you drop into the
new toolkit unprepared it can be a little scary.
So what have we learned?
We learned an aggregate is a manifestation
embodying multiple distinct expressions.
Realizes the plan for the aggregate, the
selection and arrangement of the aggregated
expressions, catalogers will typically describe
this aggregate manifestation in one or more of
the related expressions and works.
For aggregates described in Ma Ma the new RDA
toolkit can produce the same or similar results
by those produced by the current toolkit
depending on the options selected. An
aggregating expression realizes one and only one
aggregating work, and finally closely related
aggregating works can be brought together as a
workgroup. I'll try to answer your questions
about aggregates now before we move on to
diachronic works.
>> All right. The first question, do your
options for Ma Ma -- bound with items, works that
have been selected and bound together by a
private collector rather than a publisher?
>> Those would be items. I would see no
reason why not. But I'm very hesitant to go
outside my immediate area of semi-competence to
say that, but I would see no reason that that's
simply something that's taking place where the
resulting manifestation is a single item. And
there is aggregating work taking place. So yes,
11. I would see no reason why not. I would check
with others to be sure.
>> All right. Before I get to the next
question, we've had a few people mention it seems
like your audio level is going up and down. I'm
not sure if you're moving away from the mic at
times and moving closer. Some people have asked
that you get a little bit louder if possible.
>> Okay. I'm sorry.
>> No problem. Next question, how are we
supposed to treat conference proceedings in the
future?
There seems to be a conflict because on the
one hand a conference is considered to be the
creator and on the other hand the creator of the
aggregating work would be the editor?
>> Hmmm ... I'm afraid I hadn't considered
conferences. Maybe someone else could respond to
that. Is there any on?
>> All right. Well, maybe we can move on to
the next question.
>> I'll have to investigate that, yeah.
>> All right. Would a workgroup example be
ag the Christy's and then there were none and 10
little Indians?
>> I would say, well, and then there were
none and 10 little Indians, is that the exact
same expression just using a different title
proper.
>> All right. Relating members of a
workgroup through a nomen string seems tenuous.
Why not identify workgroup through IRI?
Why can't a workgroup be something like super
work?
>> Well, it was decided by the RDA steering
committee not to define another entity for
workgroups and to use instead a common nomen
string. And this was done, I believe, just to
satisfy the needs for collocation. There were
difficulties with defining an additional entity.
>> All right. Well, there are a couple more
questions, but I want to be mindful of the time.
We will keep those on the docket for our Q&A at
12. the end. We will move on.
>> Okay. Is it possible for you to break in
if people start saying my voice is fading?
Well, he's gone. All right.
>> Yes, we will do that, Ed.
>> Okay. Thank you.
So diachronic works. This will build on what
you have already learned about aggregates but
we'll be focusing just on the right-hand side of
that diagram you saw earlier. Instead of
aggregating work and aggregating expression,
we'll have a diachronic work with a diachronic
expression. And we'll have a manifestation that
may be an aggregate manifestation or plain
vanilla single manifestation. Here we go again.
By the end of this half of the webinar, I hope
you will be able to define a diachronic work,
understand the modeling of seriously and IFLALRM,
relate diachronic works to aggregates, explain
extension plans, explain transformation
boundaries and transformations. Explain the
cardality constraints on diachronic works. And
assign closely related diachronic works to a
workgroup using a shared nomen string. What is a
diachronic work?
This is the definition from the RDA glossary.
As with the aggregate definition, it's a bit
opaque on first encounter. I'll take you through
it step by step. First, it's a work, and is
planned to be embodied over time. This means
that the manifestation embodying this work is
expected to evolve over time. As we'll soon see,
there are various ways this can happen. You'll
notice that word plan too. Like the aggregating
work, a diachronic work is mainly a plan. In
this case it's a plan to be embodied over time.
Sounds a little creepy. So where did this come
from?
RDA's diachronic works originate in the
modeling of seriously in IFLA LRM. LRM treats
any serial as a distinct instance of the work
entity. Hence, these are diachronic works.
LRM also tries to be compatible with the
13. modeling of seriously in the ISSN where each ISSN
is said to identify -- all this entails what's
called a WEM lock or a work expression
manifestation lock, which we'll get to next.
In LRM, serial work is a specific kind of
aggregating work, aggregating the content of a
series of issues. Like an aggregating work, a
serial work represents a plan for aggregating
this content over time, but not the content
itself.
Now it gets tricky because serialize are
issued over time. For this reason, LRM treats
each serial as a unique work. In doing so
entails something called a WEM lock, this is like
the work expression lock we saw for aggregating
works and then some. Not only is a serial work
realized in one and only one expression, but that
expression is embodied in one and only one
manifestation.
You'll see the implications to this later in
the presentation.
While LRM models serials that is resources
issued in discrete parts over time, it doesn't
explicitly model aggregating resources or other
resources issued over time.
The makers of RDA decided to do so. RDA
strap lates from the LRM modeling of serials to
similarly model all -- this means the modeling of
any resource issued over time is based on the
modeling of serials and aggregates in LRM.
There's embodied content on the one hand,
expressions of one or more works, and a plan for
embodying this content over time on the other.
As a result, LRM serial work becomes RDA's
diachronic work, with an expanded scope.
It also means the content embodied in a
diachronic manifestation can be a single
expression embodied over time. We'll see an
expression of this -- an example of this at the
very end of the webinar.
To support this expanded scope, RDA imported
some attributes from the RDA onyx framework from
the resource categorization or ROF. You may
14. remember ROF as the source for some earlier RDA
elements, media type, carrier type and content
type. The new attributes -- how diachronic
resources change or grow over time. The first
attribute, which ROF calls extension mode records
whether extension occurs through the addition of
discrete part manifestations or by replacing
content in a single manifestation. What often
happens in a looseleaf publication. The second
attribute, what RA calls -- issued over a limited
time span from those that are expected to go on
indefinitely.
And RDA combines these two ROF attributes
into a single RDA element called extension plan.
All works have extension plans. You can see here
how the two values of the two ROF attributes have
been combined to produce four RDA extension plan
values. Successive determinant plan, successive
-- integrating indemonstrator ant plan and also a
fifth value, static plan to cover when the
extension plan is basically not to extend. The
resource comes out all at once, like most books,
most movies and most resources. Anyway, we are
not interested in these so we'll ignore the
static plan and instead introduce the four
diachronic plans one by one using examples.
Our first example is a census bullet from the
1900 -- bullets were the breaking news of their
day, published whenever new interesting
information came out from the census and the
government wanted to publicize it. They ceases
once the census itself stopped. In terms of the
ROF, its extension mode is succession that is
multiple discrete parts of the bullet -- plan to
stop at some point. This translates into an RDA
extension plan value of successive determinant.
Next example.
This is something a bit more familiar, a
periodical, the journal nature. Like the census
bulletins, it has an extension modes of
succession, comes out every week or nearly every
week, but unlike the census bulletins nature's
plan to continue indefinitely, at least so long
15. as libraries keep subscribing to it. It's
indetermine ant, and this translates into an RDA
extension value of successive indetermine ant,
which RDA calls a serial work. Bear in mind that
this serial work is not the collection of issues
you're used to thinking of as a serial. That's
the manifestation. The serial work is the stuff
involved in the plan, the title, the frequency,
the layout, editorial policy and so on. If you
recall the diagram of an aggregate from earlier,
it's the framework, the right-hand side, not the
left side. Not the content of the expressions
embodied in the manifestation.
This is sort of hard to wrap your head around
but it's important. We have two more examples.
Now we arrive at integrating works,
looseleafs and websites mainly. This is the
website of the 2017IFLA library Congress. It's
over. Its website persists because there's
interesting stuff there, though it's hiding. If
you click on the calendar and select Congress
program and pick a session, you can download the
papers presented at that session. It's still
useful but the important thing is it's over and
they always planned for it to end so its
extension determination is determine at. It's a
website, it's complicated, content comes and
goes. For us, this means its extension mode was
integrating so its RDA extension plan is
integrating determine at. Finally, there's the
website that goes on changing indefinitely.
Online newspapers are this way. They have
articles. But the articles may get updated as
the news gets updated, unlike a print newspaper.
And people comment on the articles, and there's a
lot of other stuff, weather, ads, sports scores,
share prices, it changes over time. Lots of
websites are this way. The RDA toolkit, your
local library catalog, your library website, your
University website, you get the idea.
So this has an extension mode of integrating
like the IFLA conference and an extension
determination of indeterminate. So its RDA
16. extension plan is integrating indeterminate. And
this leaves the static extension plan, which we
won't be discussing because this webinar is about
diachronic works. Sorry.
So how do you catalog a diachronic work?
Very carefully.
First, you select issue to serve as the bases
of the description. Maybe you choose the
earliest or lowest, maybe the latest. For
example, for the English speaking library
community, we choose the earliest or lower
numbered part of successive diachronic work. For
aggregating resources, we all prefer the latest
iteration mainly because the Earlier have
disappeared.
As you record this choice in a note, in RDA,
you use an element called note on issue part or
iteration used as basis for identification of
manifestation.
It's a mouthful. This is a case where you
might find yourself missing the old instruction
numbers. In MARC21 you record this field in 588
and it's common practice to record the source in
the issue, for example, the cover or caption, or
for integrating resources, the date the cataloger
reviewed the resource. This is called data
provenance and this is a good time to take a look
at how it applies to diachronic works.
You've already learned about data provenance
in a previous webinar but we are going to look at
it here too. This because data provenance is
important for diachronic works. By their nature,
diachronic works change over time and so do the
value of their elements and to track the changes
we use data provenance. Basically when the value
of an element like frequency applies only to a
limited set of issues or limited period of time,
we try to record this fact. When it applies to a
limited set of issues, we can use the scope of
validity element. When it applies to a limited
period of time, we can use the related time span
of work element. Of course you can record this
as a note on work and be done with it, but that's
17. too easy. To be truly useful in a linked data
environment, you want to record it as a metadata
statement. And then add the time span for which
this metadata is true.
Now, in that linked data environment, we do
this by transforming this three-part metadata
statement, my journal has frequency monthly, into
a metadata work. Basically we have my metadata
statement has related time span of work,
March 2012 to July 2015.
Just a note. We can do this to any element
that's subject to change, and any three-part
metadata statement can be transformed into a
single metadata work and linked to data
provenance.
This is a neat trick that's very helpful in
linked data. For this webinar, I guess it's just
interesting.
Two things about recording this stuff in
MARC21. MARC21 uses a single subfield to record
both elements. MARC21 doesn't distinguish them.
In fact it didn't know they existed because they
didn't until now.
On the other hand, due to its long and
complex evolution, MARC21 records this kind of
data provenance in a variety of places. For
variant titles it's in subfield F, for related
diachronic works, it's in subfield G and so on.
The important thing is to recognize that all
subfields contain the same sort of data, the
scope of validity for the other data in the
field. So we use this scope of validity for all
changes that aren't important enough to trigger a
new work.
So what happens if a change is big enough to
trigger a new work. Before we get to that, time
for some new vocabulary. Because of the WEM
lock, anything issued over time is a unique work
and each significant change and each new version
triggers a new work. RDA calls this triggering a
transformation. The point at which it happens
when a work is treated as a new work is called a
transformation boundary. Basically when you
18. define major and minor changes, you're defining a
transformation boundary. Fortunately for me,
such things are dependent on community
applications and policy statements, so I won't be
going into them in any detail. Suffice it to say
that a transformation boundary is a stage at
which a work is treated as a new work separate
from the existing work.
Now, for a static work, this boundary is the
familiar one between treating something as a new
expression and treating it as a new work.
Because static works can have more than one
expression. For diachronic works, because of
that WEM lock, it's simpler, everything is a new
work within reason. The main question is has the
title changed enough for this to be a new work.
Otherwise, the answer is always yes. If it comes
out on a different carrier, it's a new work, a
different language is a new work. New works
everywhere. So what about transformations?
Diachronic works, transformations generally
fall into two broad categories. Transformation
by policy, which a work created by changing the
scope or editorial policy of work. And
transformation by audience, a work created by
change the audience of the work. For this
webinar, we'll ignore the fuzzy. Transformation
by policy covers the normal range of changes. A
change of title, a journal splitting into two
journals, two journals merging into one journal,
basically the works are -- in MARC21 we record
the authorized access point in new work in field
75 the record of the -- we record the authorized
access point for the original work in 780 of the
record for the new work. Transformation by
extension plan is a little different. A change
in extension plan may signal the end of one and
the beginning of another.
The content of occasional updates now mingles
with the content of that old print work. But a
change of extension plan may not displays an
existing work. Most online versions are
integrating versions of print succession works.
19. As with aggregates, a single aggregate can
represent both a collection and augmentation, so
a transformation can represent both change of
policy, say a switch to an integrating plan and a
change of audience to attract remote readers.
You emphasize what matters most in your
environment.
This brings us to transformation by audience.
There are three common categories of audience.
Audiences for a particular carrier format,
audiences for content in a particular language,
and audiences with a particular geographic focus.
MARC21 has fields for handling the first two
of these and a more general field, 775 called
other edition entry for the third. We use the
broader transformation by audience in MARC21
field 775 and -- editions aimed at consumers,
people living in a particular zip code. There's
an endless number.
So the WEM lock succeeds at splitting these
works apart. How can we bring them back
together?
Historically we've used two mechanisms.
Related entry notes recorded in MARC21 field 760
to 787. This standard method used in the ISSN
network and in serial Saos cataloging generally.
And authorized access points for expressions
built on the authorized access point for a common
work, for example, a translation built on the
authorized access point for a work in the
original language. Due to the WEM lock, the
second method is no longer an option. A foreign
language version of a journal is identified by an
authorized access point based on its own title,
not the title of the journal in the original
language. So what do we do?
I don't know.
But we do have a new friend, appellation of
workgroup. Has been drawn from the library of
Congress' authorities of vocabulary and coding
scheme. Or had a mechanism that identified a
given value as applying to a workgroup but we are
not there yet. Another candidate for an
20. appellation of workgroup is linking of ISSN or
ISSNL. Bear in mind there's currently no
vocabulary encoding scheme that includes all
potential workgroups or relationships. So
relationships that is MARC21 field 760 to 787
will remain the primary mechanism for bringing
together related diachronic works for this time
being.
Finally, single works, where the content is
realized in one distinct expression, like a
novel, or in this case a work of investigative
journalism. Single works are also frequently
issued over time and they're more common than you
may think. In the 19th
Century works were
frequently issued first over time as faskells,
then once this diachronic work was complete, as a
static work. Works of long fiction were often
actually composed over time with the author
adjusting the ongoing plot in response to
readers' reactions. Periodicals and newspapers
often contain serialized contributions. A
television series based on a single work or the
hand maids tale which started out as a single
work and became a diachronic work in the
television version. They may first be released
as a diachronic work in installments and later
made available as a static work on DVD or some
other media. Finally the internet age serialized
-- ongoing interaction with readers. How will we
handle diachronic works that are single works?
That remains to be seen. But they will have
a transformation by extension plan relationship
with any closely related static works and maybe
we'll share an appellation of workgroup with
them. We haven't gotten there yet.
And that brings us to the end. To summarize
what I hope we've learned in this half of the
webinar, a diachronic work is a work that is
planned to be embodied over time. The RDA
modeling of diachronic works is extrapolation
from the IFLA model of serials. A diachronic
work may be determinant or -- a diachronic work
is realized in one and only one expression and
21. that expression is embodied in one and only one
manifestation.
We call this the WEM lock.
All relationships among diachronic works are
among instances of the work entity and involve
some transformation of the content or the plan.
Any diachronic element can have an associated
time span or scope of validity. An agency can
associate the preferred title with a particular
relative time span such as earliest or latest
based on community preference. And finally,
closely related diachronic works, like closely
related aggregates, may be brought together as a
workgroup using a common nomen string.
And finally, questions?
>> All right. Thank you so much. We have
had some questions come in. We are going to go
down what we have left and please feel free to
put your questions into the chat space. Could
you explain why the creator of the aggregating
work is not put in MARC100.
>> The creator of the aggregating work can
be put in MARC100.
If, for example, a single person is -- the
translation, say, that were done by a single
translator are brought together, that translator
creates the -- well, no, no. That's not the
aggregating work, is it?
I believe the creator of the aggregating
work, if known, they're usually not even known in
many cases, can be used as the creator -- can
have a creator relationship, will have a creator
relationship to the aggregating work. So that
when you have the aggregating work as an entity
in linked data, the creator of that aggregating
work won't be identified. It's simply not one of
these choices in the condition option group for
creating the authorized access point. The
authorized access point has to be viewed not as
it currently is in MARC where you have a 1XX
field and 245 field and the 1XX automatically
implies something. If you wanted to create a
bizarre authorized access point for something,
22. you could do that. The authorized access point
doesn't imply that the first part of that
authorized access point refers to the creator.
It's simply used for facilitating access to the
work, to the manifestation, to whatever the
library's making available.
>> All right. This is Dan. Can a workgroup
have several aggregating expression?
>> Can a workgroup have several aggregating
expressions?
Yes. Well, workgroup, again, you have to
bear in mind that a workgroup is simply a group
of aggregating works that share a common nomen
string.
And so there is no entity called a workgroup.
It is simply a bunch of works that are brought
together or can be brought together by using this
common nomen string. And that common nomen
string identifies the workgroup only because it's
associated with a vocabulary encoding scheme that
clarifies it as being a nomen string for a
workgroup.
>> All right.
>> So the relationship with the aggregating
expressions at all except in that they're
realized in the aggregating works that share the
nomen string.
>> All right. Next question. Is the WEM
lock for diachronic works really necessary if,
for example, a print and electronic versions of a
serial might show certain differences over time,
wouldn't it be enough to say these are different
expressions and might roll a second question into
that?
Someone else said I still find the WEM lock a
mystery. Why?
And what are the implications for a user
searching a catalog?
>> The WEM lock originates with the
IFLA LRM. It in fact is pragmatically speaking,
what we do today. It's very rare with serials
cataloging that we have where we represent things
as expressions of works. When we have different
23. editions aimed at different audiences, those are
currently distinguished by treating them as
separate works. I think it becomes more
problematic in the case of the single works that
we're used to thinking as monographs that simply
have come out over time.
And I think a lot of this will be dealt with
in specific applications. That is, particular
communities will determine how they are going to
approach this and what remedies they may come up
with for their own users in terms of bringing
things together and that sort of thing. It's not
something that RDA itself addresses, but policy
statements and so on can be used to deal with
specific possibly problematic situations.
>> All right. Next question. Would
multi-part monographs be successive determinant?
>> If they come out over time, yet.
>> Would every DVD released and every
release by a service like Hulu or Sling be a
different work?
>> It all depends on the expressions that
are in those different things. If, say, a
streaming service makes available the parliaphone
version of Rubber Soul, then those both represent
the same expression of Rubber Soul. If they do
some fiddling with it, that becomes a different
aggregating work.
And it all depends where you want to draw the
line. That is, you can look at, even today, we
can look at things and say how much fiddling with
something is necessary before we treat it as a
separate work?
How much fiddling is necessary before we
treat it as a new expression?
We make compromises all over the place and I
assume we'll continue to do so.
>> Okay. Are print on demand vanity press
novels that the author frequently edits
considered diachronic works?
>> Yes, they would be integrating works. In
fact, integrating works are -- they're something
that I think we really need to explore in much
24. greater detail than we have already, especially
in scientific journals, even things that we think
of as equivalent to the print journal, the online
journal is drifting away. A medical journal
where an article gets corrected, the text of the
article itself is now getting corrected online.
And so the online article no longer corresponds
to the article in the print issue of the journal.
So integrating resources, basically the
consequences of things occurring online is
creating all sorts of interesting new situations.
>> Okay. Not all diachronic works are
serials, only indeterminate diachronic works are,
so why does the WEM lock apply to all diachronic
works?
>> It applies because the RSC decided to
extrapolate from the LRM modeling of serials to
apply to all works issued over time. I mean,
that's the short answer.
>> All right. Well, we're actually at the
end of our list of questions, but we've still got
plenty of time. So if we've got any other
questions -- here's one. Can you give an example
of cataloging event related publications such as
a festival, catalog, now that the event no longer
exists in RDA, assuming I have got that right.
>> Could you repeat that?
>> Can you give an example of cataloging
event related publications such as a festival
catalog, now that event no longer exists in RDA?
I'm assuming I have got that right.
>> Cataloging an event ...
Hmmm.
I'm having trouble understanding exactly what
that's referring to.
>> Maybe we can ask for a clarification
there. I have got another question. She said
she may email with that later. Coming back to
the aggregates, I don't see the advantage of
splitting up two editions of the same essays as
belong -- essays has been revised. Is it a lot
of work comparing the two editions?
What's the benefit?
25. >> That's what I meant when I said that
people will decide how much fiddling is
necessary. That is, of course not. It all
depends, if you're a specialist library
collecting all the editions of this thing, then
yes, maybe you would want to track at that level
of granularity. But in most cases we are going
to catalog without even being aware that the
difference exists. And I expect that individual
cataloging communities will make pragmatic
decisions about how much change is a significant
difference from one aggregating expression to
another.
I think Rubber Soul is a case where there is
significant difference, but in other cases, like
the case given of one of the essays is revised,
probably not. But again, that would be up to a
local cataloging community.
>> [Away from mic].
>> I think you answered the last question
there about Rubber Soul. Again, we're at the end
of the list of questions. I guess before we wrap
up, Ed, do you have any closing thoughts?
>> Just that this is all a brave new world
to me too. If people would like me to just talk
about anything in particular in more detail, I
would be willing to do that.
>> All right.
>> When I went through this on my own, it
lasted longer.
>> That is always the case, yes. A question
from Timothy. Do you foresee fewer new records
or title changes?
>> Well, for serials, the thing is that
everyone everywhere is pretty much locked into
what the ISSN network does. That is, the things
that qualify as a major title change for creating
a new serial now are generally based on the
instructions in the ISSN manual, which in the
past were incorporated into the toolkit. But I
don't enadvicage any dramatic change there. I
think the main difference I see for serials is
simply what occurs with translations and few
26. other cases where currently we treat things as
expressions that will now be treated as separate
works.
>> All right.
Here's another one. To clarify, what we now
call multi-part monographs can be either static
or diachronic depending whether the entire
publication is in hand at the time of cataloging?
>> Yes. And that's again, why I say that I
think there will be pragmatic solutions to many
of these questions. In another presentation, I
brought up the encyclopedia Britannica, which in
its earlier editions came out in fasicles and
subsequently the publisher put out volume title
pages. You know, you were supposed to accumulate
your fasicles, bind them together and slap on the
volume title page you got. So it came out as a
diachronic work, but once it was complete, its
initial, whatever you want to call it,
diacrownisity, became invisible. If your library
required that set from a collector with the
volume title pages all saying 1797, even though
it had come out over the previous several years,
there would be no way you could know that. As I
say, I think there will have to be pragmatic
solutions to some of the cases where something
comes out initially diachronicly simply because
it was a pragmatic decision by the publisher to
put it out diachronicly, like say a dictionary
where you start putting it out as soon as you can
so the first volume is available, but you have no
plan to ultimately change the way you do it once
you get to Z. So initially it comes out
diachronicly but subsequently it comes out
static.
I imagine there will be pragmatic decisions
taken by cataloging communities on how to
approach situations like that.
>> All right. Will your updated edition of
RDA serials catalog include more with cataloging
with linked data in mind?
>> That depends on how long I can delay
publication.
27. Right now, I think there are just initial
talks taking place between the RFC and the
BIBFRAME community over the relationship between
the vocabularies. I would hopefully my plan now
is to start out with MARC. And if a stable
version of BIBFRAME reflecting the new toolkit is
available before publication, then I'll certainly
incorporate that.
So it's all a question of timing, really.
>> Okay. I think we're at the end of the
list, unless there's something that I missed. So
we will say thanks to you, Ed, thank you to our
audience and to so many of you who responded and
helped out and discussed things in the chat
space. We really appreciate everything and we
hope to see most of you back next week for the
next session in our special topic series. I hope
everyone has a wonderful rest of your day or
evening, depending where in the world you are.
And we'll see you again soon. Thanks, everyone.