Associate, collocate, disambiguate, infuriate

Regular followers of my twitter account will know that I regularly complain about uniform titles. I know that’s not an RDA-approved term, but I don’t currently have the luxury of a wholly RDA-approved catalogue, and time passes particularly slowly in the tech services department. It’s also the term currently used for the 130 and 240 MARC fields, a format to which we remain shackled, and in which someone will probably write my eulogy.

In my view, uniform titles are some of the most misunderstood and misused fields in cataloguing. I say this not to look down on those who remain baffled (for I was myself baffled right up until last week) but because they don’t really serve the purpose for which they were intended. I’ve seen so many records with uniform titles they didn’t need, inserted by cataloguers who were no doubt simply following someone’s rules.

According to the 2005 revision of AACR2, a uniform title had the following functions:

Uniform title. 1. The particular title by which a work is to be identified for cataloguing purposes. 2. The particular title used to distinguish the heading for a work from the heading for a different work. 3. A conventional collective title used to collocate publications of an author, composer, or corporate body containing several works or extracts, etc., from several works (e.g., complete works, several works in a particular literary or musical form).1

In other words: a cataloguer might choose, create, or otherwise determine a particular title to associate with a given work; to disambiguate from other works of the same name; and to collocate works with different names within a title index. It’s a form of authority control; titles and author/title combinations are often given authority records of their own. (Hence the tie-in to this month’s GLAM Blog Club theme, ‘control’.)

Association, disambiguation, collocation: that’s a lot to ask of one field, and I can grudgingly accept that most of it made sense within a book or card catalogue. Remember, this refers not to collocation of books on a shelf (that’s what classification schemes are for) but for collocation of entries within a catalogue. Until quite recently, a catalogue was simply a collection of indexes: title, author, subject. Librarians wanted these entries arranged in a particular order, and created filing rules to ensure this order was adhered to.

During the development of MARC in the 1960s (led by the incredible Henriette Avram), a format originally designed to automate the production of catalogue cards, the layout of a MARC record mirrored the layout of an AACR-compliant catalogue card.2 The first paragraph, mapped to the 1XX set of tags, included the main entry—an author, corporate body or meeting, but also uniform titles where the work in question had no author, but another, different, title proper. The second paragraph featured the title proper and edition statements, and were recorded in the 2XX set of tags. Because a uniform title could conceivably end up in multiple places on a catalogue card, depending on whether there was an author attached, the developers opted to create two fields.

(In the first example, it’s considered advantageous to have all the Bibles entered directly under title, collocated in the title index, then disambiguated by language, version, year [in that order]. In the second example, we already have an author, but the item in hand has a different title to that by which it is more commonly known.)

130 1# $a Bible. $l English. $s New Revised Standard. $f 2003 
245 14 The new interpreter's study Bible : $b New Revised Standard version 
       with the Apocrypha. 

-------------------

100 1# $a Xenophon $e author.
240 10 $a Hellenica. $l English.
245 12 $a A history of my times / $c Xenophon ; translated by Rex Warner.

Apropos of nothing, this also explains how the title statement, arguably the best-known MARC tag, was assigned the odd number 245:

To represent the second paragraph of the catalog card (title and edition), the MARC developers logically chose the 200 range of tag numbers. Because they had reached 130 in the first paragraph, and were trying to proceed by tens, the first choice for the title tag was 240. Continuing by tens, the 250 for edition and 260 for publication information were also defined. Law librarians, however, asked where the uniform filing title, which they used for filing, should be placed. Since uniform title preceded title proper on catalog cards, it seemed logical to maintain this arrangement in the MARC record, so the 240 was reassigned for uniform title, and the 245 tag (halfway between the filing title and the edition) was created for the title proper.3

Law librarians: warping MARC logic since 1965.

Anyway, back to titles. When I was learning to catalogue, I struggled with the reasoning behind uniform titles, as I had no concept of a title index to base them on. Never having used a card catalogue in my life, I saw no reason why anyone would use a browse function instead of a keyword search. (I stand by this view.) Even the idea of collocation doesn’t work in a keyword-based OPAC setting, because I can dive straight to the record I want, with no reason (or, indeed, ability) to view records on either side in any index. Viewing a list of records in browse mode is so… old-fashioned. (Besides, if there is no reason to do this, there is also no reason to create uniquely identifying main entry headings… (taps noggin))

The main cause of my frequent twitter complaints about uniform titles are the preponderance of unnecessary titles in our catalogue, specifically those relating to online resources. Because MARC-based catalogues entail a flat record structure, we can’t (yet) nest different expressions of a work, to use RDA parlance, within a work-level authority record. Instead we’re stuck with one record per manifestation, whose titles we have to disambiguate. Because so many resources exist in both print and online versions, and often a library has access to both, the obvious differentiating factor is whether it’s online or not. Therefore a resource might be titled: 130 1# $a Economist (Online) to distinguish it from the print version.

The problem is when cataloguers take this to mean that every online resource must be so titled, even when it has no print equivalent. This has the effect of 1) cluttering the catalogue with unnecessary uniform titles and 2) furthering the antiquated narrative that print resources are the norm, and online is the exception. There’s no point in creating 130 1# $a Digital humanities quarterly (Online) if it has only ever existed in an online format.

Associate, collocate, disambiguate, infuriate! (sigh)

Tell you what, I can’t wait for my wholly RDA-compliant, IFLA-LRM-based, fully FRBR-ised catalogue of the future (now with 25% more acronyms!). I look forward to being able to bring expressions together under a work-level authority, and have this tree display intuitively in an OPAC. I look forward to not having to use titles as disambiguators for indexing purposes. I look forward to relinquishing some of my control over the form and display of titles within a catalogue.4

I look forward to dispensing with uniform titles, for they have well and truly reached their use-by date.


  1. As reproduced in the RDA Toolkit, accessed 29 April 2018. 
  2. For more on this fascinating topic, see Jo Calk & Bob Persing (2000). From Catalog Card to MARC, The Serials Librarian, 38:3-4, 349-355. DOI: 10.1300/J123v38n03_20 
  3. Ibid, p. 351. 
  4. For more on the history of uniform titles, see Jean Weihs & Lynne C. Howarth (2008). Uniform Titles From AACR to RDA, Cataloging & Classification Quarterly, 46:4, 362-384, DOI: 10.1080/01639370802322853 

I have some thoughts about ILL and document supply

During the last few weeks I’ve been helping out the Document Delivery (DocDel) team at work with some of their workload, chiefly processing other libraries’ physical items in and out, plus a bit of journal article delivery from our collections to other libraries. I am an almost complete newcomer to the inter-library loan and document supply side of the library business, and it’s been very interesting to see how the sausage gets made, so to speak.

Knowing the Share It! Resource Sharing Futures conference is coming up in a month’s time; noting that all the speakers are managers rather than coalface staff; and wondering if a newcomer’s experience of resource sharing work might be valuable, I decided to jot down my experiences of working in DocDel for the first time.

As always, these views are entirely my own and not those of my employer (but hey, if you wanna send me to #shareit2018 I won’t say no).

  • Resource sharing is surprisingly labour-intensive. I was astonished to find out just how much human intervention goes into an ILL request. When someone lodges a request through the online portal, or gives up and emails, or can’t wait until yesterday and calls up directly, that request is handled by at least three different people:

1) branch staffer at lending library receives book request, confirms book is present, retrieves book, applies book strap, posts book, tells ILL system this has been done
2) docdel staffer at receiving library receives book, applies second book strap with a barcode, writes down patron details on bookmark for easy reference, pops bookmark in book, sends book to pickup branch, tells ILL system this has been done
3) branch staffer at receiving library uses bookmark to locate book on hold shelf, checks out book to patron

And that’s if everything works properly. None of the above is particularly difficult work, but there are a lot of moving parts, and a lot to get right. It’s hard to automate attaching a barcode strap to a book, but we also have to include a little paper note reminding circ staff to change the item’s due date, and I despair every time I reach for a paperclip.

Within consortia using the same ILS, internal resource sharing has a lot less overhead. The BONUS+ consortium, comprising Australian and New Zealand university libraries running an Innovative system (Millennium or Sierra), allows patrons to request items from participating libraries without needing the involvement of document supply staff. The kicker is that you can use the lending library’s barcode without having to attach a new one, because all the ILSs can talk to each other. It’s brilliant! It also shifts a fair bit of the resource sharing workload from DocDel to branch library staff, who incorporate it into their usual circ workflows.

  • Human labour is often compensating for poor metadata and system design. In any given day, I routinely deny over 80% of article requests. It’s not because I get a kick out of saying no, but rather that the system we use, which hooks into the Libraries Australia Document Delivery network, thinks our collection includes items it doesn’t. From what I can tell, it matches a request to a holding library using a title, author, ISBN, ISSN, or a combination of these. It doesn’t appear to consider our actual holdings of a journal (i.e. the span of years or issues we have) or whether our licensing agreements permit supply of electronic items.

Having said that, our electronic journals have their holdings data recorded in an 856 $z field. Having this data in free text rather than fixed or coded fields surely makes it far harder for an ILL system to figure out what we actually hold. The licensing information is not recorded in a MARC format, is almost certainly inaccessible to the ILL system, and would differ hugely between libraries. Consequently I waste large amounts of time figuring out that no, we can’t actually supply items after all. It would be far better—not to mention faster—for all involved if the system knew not to ask us in the first place.

My colleagues in DocDel work very hard, but I can’t help but feel a lot of their labour is wasted. Surely we can do better?

  • Resource sharing is interesting and engaging work. While I wish I didn’t have to think so hard about whether or not we can supply an article, there’s still a lot of problem-solving in document supply, especially for rare and unusual materials. Most of that work is handled by the DocDel team leader, who understandably hasn’t let me anywhere near it yet. People ask for all sorts of weird stuff—theses, manuscripts, books that no library in Australia has, increasingly esoteric foreign-language films… if nothing else, it’s a great insight into certain academics’ current research interests. I can almost guess who has requested certain items just by looking at them!
  • For undergraduates, ILL and document supply are irrelevant. For higher degree students and staff, they’re a lifesaver. I never used ILL as an undergrad, chiefly because they were gonna charge me for it, and I was too broke for that. Ten years and two degrees later, I’m far more aware of the resources beyond my library’s walls, but I’m not sure all our patrons are. If users consider the first page of discovery layer search results to be the sum total of a library’s resource provision—and let’s be honest, most of them do—what incentive do they have to look elsewhere, or ask a librarian for help, or stumble upon that wondrous part of the library website titled ‘Document Supply’? I wouldn’t be at all surprised if most of our fervent ILL users discovered the service by accident. It’s usually followed by a ‘Where has this been all my life?!’ moment. It’s been right here, the whole time, but we could do with some more advertising. Also a better funnel for disappointed catalogue users.
  • There will always be a need for resource sharing services. In an age of open-access, preprint servers and Sci-Hub, where lots of people go out of their way to make their work (and others’) available online for free, plenty of people are questioning the need for library resource sharing services at all. As I see it, there are two main problems: people don’t know resource sharing exists; and when they do know, they often aren’t prepared to wait. Document supply is not a quick process—quite apart from being at the mercy of Australia Post, the use of substandard ILL systems and reliance on human labour means stuff just takes a while. That might work for document supply staff, but it doesn’t work for users.

No matter how often we might hear it, we will never have everything available for free on the internet. As long as there are paywalls, print books and rare items, there will be a need for resource sharing. But our systems and processes must improve. Placing a request and waiting for an item has a decidedly old-fashioned sheen to it. Like I said, I am new to this work, and I don’t profess to have any of the answers. But I sure wouldn’t mind trying to come up with some.

Classifying works on Indigenous Australian languages in DDC, UDC, LCC and Bliss

Wiradjuri to English dictionary

It’s no secret that the Dewey Decimal Classification (DDC) is racist, sexist, classist and every other -ist under the sun. I can’t truly say it’s of its time, because it’s still in wide use. Libraries around the world use DDC as their main classification system for physical materials. Aren’t we supposed to be better than that?

As a former local history librarian, our collecting remit naturally included materials by, for and about the various Aboriginal nations on whose land our city was built. In particular, my former workplace has a modest collection of works about local Aboriginal languages, both in-language and in English. Unfortunately they are a DDC library, and so these materials are all classified with the same call number. In a bigger library with materials on many different Indigenous languages, this would render the call number virtually useless.

In the interests of advocating for a better solution (not just for me personally but for other DDC libraries), I thought it might be interesting to briefly compare and contrast how Indigenous Australian languages are treated by four widely used general classification systems Dewey Decimal (DDC), Universal Decimal (UDC), Library of Congress (LCC) and Bliss. (Okay, so Bliss isn’t as widely used, but it deserves some attention.)

DDC: 499.15 (source: DDC 22, hardcopy)
– Treated as an afterthought. After devoting the majority of the 400s and Table 4 to Romance languages, the rest of the world is unceremoniously shoved in the 490s at the end of the schedule. The entirety of Indigenous Australian languages are accorded /9915 in Table 4. Clearly inadequate to describe the language diversity of an entire continent

UDC: 811.72 (source)
– UDC is broadly based on DDC but with a few major structural changes; here, languages and literature are co-located, and the 4XX schedule is not used.
– ‘Australian languages’ more clearly integrated into Table 1c, but again there is no subdivision or faceting for any individual regions or languages

LCC: PL7001-7101 (source, p. 324)
– PL LANGUAGES OF EASTERN ASIA, AFRICA, OCEANIA > Languages of Oceania > Austronesian, Papuan, and Australian languages > Australian languages
– This schedule is a bit better thought out, and less squashed. Also specifically lists about three dozen individual languages (though not my local ones, sigh)
– Individual languages are cuttered, not classified by any real measure—still essentially lumping all Australian languages together as one entity, which taxonomically isn’t much better than DDC / UDC, but if you needed to create a local cutter for a particular language that wouldn’t be difficult

Bliss: XJE (source, p. 43)
– X classification is still in draft (after all these years) so I will give the authors a pass on this, but just FYI: ‘Austronesian languages’ ≠ ‘Australian languages’. The former refers to a language family roughly around the Mekong Delta.
– Astonishingly—and I was really not expecting to see this in Bliss, of all places—the authors have actually properly classified individual languages! It’s a bit piecemeal, to be sure: initially a prefixing / suffixing divide, then the prefixed ones by multiple, dual or non-classifying (in a linguistic sense), followed by suffixing languages by geographic region. It is… idiosyncratic, but it’s a damn sight better than anything the other three came up with. I appreciate that this was given serious thought

In summary, it looks like Bliss is your best bet for classifying materials related to Aboriginal Australian languages. But if you’re in a position to create a local, culturally appropriate classification system (as is being done up in Galiwin’ku), totally go for it!