Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
J. McRee Elrod wrote: How nice to have Heidrun join Bernhard as a voice of reason from Europe. Germany may save more than the euro zone! Mac had me blushing violently here... I'm not so sure about the euro zone, but I believe it is a very helpful experience to find out that there is more than one way of doing things, that things are actually being done differently elsewhere. When I once spent some months in the UK I found out that I wasn't able to exchange a broken light bulb, and I had to get a friend to help me. For him it was just one short movement of the hand to get the light bulb out - which I hadn't managed on my own. Then we found out why: I had expected light bulbs to be affixed in the same way as they are in Germany (where you have to swivel them), but they're fitted in differently in the UK. It was a simple, yet memorable experience. I've found only one thing with which to disagree. "Nabokov, Vladimir, 1869-1922. Bend sinister (English. With additional materials)", So nice to see the "preferred title" include main entry. I do think "preferred title" is misleading as a term, when it includes more than a title. "Preferred citation" would make more sense, as well as being in accord with scholarly practice. I like "preferred citation" very much. Actually, the thing Mac and I disagree about (but haven't had time to go into more deeply yet) is the question of main entry as such. Although I had given the title of the work in the form that RDA constructs access points for titles of works here, I've been thinking for a long time that we should get rid of the concept of main entry altogether - aas RDA hasn't really managed, I believe. Although there is no more talk about "main entry" in RDA, the basic distinction between works which are entered under author, under corporate body, or under title is still there in the rules for constructing authorized access points for works and expressions. I'm convinced that users don't need this information in order to help them with their bibliographies. In many German catalogs you won't even see what the main entry is (unless you are a librarian), and nobody seems to miss this. Compare the following two entries in the Southwest German union catalog: http://swb.bsz-bw.de/DB=2.1/PPNSET?PPN=112695671&INDEXSET=1 http://swb.bsz-bw.de/DB=2.1/PPNSET?PPN=276186850&INDEXSET=1 (I hope these links are really persistent as they should be.) The first has main entry under author, the second has main entry under title (it's a collection of essays) - does the difference seem striking to you? In the case of edited collections, we also have a general discrepancy between what librarians think the main entry should be (the title) and how scholars construct their citations (starting with the editor). When Group 1 entities are mentioned in cataloging, e.g. in added entries or footnotes, I think this should (at least in the medium term) be all changed to links via a control number. Have a look at this entry: http://swb.bsz-bw.de/DB=2.1/PPNSET?PPN=355236370&INDEXSET=1 which is the printed version of a doctoral thesis. Under "bibliographic context" there is a link to the corresponding e-book edition. What lies "behind" that is not a standardized text string, but simply the control number for the other record (called a PPN, Pica production number) 35523503X. Obviously this should still be shown in some textual way to the users. But which textual form to use does not necessarily have to be fixed by rules. It could be handled quite flexible in each catalog (perhaps even according to the preferences of each individual user). In this catalog, in the link to the other manifestation the work is not "named" in the conventional form (which in our cataloging tradition would be: "Kostrzewa, Krzysztof: Advanced computational methods in identification of thermo-acoustic systems"). Instead what's taken automatically from the linked record and shown here is simply the title and statement of responsibility (not altoghether a bad idea, I think). As I've already hinted at in some earlier post, we do not use standardized text strings (which in RDA are called "authorized access points") in order to record relationships, but instead we make links to different records voa the control number. E.g. all bibliographic records belonging to the same author are linked to his or her authority record. And all parts of a multi-part work or a monographic series are linked to the corresponding main record for this multi-part work or series itself. Let me openly admit that there is considerable self-interest in my campaign for getting rid of main entry altogether: The reason is that the German and the Anglo-American cataloging tradition quite often differ not on the entries as such, but on which of these entries is the main one. I'm afraid that this will cause a lot of problems and a huge amount of work when Germany will switch over to RDA. T
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
Quoting "Tillett, Barbara" : Quick note to mention that the manifestation to work bit can be handled with a placefolder at the expression level. Yes, of course. But I don't think that affects the issues here. As for the whole/part relationships and mapping to 505, that also is covered in RDA. Whether it would be displayed as a note as now with MARC or done otherwise in the future with links between the whole and parts will depend on systems. I don't think that's accurate. I think whether systems can display it will depend on how the bibliographic data is structured. It's data that drives systems, not the other way around. What we're trying to figure out is how to structure the data so that the user display will make sense. It appears that if the data for aggregates is not explicitly structured in some whole/part relationship it may not be possible to make that clear to users. Plus, we don't seem to be able to find a defined data structure that corresponds to the instructions in RDA. (I personally think that a contents note would be very useful for some situations, like listing the chapter headings of a book by a single author. I think this is useful information but it shouldn't have to be structured like an embedded work in order to be included.) You may be interested in seeing a training tool used by The MARC of Quality folks (Deborah and Richard Fritz - they just did a demo here at LC yesterday) which beautifully demonstrates such links in a non-MARC environment - I hope they can show their views to others at ALA or soon thereafter. It would "show" you how all of your questions in this thread work nicely with RDA and FRBR. Yes, I'm familiar with their product. Deborah and I talked recently about trying to create data for some aggregates, especially ones having the same work appear both in an aggregate and separately. After that, though, I think we need to find someone who can load the data into a triple store so we can run some actual linked data processes on it. For a while I've been wishing we had a test suite of RDA data in RDF. That would help us try out some of these ideas and see if the data elements as defined can support the retrieval and displays that we might want. It seems that it would really help if folks could see some results. We may be getting closer to that. kc - Barbara Tillett (personal opinion) -Original Message- From: Resource Description and Access / Resource Description and Access [mailto:RDA-L@LISTSERV.LAC-BAC.GC.CA] On Behalf Of Karen Coyle Sent: Friday, January 06, 2012 4:46 PM To: RDA-L@LISTSERV.LAC-BAC.GC.CA Subject: Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates Quoting JOHN C ATTIG : - Original Message - | Karen said: | >RDA does not have a data element for contents; there is nothing | >similar to the MARC 505. Karen is not quite correct. The contents (parts) of a resource are considered Related Works in RDA. The formatted contents note is a structured description of the related work -- a list of the titles of the parts of the resource. If you look at the MARC to RDA mapping provided in the RDA toolkit, you will find that field 505 maps to RDA 25.1 (Related work). In the examples of structured descriptions of related works under 25.1, you will find examples of contents notes with the relationship designator "Contains" used as a caption. Note: I am looking at this from a data creation point of view. Data creation is not nearly as maleable as notions and ideas. My question is: can we create valid data using FRBR and the published RDA properties? RDA: http://rdvocab.info/ FRBR: http://metadataregistry.org/schema/show/id/5.html John, there is no contents note in the list of RDA elements. In that I am sure I am correct. And MARC 505 is a note. Therefore, nothing that is the same as the 505 exists in RDA *as defined*. It might seem the same conceptually, but I am struggling to find data definitions that support it. If the RDA 25.1 (and I note that in an earlier message to me you were the one who referred me to 27.1.1.3) is a work/work relationship then it cannot be used to indicate a relationship between a manifestation and a work. It isn't clear to me how a manifestation can have a related work, since manifestation in FRBR must manifest an expression, not a work. It isn't clear to me what kind of relationship a Work can have to a manifestation given the way that they are defined in FRBR. Also note that FRBRer, as defined in the metadata registry, has no "related Work" property. It does have a work/work whole/part relationship. The RDA definition of related Work is: "A work related to the work represented by an identifier, a preferred access point , or a description (e.g., an adaptation, commentary, supplement, sequel, part of a larger work)." I read this as a set of work/work relations
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
Quoting Heidrun Wiesenmüller : Predominant and non-predominant would need to be relationships between the expression and the manifestation. It's not a characteristic of the work or the expression. This may be true for different ways of modeling aggregates. In my model I'd have an aggregate work with two parts; I don't see why it shouldn't be possible to give these parts of works attributes like "main component of aggregate work" or "secondary component of aggregate work" (I admit this would be a new attribute to FRBR, something which could only be applied to aggregate works). What type of entity would be "part" be? I'm thinking that there is no such entity as "part" but that a work can be a "is part" of another work. Taking into account that the work is a single entity that may be related to any number of expression/manifestations it cannot be "secondary" since that is what it is only in relation to the manifestation being cataloged. Primary and secondary, therefore, have to be relationships. In a sense, a Work is always whole, even if it is part of another work. If it didn't have "wholeness" it couldn't be a work. Note that this would not affect the work "Introduction" as such, but only in its role as part of the aggregate work. The supposedly clever thing in my model (it may turn out not be that, of course) is that the "Introduction" is wearing, so to speak, two hats at the same time: One for its role as an individual work and one for its role as a part of the aggregate work. If the introduction were to be published independently later on, this would give you an ordinary FRBR tree of a work (the introduction), an expression, and a non-aggregate manifestation. In my diagram, this would mean another arrow from the node E (W2) to a new manifestation which would only embody this single expression. Yes, that is how I imagine the graph to "grow." But I guess I'm not sure what the "part" box is in your model -- it appears to be a Work that has the characteristic of being a part of the aggregate. I also note now that your Fig. 3 has an expression that realizes more than one work, which I believe is problematic. It definitely violates the current FRBR model, but then you are advocating for change in that model. Of course my model might turn out not be feasible at all. It's certainly still at an experimental stage, and new aspects are bound to come up. But up to now I haven't seen an argument in this thread convincing me that I'm on a completely wrong track. Would the "Work part" have the same properties as the work described on its own? W1 type: Work editor: Jones, Jane work title: Ecology collection subject: trees subject: streams W2 type: Work author: Smith, John work title: Essay on trees subject: trees WP7 type: Work part part of: W1 author: Smith, John work title: Essay on trees subject: trees Is this what you were thinking of? I'm not sure what you mean with "title search" here. Do you perhaps mean a title search on manifestation level? That's not what I have in mind. I rather imagine a system like OCLC's FictionFinder (by the way: will that ever go online again?), which at the first step presents not manifestations, but only works. But I believe it searches on all titles. Otherwise, one would have to know the original language title in order to retrieve the work. Unfortunately Fiction Finder doesn't seem to be running at the moment so I can't check that. The other option is that all manifestation titles would need to be alternate titles in the work. However, I don't think we can design for a single system structure. Surely some systems will provide a full keyword access on any entities. Sorry I can't follow your argument any better than this (which has probably not been satisfactory). We must have got our wires crossed somehow. No, I actually think we're getting very close. It would be useful to have examples, so if you can mock up examples of your ideas I think that would help. Then we can refer to specifics. What I really want is a real time white board for drawing diagrams... this kind of thing is very hard to do in email. (And I greatly appreciate your excellent command of English, as there would be no communication at all without it.) kc In the end I think I am agreeing with you that we need a whole/part relationship that connects the contents of manifestations to the manifestation. The current whole/part relationships in FRBR may not be sufficient, or it might be that we aren't clear about how they work in RDA. Yes, I think it's obvious that we can't do without a whole/part relationship _somewhere_. The question of where is still open to debate, I think. My proposal is to have it neither on manifestation nor on expression level, but modeled as an aggregate work with separate parts. Heidrun -- - Prof. Heidrun Wiesenmu
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
Karen Coyle wrote: Quoting Heidrun Wiesenmüller : Firstly, the system should be able to distinguish between an aggregate work and an "ordinary" work. The whole/part relationship (from my approach) would not be enough as ordinary works can have parts as well. So there should be some sort of flag for an aggregate work, perhaps a new attribute (aggregate / non-aggregate). This would require a new FRBR concept, I believe. It would require a new attribute for the work entity. This would certainly have to be approved by the FRBR Review Group. I don't think it would upset the FRBR universe in any dramatic way, though. The aggregate work, as it is a work, needs -among other things - a preferred title of its own (core element in RDA). This might be something like "Bend sinister (With additional materials)" (perhaps also: "Nabokov, Vladimir, 1869-1922. Bend sinister (English. With additional materials)", taking into account which expression of the novel has been used of the aggreagte work. I'll have to think on that some more). I don't think it can have the language in it, since language is an Expression-level concept. That makes this quite complex, though, because now I don't see a clear relationship between the translation and the original. Yes, that got me thinking as well. It seems somehow wrong to have a typical attribute on expression level like the language in the name of the aggregate work. On the other hand, the alternative model deliberately does _without_ an aggregate expression (there are only expressions of parts of the aggregate work). The language could be deduced from the expression which is embodied in the aggregate manifestation, though. I grant that there is a complexity here which needs to be explored some more. In the case of augmentations, it might be useful to flag the predominant work in the aggregate work somehow (Casey A. Mullin suggested that in one of her posts in this thread). Then we'd also have the possibility to present non-predominant works at the end of such a list, or perhaps present them to the user only via a separate link (e.g. saying: "There are also minor works of Ms Famous, such as: Introduction, in: Nabokov, Vladimir, 1869-1922. Bend sinister (With additional materials). Show minor works as well?" Predominant and non-predominant would need to be relationships between the expression and the manifestation. It's not a characteristic of the work or the expression. This may be true for different ways of modeling aggregates. In my model I'd have an aggregate work with two parts; I don't see why it shouldn't be possible to give these parts of works attributes like "main component of aggregate work" or "secondary component of aggregate work" (I admit this would be a new attribute to FRBR, something which could only be applied to aggregate works). Note that this would not affect the work "Introduction" as such, but only in its role as part of the aggregate work. The supposedly clever thing in my model (it may turn out not be that, of course) is that the "Introduction" is wearing, so to speak, two hats at the same time: One for its role as an individual work and one for its role as a part of the aggregate work. If the introduction were to be published independently later on, this would give you an ordinary FRBR tree of a work (the introduction), an expression, and a non-aggregate manifestation. In my diagram, this would mean another arrow from the node E (W2) to a new manifestation which would only embody this single expression. Of course my model might turn out not be feasible at all. It's certainly still at an experimental stage, and new aspects are bound to come up. But up to now I haven't seen an argument in this thread convincing me that I'm on a completely wrong track. Now if somebody looks for the work "Bend sinister" in an English version, the system would look for the English expression (in my diagram: E (W1)) and show all three manifestations linked to this. The system would also note that one of the manifestations is an aggregate one (there would not have to be an attribute "aggregate" on this level, I believe, as the aggregation is obvious from the fact that more than one expression is embodied). In this case, it would display further information about its environment. The display might look somewhat like this English version of: Nabokov, Vladimir, 1869-1922. Bend sinister - Published: New York : Vintage International, 1990 - Published: Alexandria, Va. : Time-Life Books, 1981, c1947. Together with: Ms Famous: Introduction. In: Nabokov, Vladimir, 1869-1922. Bend sinister (With additional materials) - Published: New York : H. Holt, [1947] Would that be an answer to your concerns or have I misunderstood the problem? I think your example works if there is a whole/part relationship between Bend sinister and the introduction, but not if the introduction is coded as "embodied in"
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
Thomas said: >The lack of an authorized access point doesn't mean the entity >disappears or can't be accounted for. Control numbers and >identifiers, as well as the collection of associated elements >(including title by itself), can be used to point to an entity. I'm trying to picture this in a footnote or bibliography. I thought one goal of RDA was to "play with others". This turns our back on centuries of scholarly practice. Codes and\or "title by itself" would not work in the larger world. And why all the new terminology? What's wrong with "edition", "citation", "main entry", "subject and added entries", etc.? Are we using new jargon to make ourselves feel important? Mystify the uninitiated? I don't suppose reverting to known terms is part of the mandate of the RDA rewrite? __ __ J. McRee (Mac) Elrod (m...@slc.bc.ca) {__ | / Special Libraries Cataloguing HTTP://www.slc.bc.ca/ ___} |__ \__
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
> -Original Message- > From: Resource Description and Access / Resource Description and Access > [mailto:RDA-L@LISTSERV.LAC-BAC.GC.CA] On Behalf Of J. McRee Elrod > Sent: January 7, 2012 11:12 AM > To: RDA-L@LISTSERV.LAC-BAC.GC.CA > Subject: Re: [RDA-L] Some comments on the Final Report of the FRBR Working > Group on Aggregates > ... > So nice to see the "preferred title" include main entry. I do think > "preferred title" is misleading as a term, when it includes more than > a title. "Preferred title" only includes the title element. Additional elements can be added to construct the authorized access point for the work, which is what the entire string as a heading, including the creator prepended, is called. The authorized access point itself is only one method for identifying an entity. It carries the baggage of all the old main entry rules, which apply to works (series included). The lack of an authorized access point doesn't mean the entity disappears or can't be accounted for. Control numbers and identifiers, as well as the collection of associated elements (including title by itself), can be used to point to an entity. For example, RDA envisions scenarios in which one is not forced to create a name-title heading for a series as the only means of identification. > > In this thread, the WEMI relationship has been spoken of as vertical, > and the whole part one as horizontal. It seems to me we need a third > term for the whole part relationship; the whole part relationship is > not horizontal; as Heidrun has pointed out in other posts, the part is > secondary to the whole. Relationships are reciprocal and can convey this meaning of main and secondary. For example: "Contains" and "Contained in" convey very well the nature of the relationship as to which is whole and which is part. In addition, Numbering of Part is an RDA relationship element that can be added to qualify even further the relationship with a numeric designation, which only adds to the clarification of what is whole and what is secondary. > Translations and editions are horizontal, not > parts. They can be, but only as expressions to their expression counterparts. All cataloging conventions to date have assumed a primary relationship from the work down to the different language translations and editions. Thomas Brenndorfer Guelph Public Library
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
Quoting "J. McRee Elrod" : In MARC, adding a code for aggregate to LDR/06 should do it. Code "c", I assume, means a collection of separate items, as opposed to bound withs. We use it for, as an example, a collection of manuscript letters or sermons. We have to consider that we may not be creating "records" in the sense of MARC, but "graphs" that bring together data entities. The "Work" will be used in a lot of different contexts. So there is no code that will cover the whole graph. That information must be carried in the relationships between things. In this thread, the WEMI relationship has been spoken of as vertical, and the whole part one as horizontal. It seems to me we need a third term for the whole part relationship; the whole part relationship is not horizontal; as Heidrun has pointed out in other posts, the part is secondary to the whole. Translations and editions are horizontal, not parts. Absolutely! Thanks, Mac, for teasing this out. kc __ __ J. McRee (Mac) Elrod (m...@slc.bc.ca) {__ | / Special Libraries Cataloguing HTTP://www.slc.bc.ca/ ___} |__ \__ -- Karen Coyle kco...@kcoyle.net http://kcoyle.net ph: 1-510-540-7596 m: 1-510-435-8234 skype: kcoylenet
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
On 06/01/2012 20:34, J. McRee Elrod wrote: James Weinheimer said: Probably, the issue of aggregates is also more related to physical materials than to virtual resources. Absolutely not. While we first encountered the aggregate work problem with papers given at continuing education symposia, we now encounter it with constituent parts of websites. Many electronic publishers have parts of their websites for particular series, subjects, types of users, etc. But if it is just the conference papers etc., everything can be handled as they have always been done, as you point out. What I meant was that with physical materials, it is much easier to know what actually is the "aggregating entity" because you are looking at a book with lots of conference papers, the journal issue with different articles, and so on. From my experience, it is much more difficult for the cataloger to discover precisely what is, or is not, part of the same website, especially if you are looking at specific parts. The webmaster of the specific site knows this much better than anyone else. I am still trying to find better examples, but here are a couple that should illustrate it. You may catalog an electronic document such as this http://library.usask.ca/theses/available/etd-09192010-154127/unrestricted/dissertation.pdf, but you remain completely unaware that it is actually part of this: http://library.usask.ca/theses/available/etd-09192010-154127/. Many times because of the structure of the site, you are looking at a specific article or section, and there is no indication that the item is part of a series. Here's another example: http://www.spunk.org/texts/intro/sp000281.txt, is actually part of "The Spunk Library" http://www.spunk.org/, but you would not know it except through creatively playing with the URL. Frame sites (i.e. using the or coding) can be especially confusing, since it can turn out that you are only looking at one part of a whole. Here is an example. You see this page and everything looks OK http://www.gooddocuments.com/philosophy/skimming_m.htm, but it is actually designed to be seen in this way: http://www.gooddocuments.com/philosophy/skimming.htm. With printed materials, the "aggregating entity" will almost always be much more obvious but online, can easily be hidden. And, to return to dynamically-created mashups, while it may be theoretically possible to catalog them according to FRBR, to do so in reality would be more tedious than finding needles in a haystack and probably not worth the effort. So, in a case of an online conference with multiple papers (all virtual), the current methods can be used. But the methods can fall apart for many materials online. -- *James Weinheimer* weinheimer.ji...@gmail.com *First Thus* http://catalogingmatters.blogspot.com/ *Cooperative Cataloging Rules* http://sites.google.com/site/opencatalogingrules/
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
Quoting Heidrun Wiesenmüller : Firstly, the system should be able to distinguish between an aggregate work and an "ordinary" work. The whole/part relationship (from my approach) would not be enough as ordinary works can have parts as well. So there should be some sort of flag for an aggregate work, perhaps a new attribute (aggregate / non-aggregate). This would require a new FRBR concept, I believe. The aggregate work, as it is a work, needs -among other things - a preferred title of its own (core element in RDA). This might be something like "Bend sinister (With additional materials)" (perhaps also: "Nabokov, Vladimir, 1869-1922. Bend sinister (English. With additional materials)", taking into account which expression of the novel has been used of the aggreagte work. I'll have to think on that some more). I don't think it can have the language in it, since language is an Expression-level concept. That makes this quite complex, though, because now I don't see a clear relationship between the translation and the original. In the case of augmentations, it might be useful to flag the predominant work in the aggregate work somehow (Casey A. Mullin suggested that in one of her posts in this thread). Then we'd also have the possibility to present non-predominant works at the end of such a list, or perhaps present them to the user only via a separate link (e.g. saying: "There are also minor works of Ms Famous, such as: Introduction, in: Nabokov, Vladimir, 1869-1922. Bend sinister (With additional materials). Show minor works as well?" Predominant and non-predominant would need to be relationships between the expression and the manifestation. It's not a characteristic of the work or the expression. Now if somebody looks for the work "Bend sinister" in an English version, the system would look for the English expression (in my diagram: E (W1)) and show all three manifestations linked to this. The system would also note that one of the manifestations is an aggregate one (there would not have to be an attribute "aggregate" on this level, I believe, as the aggregation is obvious from the fact that more than one expression is embodied). In this case, it would display further information about its environment. The display might look somewhat like this English version of: Nabokov, Vladimir, 1869-1922. Bend sinister - Published: New York : Vintage International, 1990 - Published: Alexandria, Va. : Time-Life Books, 1981, c1947. Together with: Ms Famous: Introduction. In: Nabokov, Vladimir, 1869-1922. Bend sinister (With additional materials) - Published: New York : H. Holt, [1947] Would that be an answer to your concerns or have I misunderstood the problem? I think your example works if there is a whole/part relationship between Bend sinister and the introduction, but not if the introduction is coded as "embodied in" the manifestation. In the latter case you have: W Nabokov.Bend sinister E Bend sinister. English M Bend sinister. NY, vintage, 1990 M Bend sinister. Alexandria, T-L. 1981 M Bend sinister. NY, Holt, 1947 W Ms Famous. Introduction E English M Bend sinister. Alexandria, T-L. 1981 Do a title search on "Bend sinister" and you retrieve the introduction if it has been coded in this way. Even if you can find an efficient way to "de-duplicate" at this point, the information does not exist to determine that the Introduction is a "minor" work, because every work is a work, and major and minor depend on the context. I believe that at this moment we do not have a way to make that distinction using FRBR. In the end I think I am agreeing with you that we need a whole/part relationship that connects the contents of manifestations to the manifestation. The current whole/part relationships in FRBR may not be sufficient, or it might be that we aren't clear about how they work in RDA. kc Heidrun -- - Prof. Heidrun Wiesenmüller M.A. Hochschule der Medien Fakultät Information und Kommunikation Wolframstr. 32, 70191 Stuttgart Tel. dienstl.: 0711/25706-188 Tel. Home Office: 0711/36565868 Fax. 0711/25706-300 www.hdm-stuttgart.de/bi -- Karen Coyle kco...@kcoyle.net http://kcoyle.net ph: 1-510-540-7596 m: 1-510-435-8234 skype: kcoylenet
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
Heidrun said: >I don't see any problems here which couldn't be solved by sound >underlying data structures on the one hand and a proper design of the=20 >display on the other. How nice to have Heidrun join Bernhard as a voice of reason from Europe. Germany may save more than the euro zone! >Firstly, the system should be able to distinguish between an aggregate >work and an "ordinary" work. In MARC, adding a code for aggregate to LDR/06 should do it. Code "c", I assume, means a collection of separate items, as opposed to bound withs. We use it for, as an example, a collection of manuscript letters or sermons. >The whole/part relationship (from my approach) would not be enough as >ordinary works can have parts as well. YES. We do chapter level records, including records for prefaces and bibliographies, for some electronic publishers. They offer parts of their works in mix and match packages. It is so refreshing to read a post from someone who seems to occupy the same bibliographic world as SLC. In offlist correspondence with this brilliant woman, I've found only one thing with which to disagree. >"Nabokov, Vladimir, 1869-1922. Bend sinister (English. With >additional materials)", So nice to see the "preferred title" include main entry. I do think "preferred title" is misleading as a term, when it includes more than a title. "Preferred citation" would make more sense, as well as being in accord with scholarly practice. On the other hand, series citation should only include series title. We know who wrote the past issues of a series, but not who will write the next one. >There may be also a way to record the title of the introduction not >simply as "Introduction", but perhaps in a more meaningful way as >"Introduction [to Nabokov, Vladimir, 1869-1922. Bend sinister]" When we prepare part records for electronic monographs, and the part title is not distinctive, we use 245 10 $a.$p, e.g., $pIntroduction, Preface, Bibliography. It seems better to me to gather by title the nondistinctive parts of a monograph, rather than to gather all the prefaces, introductions, and bibliographies. In this thread, the WEMI relationship has been spoken of as vertical, and the whole part one as horizontal. It seems to me we need a third term for the whole part relationship; the whole part relationship is not horizontal; as Heidrun has pointed out in other posts, the part is secondary to the whole. Translations and editions are horizontal, not parts. __ __ J. McRee (Mac) Elrod (m...@slc.bc.ca) {__ | / Special Libraries Cataloguing HTTP://www.slc.bc.ca/ ___} |__ \__
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
Quoting Karen Coyle : Manifestation 1 is embodiment of Expression A Manifestation 1 is embodiment of Expression B Manifestation 1 is embodiment of Expression C something else occurs to me about this model: there is no place for a title proper for each of the expressions -- If A is the whole, and B and C are individual works in A, then where are the titles proper for B and C? Casey, you might be able to answer this one since this seems to be a common situation in music data. kc This to me seems inferior to a whole/part relationship, but perhaps it is sufficient. The other option is to have (and this is hard to do without diagrams) w1 e1 m1 (the aggregate work) w2 e2 m2 (one of the essays) w3 e3 m3 (another essay) m1 has part m2 m1 has part m3 Again, without mocking this up it's hard to imagine what users would see. However, I think this is conceptually valid linked data. kc Of course there will always actually be an expression, but a cataloger may choose not to identify it for local reasons, and if someone needs it later, it can be added. This has been discussed by the JSC and with Gordon Dunsire when looking a the element set on the Open Metadata Registry, and we felt this was a workable approach that enables practice while allowing the structure to be complete in systems. As for the whole/part relationships and mapping to 505, that also is covered in RDA. Whether it would be displayed as a note as now with MARC or done otherwise in the future with links between the whole and parts will depend on systems. You may be interested in seeing a training tool used by The MARC of Quality folks (Deborah and Richard Fritz - they just did a demo here at LC yesterday) which beautifully demonstrates such links in a non-MARC environment - I hope they can show their views to others at ALA or soon thereafter. It would "show" you how all of your questions in this thread work nicely with RDA and FRBR. - Barbara Tillett (personal opinion) -Original Message- From: Resource Description and Access / Resource Description and Access [mailto:RDA-L@LISTSERV.LAC-BAC.GC.CA] On Behalf Of Karen Coyle Sent: Friday, January 06, 2012 4:46 PM To: RDA-L@LISTSERV.LAC-BAC.GC.CA Subject: Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates Quoting JOHN C ATTIG : - Original Message - | Karen said: | >RDA does not have a data element for contents; there is nothing | >similar to the MARC 505. Karen is not quite correct. The contents (parts) of a resource are considered Related Works in RDA. The formatted contents note is a structured description of the related work -- a list of the titles of the parts of the resource. If you look at the MARC to RDA mapping provided in the RDA toolkit, you will find that field 505 maps to RDA 25.1 (Related work). In the examples of structured descriptions of related works under 25.1, you will find examples of contents notes with the relationship designator "Contains" used as a caption. Note: I am looking at this from a data creation point of view. Data creation is not nearly as maleable as notions and ideas. My question is: can we create valid data using FRBR and the published RDA properties? RDA: http://rdvocab.info/ FRBR: http://metadataregistry.org/schema/show/id/5.html John, there is no contents note in the list of RDA elements. In that I am sure I am correct. And MARC 505 is a note. Therefore, nothing that is the same as the 505 exists in RDA *as defined*. It might seem the same conceptually, but I am struggling to find data definitions that support it. If the RDA 25.1 (and I note that in an earlier message to me you were the one who referred me to 27.1.1.3) is a work/work relationship then it cannot be used to indicate a relationship between a manifestation and a work. It isn't clear to me how a manifestation can have a related work, since manifestation in FRBR must manifest an expression, not a work. It isn't clear to me what kind of relationship a Work can have to a manifestation given the way that they are defined in FRBR. Also note that FRBRer, as defined in the metadata registry, has no "related Work" property. It does have a work/work whole/part relationship. The RDA definition of related Work is: "A work related to the work represented by an identifier, a preferred access point , or a description (e.g., an adaptation, commentary, supplement, sequel, part of a larger work)." I read this as a set of work/work relationships. There are no Manifestation to Work relationships in FRBR. There is a whole/part relationship between manifestations in FRBR 5.3.4.1. While it might make logical sense to point from a manifestation to "related works" the underlying structure of FRBR does not support this as far as I can tell. Therefore, if the RDA properties are associated definitionally each with a FRBR en
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
Karen Coyle wrote: I need to back up here and say that we are talking about a linked data model, not a fixed record, so the idea of "marking" a W as "secondary" simply doesn't exist. Just noting that in my "alternative model", I think this could be done after all. If you look at figure 2 in my "additional diagrams" paper, the place to record an attribute "secondary" would be the entity marked "Part 2 of Aggr. Work". Actually, I believe this may be a good argument for having the model like this (although it looks a bit complicated by having "Work 2" and "Part 2 of Aggr. Work" together at the same time), and not simply having a simpler arrangement like this: Aggregate work Part 1: Work 1 Part 2: Work 2 Indeed Work 2 couldn't then be marked "secondary" as it is not secondary "as such". It is secondary only with regard to the part it plays in the aggregate work - and this can be captured, I think, in my model. Heidrun -- - Prof. Heidrun Wiesenmüller M.A. Hochschule der Medien Fakultät Information und Kommunikation Wolframstr. 32, 70191 Stuttgart Tel. dienstl.: 0711/25706-188 Tel. Home Office: 0711/36565868 Fax. 0711/25706-300 www.hdm-stuttgart.de/bi
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
Quoting Casey A Mullin : But regardless of whether the aggregate work and constituent work are directly related, or related by virtue of a common manifestation, W/E 2 and 3 need not be identified for the user in this example. As I stated previously, we may construe their existence, but the user need only be presented with W/E 1 and the three M's that embody it. I don't see how this could be done, algorithmically if the parts have been given a relationship of "embodied in/expressed/" from the M to the W. Note that each W could be expressed and manifested in a number of different instances, so this is not a property of the work nor of the expression. Nor, in the case of a main work and a secondary work, is there any visible difference in the coding of this primary relationship. If 1, 2 and 3 are all coded identically, there is no way to know which one is the aggregate and which are the individual works. I need to back up here and say that we are talking about a linked data model, not a fixed record, so the idea of "marking" a W as "secondary" simply doesn't exist. Any such information needs to be in the relationship of the W to the M. That was the example that I gave with this: w1 e1 m1 (the aggregate work) w2 e2 m2 (one of the essays) w3 e3 m3 (another essay) m1 has part m2 m1 has part m3 I believe this is the only way to convey the information such that it can be displayed as you wish to the user. kc I hope that makes sense. Casey On 1/6/2012 1:52 PM, Karen Coyle wrote: Quoting Casey A Mullin : Manifestation 1 (embodies E 1) Manifestation 2 (embodies E 1) Manifestation 3 (embodies E 1,2,3) Is "embodies" a part/whole relationship? Because you only have one option: Manifestation expresses Expression So this would be: Manifestation 3 (expresses E1) Manifestation 3 (expresses E2) Manifestation 3 (expresses E3) and each of those is a separate declaration of a relationship. Without a whole/part relationship in there somewhere there is nothing that says that one of them includes the others. They are all equal. The M -> E relationship is not a whole/part relationship. That might be ok, but again I ask about the user view - would all three of these be displayed to the user if a search retrieved them all? And would there be anything to indicate to the user that one of them is a larger package for the other two? kc Entities we IDENTIFY (that is, fully so, beyond oblique mention in statement of responsibility or other notes): Work 1 Expression 1 Work/Expression 2-3 definitely exist, but their existence is implied, and need not be identified using RDA's methods (access points, identifiers) Manifestations 1-3 The use case would be thus: User is presented with Work/Expression 1, then the 3 Manifestations embodying it. (Presumably, W/E 1 are the primary entities of interest.) If the user wanted to probe deeper, they could learn about the existence of W/E 2 (the supplemental material) through its oblique mention in the description for M 3. As for how RDA turns this model into practice, the answer lies in Chapter 17. Whatever the nature of a resource (aggregate or not), RDA only requires at a minimum that the "predominant or first-named" work/expression be identified. This language ought to be clarified in light of this expanded understanding of aggregates; that is, what is "predominant or first-named" in an aggregate resource? For example, in a compilation, the aggregate W/E is favored in our current MARC implementation scenario (resulting in title main entry), but it needn't be. Rather, the encoding should be agnostic as to which entities are selected as the most salient for identification. It is not that FRBR is incompatible with our needs going forward, it is that MARC is inadequate to encode FRBRized data (which is probably why LC is ignoring Chapter 17 in the current implementation scenario; it just can't be applied correctly). Casey On 1/5/2012 5:36 PM, Karen Coyle wrote: Maybe what we need to do is develop some use cases and see how they would turn out. I'm less concerned about the cataloger view than the user view. You've probably run into some description of looking at FRBR from "bottom-up" vs. "top down." Some folks consider the cataloger view to be bottom-up (from the thing in hand to the Work) while the user view is top down (from the Work to the item on the shelf). Here are three items. I don't know if they are enough to illustrate what worries me: 1. LC control no.: 47003534 LCCN permalink: http://lccn.loc.gov/47003534 Type of material: Book (Print, Microform, Electronic, etc.) Personal name: Nabokov, Vladimir Vladimirovich, 1899-1977. Main title: Bend sinister [by] Vladimir Nabokov. Published/Created: New York, H. Holt [1947] Description: 242 p. 21 cm. 2. LC control no.: 89040559 LCCN permalink: http://lc
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
Diane Hillmann wrote: I keep hearing a couple of threads in this conversation that I think need further examination. The first is that there needs to be 'agreement' on how to handle these situations, before anyone can do anything. This implies that we need to retain the notion that it's critically important that we minimize the impact of those who stray from the 'true path' because they make our jobs harder. I really think this idea needs to hit the dumpster now, if not yesterday. If we're entering a world where the FRBR model is used to help us link together information at a number of levels of description, it seems to me that we all benefit from those who add important detail to the shared environment. That old straightjacket 'granularity consensus' is one of the things that marginalize us in the world where the old boundaries around what we do and don't do gets in our way. We certainly should think of FRBR as a dynamic system which is not "finished" and closed once and for all, but will have to evolve and expand. It also should be flexible enough to provide a variety of approaches, so there is certainly nothing wrong with having modeling variants. Still, I believe it would be a good thing to have those variants moving within certain boundaries marked out by the FRBR system, so that they adhere to the "FRBR basics". If an application does not completely follow the "FRBR basics" this would not be something inherently "bad". It might be absolutely useful and fitting for the application in question, and, of course it might still be possible to provide meaningful connections between this application and other applications which move fully within the FRBR boundaries. Perhaps the discrepancies can also point to aspects where the "FRBR basics" should be improved, and in this case the community might want to incorporate them. But as long as this hasn't happened, the application should be openly called a "non-completely FRBR" application. To my mind, aggregates are such an important thing that the modeling of them should be included in the FRBR basics. This does certainly not mean there can only be one way of doing it; we might accept more than solution as being within the boundaries of FRBR. I should also point out that the DCMI/RDA Task Group built a number of cataloger scenarios, including one that included a festschrift (http://wiki.dublincore.org/index.php/Cataloger_Scenarios#Scenario_2:_A_collected_work). The TG name has been changed to the DCMI Bibliographic Metadata Task Group, but the wiki data from the old group has been moved (is in the process of being moved anyway, but the Cataloger Scenarios are all moved). I'd be happy to entertain discussion on whether or not this scenario makes sense (leaving aside the question of whether anyone will do it), but suggest that maybe a new subject line would make sense. Thanks for pointing that out. If I understand the scenario correctly, it shows an "aggregate as work-of-work" approach, making use of whole/part relationships. This is, I find, an entirely plausible and intuitively reasonable way of looking at something like a festschrift (by the way: my students always find it quite hilarious when I tell them about this beautiful Germanism). My "alternative model" is rather similar (but not identical) to this. But what bothers me is that this approach is the one _not_ presented in the Final Report for the modeling of aggregates: The Working Group's general model does not have a part/whole relationship at any stage which seems counterintuitive. Note that there is mentioning of part/whole relationships in Appendix B, reflecting the view of some members of the Working Group. Looking at the proposed FRBR amendment on p. 6-7 of the report, I'm at a loss to decide whether a modeling using whole/part relationships would be acceptable (in the sense of: being within the boundaries of FRBR basics) as an alternative to the main model of the Working Group, or not. One of the things I dislike about the report is that it very often doesn't spell out things clearly. So when trying to find out what they _really_ mean, there is a lot of speculation and exegesis involved. This is not only my own impression but that of some of my German colleagues, as well. Sorry about putting this criticsm so plainly. It is not meant in a personal way at all. Of course I understand that the problem is a "devilish" one indeed, as Karen put it, and I'm also sure that the members of the Working Group did their very best in a difficult situation when, obviously, a consensus was hard to reach. Still, the result is not something I can be comfortable with. Heidrun -- - Prof. Heidrun Wiesenmueller M.A. Stuttgart Media University Faculty of Information and Communication Wolframstrasse 32, 70191 Stuttgart, Germany www.hdm-stuttgart.de/bi
Re: [RDA-L] Some comments on the Final Report of the FRBR Working Group on Aggregates
Karen, My concern is about examples like the one I gave, although it may have been imperfect. Assume that the preface is one that is considered important enough to be noted in the catalog record, one that is written by someone famous. You want to include an entry for that preface under the name of Ms. Famous. It's a Work, so you need a Work entry. (Also, you can't indicate a creator without having a Work entity.) You want to indicate that the Work is a part of the Manifestation along with the main text. Adding a new Expression-Work unit is not a clear part/whole relationship (which is what Heidrun is pointing out). And again I'm interested in how this would be displayed to a user, how this set of relationships will be brought together in a display. Perhaps one could treat this secondary "Work" as a related manifestation? However, in FRBR structural terms, all Works are Works, there are no "lesser Works," so there would be no difference between this preface and an essay in a set of essays. I don't see any problems here which couldn't be solved by sound underlying data structures on the one hand and a proper design of the display on the other. Firstly, the system should be able to distinguish between an aggregate work and an "ordinary" work. The whole/part relationship (from my approach) would not be enough as ordinary works can have parts as well. So there should be some sort of flag for an aggregate work, perhaps a new attribute (aggregate / non-aggregate). By the way, if one were to transform AACR/MARC data into FRBR/RDA data by means of algorithms, I think there would be lots of indicators in the records (like 505 or 490/8XX) pointing out whether something is an aggregate or not. Augmentations are different in that respect (you'd have a hard time analyzing them mechanically, as probably the only information which could be used are things like "edited with an introduction and notes by ..."). Therefore, for something like the augmented edition of Nabokov's novel the flagging would be something which has to be done by the cataloger who has decided to treat it as an aggregate work in the first place. The aggregate work, as it is a work, needs -among other things - a preferred title of its own (core element in RDA). This might be something like "Bend sinister (With additional materials)" (perhaps also: "Nabokov, Vladimir, 1869-1922. Bend sinister (English. With additional materials)", taking into account which expression of the novel has been used of the aggreagte work. I'll have to think on that some more). There may be also a way to record the title of the introduction not simply as "Introduction", but perhaps in a more meaningful way as "Introduction [to Nabokov, Vladimir, 1869-1922. Bend sinister]" or some such like. This would not be imperative, thpugh, as it can be made clear in a different way as well: The environment of the work "Introduction" (i.e. the aggregate work and/or the other works) can be displayed to the user. So, assuming the introduction in question is by a Ms Famous, and that's why we want to bring it out in the catalog in the first place (by the way, I'd rather like to think of catalogers as not using criteria like this), and somebody is looking for all the works of Ms Famous, they might get: Famous Work #1 Famous Work #2 Introduction, in: Nabokov, Vladimir, 1869-1922. Bend sinister (With additional materials) Famous Work #3 In the case of augmentations, it might be useful to flag the predominant work in the aggregate work somehow (Casey A. Mullin suggested that in one of her posts in this thread). Then we'd also have the possibility to present non-predominant works at the end of such a list, or perhaps present them to the user only via a separate link (e.g. saying: "There are also minor works of Ms Famous, such as: Introduction, in: Nabokov, Vladimir, 1869-1922. Bend sinister (With additional materials). Show minor works as well?" Now if somebody looks for the work "Bend sinister" in an English version, the system would look for the English expression (in my diagram: E (W1)) and show all three manifestations linked to this. The system would also note that one of the manifestations is an aggregate one (there would not have to be an attribute "aggregate" on this level, I believe, as the aggregation is obvious from the fact that more than one expression is embodied). In this case, it would display further information about its environment. The display might look somewhat like this English version of: Nabokov, Vladimir, 1869-1922. Bend sinister - Published: New York : Vintage International, 1990 - Published: Alexandria, Va. : Time-Life Books, 1981, c1947. Together with: Ms Famous: Introduction. In: Nabokov, Vladimir, 1869-1922. Bend sinister (With additional materials) - Published: New York : H. Holt, [1947] Would that be an answer to your concerns or have I misunderstood the problem? Heidrun -- --