Sunday, 9 October 2011

#libcampuk11 session 1: cataloguing and classification (#hvcats)

(Library Camp UK 2011 home | my other posts about it)

I suggested this session in advance on the wiki. Adrienne stood up to pitch for it on the day - thank you!  The session pretty much ran itself - after intros (some 'expert' cataloguers, some who 'only catalogue to in house rules', some who 'will have to start cataloguing soon and haven't much experience yet', some systems folk and some just interested) various ideas and problems were discussed by the group.  Here are my notes (I've removed identities to protect the innocent...):

We started with a comment inspired by the introductions: "Lots of people say that they 'don't follow a proper standard'. So what is the point of major international standards? Maybe it's only the national libraries that need/want to follow such standards.  Perhaps RDA moves some way, with the idea of 'cataloguer's choice', towards allowing greater freedom for individual organisations to creat their own scheme."

"Maybe the most important thing is to get the data in in some form or other - 'mark it and park it'.  We need to ask what the motivation is for creating 'beautiful' records using loads of different obscure fields and all the right punctuation."

"Some places aren't doing that now. [E.g. libraries in private companies] But there is value in big institutions 'doing it properly'."

"But what is 'doing it properly'? Different users have different needs [example given of a researcher wanting to look at e.g. female authors of poetry in a given century, but we don't currently record genders of authors.] MARC is difficult to use - it's hard to meet the users' needs with it."

"It's hard to meet user needs because cataloguers don't get much feedback about how their cataloguing is used.  [Example: not having it explained that the date in the leader is what's used for date-limited searches, so it's really important that it's there... There are lots more examples like this]"

"Do we have feedback buttons on OPACs so that users can report mistakes in records? Are we keeping our user data and actually examining it meaningfully?"

"So how to standards fit into this? Do we want standards or not?"

"We want good standards. Web-friendly standards, e.g. URLs  for authors (i.e. linked data).  That's a different issue to on-the-ground cataloguing issues which are about how to identify which author a given work has, for e.g. (matching to authorities)."

"How good/useful is Dublin Core in this context?"

"There are lots of things available (Dublin Core, SXML, JSON...?) - stuff that helps cataloguers.  But some people (e.g. rare books) nee more fields/more granularity than available in DC."

"Dublin Core, or whatever, won't be the solution if we just use it in the old way - with a reliance on punctuation, not separating components of an author's name, etc."

"What about classfication? [Insert institution name here] has a frightful mess."

"[Company library] has a bad classification system because it's not fine-grained enough.  There are lots anf lots of not very similar books all at one number."

"What the point of classification? Why not just label the shelves?"

"If it's done well, classification marked in the catalogue can be used to subject retrieval."

"Does anyone actually do that?"

"Perhaps it could be done retrospectively for closed access material - classify, and then extract textual subject information from the class no. and display that on catalogue.  Then shelve by running nos, etc. as is more efficient."

"[A library] use complicated UDC but isn't browsed much - a mis-match of effort vs use."

"OCLC are using DDC for book recommendations.   But what is the point of DDC?"

"A legacy of the physical library. Class at the shelf is useful in a physical library for browsing.  The OPAC/online world hasn't yet replicated this yet, and that's a big issue."

"Also an issue for e-books."

"[A question from a systems librarian:] Do we enjoy cataloguing?"


"We need to see more of the systems people so that we can collaborate and explain our side of the story."

"Cataloguers make good systems people - if you can write an all-singing, all-dancing MARC record, you can probably code just fine!"

"So how do we get a systems job?"

"Go to a mashlib event!"

"Learn and practice in your own time. Try building something! Fashionable programming languages are: python, php, perl [and one that I wrote so poorly I really can't decipher it... sorry.] [ETA Adrienne has identified it as ruby]"


Conclusions: we need to re-evaluate what we're doing and how in collaboration with the people who have the data about what people are doing and trying to do with our catalogues, and can build the systems we want.

Resources (plugged at the end of the session):


  1. Thanks for writing this up, very useful as I just really wanted to be there. Lots to think about, even though I don't quite know what was said - am pondering some stuff about standards so would love to hear more from other participants as get the impression there was a bit of a "what's the point" type of feeling about cataloguing standards....

    Thanks also to whoever gave HVCats a plug at the end :)

  2. I was scribbling as fast as I could to keep track of what was said - should have explained above - the notes here are pretty much what was said in order. Obvs not all the details are there, though... I'd be very happy to hear other participants' write-ups too, though (hint hint!).

    I think the general mood was that standards for the sake of standards are silly and pointless. And that writing good standards is hard. Also - standards will change with time, so it's maybe counter productive to obsess over every aspect of records you create today, as they won't stay in the format you create them indefinitely.