catalog system

Post by **hugh** » November 7th, 2005, 2:05 pm

The catalog project is in progress.
More information on current developments can be found here:
http://librivox.org/wiki/moin.cgi/CatalogProject

---------

ok here's some stuff to chew on, from Jessamyn west, of http://librarian.net, in response to an email I sent her:

1. since you're mostly, for now, dealing with monographs, having standard
cataloging in addition to fruity 2.0 cataloging is a must. So, make sure
you have

- title
- author
- authority control [so if you are in the happy state whereby someone
records a different edition of a work in your catalog, they can be
co-located in the catalog with the first one. in the library see:
Bible/Shakespeare]
- isbn [see above, these can work together in the new 13 char ISBN which
is coming out I think this year]
- publisher [good code can make the publisher appear once the ISBN is
given, HOT!]
- reader [seems odd but some people may really like the voice of one
reader and hate others]

2. I like facets. Especially with a known scope, like "out of copyright
books" I think there are some inputs that should be controlled only
because I think they add to findability. So think category instead of
tags, or in addition. Ideas in this area are

- type: fiction/non-fiction/bio/travel
- subtype: genre (sci fi, history, fantasy, )
- kid/teen/adult
- language?
- length: short medium long?
- voice (male/female) might matter to some

3. Tagging will deal with some of the above but not all of it. If a tagger
misses calling something a biography, that's a big miss. Tagging is
really good for things that are either more mercurial (tags for colors,
appearance of town/city, etc) or subjective (creepy, romantic) or only
matter to some people when they choose books (Vermont, disabled,
waterfall). So you can have inputs for the above [I don't know what your
input area looks like, I'd happily check it out] and then a free form area
for tags with some serious flickr-like limits on how they work. The
biggest thing with tags is making them useful, clickable, cloudable, etc.
I have tags on librarian.net and they're just "eh" because the tool I used
wasn't very robust. I'm playing with a much neater tool which, in
hindsight I maybe should have started with.

So, tags need to, at minimum

1. combine using boolean or preferably in-URL ways
2. be hyperlinks
3. complement other classification
4. be easy to input, edit and remove if necessary (and make sure editing
them edits ALL the tags in that namespace) and I strongly suggest tag
adding by users

maximally they could

1. be used for findability in sites like technorati &c [do you have an RSS
feed of new titles? get one!]
2. be a way that users/readers could see each others interests by the tags
of what they've read, contributed to, downloaded, I dunnow
3. be used visually to show areas of specialty or areas that need work,
when viewed cloud-like overall

You can view my micro-presenation here

http://librarian.net/talks/flickr

but there aren't a lot of words in it. I'm not much of an expert on tags
per se, I've just thought an awful lot about them because I like
organizing things. When you have an idea of what you might want your tag
universe to look like, drop me another line.

Post by **hugh** » November 7th, 2005, 2:07 pm

oh, one thing would be nice is a database of users/volunteers/listeners, something like flickr. So I can check out kristen's page, eg, and see:

a) what she has downloaded (? maybe not so much?)
b) what she has recorded
c) what she is planning to record...

kayray · Post by **kayray** » November 7th, 2005, 3:05 pm

hugh wrote: b) what she has recorded
c) what she is planning to record...

yeah, now _that_ would be cool! I'm keeping track of my assignments in a notebook on a piece of paper, for heaven's sake... how archaic.

kara

thistlechick · Post by **thistlechick** » November 30th, 2005, 1:48 pm

ahhh... i just found this discussion after posting my ideas in the Comments/Suggestions forum ( http://librivox.org/forum/viewtopic.php?t=294 )... I'm glad I'm here so early in the discussion

I am a cataloging librarian and would love to help out with establishing a consistant metadata procedure.

Kristen · Post by **Kristen** » November 30th, 2005, 4:48 pm

http://www.mediatinker.com/librivox/catalogBack.html

The above is a draft (not very pretty yet) of the administrators data page for a book in our new catalog. Summaries have been suggested, so we'll add those to the book section.

Further suggestions welcome...

Stephan · Post by **Stephan** » December 18th, 2005, 10:57 am

Hi Hugh, i?d very much like to participate when it comes to the visual end of the catalog/database, its usability, its design.
Can?t code, but studied design....
I once worked out sort of an online calender togather with a web-database programmer - a tool to help a group of people to coordinate common dates. Making it easy to use and understand was a great joy.
Where are you in the process? Any Subforum i can read into?

(I do now understand difference between tagging and catergorizing. You got a discussion about that?)

Arjun · Post by **Arjun** » February 14th, 2006, 11:44 am

Hello,

I am (still) new to the list, and have just come up on this thread about cataloging.

What software do you use for cataloging? Are you planning to set up a searchable catalog (there are open-source solutions)? Whom can I contact about cataloging?

Thanks!

Post by **hugh** » February 14th, 2006, 12:17 pm

we have a new (today!) catalog project set-up on sourceforge:

https://sourceforge.net/projects/librivox

you can take a peek. Otherwise, kara, thistlechick (betsie), gesine, librarylady (annie), and kri take care of cataloging; tis (chris) is our main back-end tech guy. cataloging now is done more or less manually at the moment.

Hugh.

Peter Why · Post by **Peter Why** » February 14th, 2006, 12:48 pm

I've not looked at enough of these posts to know whether this has been discussed, but books, stories, poems or whatever, often can't neatly be classified by subject/style, so ideally we should be able to assign a book to more than one ... humour, war, biography could all apply to the same book.

You can't even be sure about fiction/non-fiction sometimes! ... myths? fairy tales? ghosts?

It would be like assigning more than one Dewey classification number to each work.

Post by **hugh** » February 14th, 2006, 1:36 pm

peter I think we are supposed to be able to "tag" the books - meaning they might have a "top-level" classification (fiction), but also some less-official tags (humour, clown, pie etc) to allow what you suggest. But for this i defer to all our librarians.

vee · Post by **vee** » February 14th, 2006, 1:38 pm

Just a note on the sourceforge site. I think they spelled the URL wrong "http://Libravox.org" in the description

LibraryLady · Post by **LibraryLady** » February 14th, 2006, 2:36 pm

Oooo, cool, the database is finally starting to become a reality!!

As for how we will classify things, I'm thinking we'll put each work into one of the following collections:
Adult Fiction
Adult Non-Fiction
Children's Literature
Other Language
Short Works
Poetry
Basically, the same categories we (MC/catalogers) currently use to identify and tag the projects.

But ideally, within the item record for a project, we will add subject headings that will also be searchable. I think Betsie has already begun keeping track of LCSH stuff for some of the finished projects. You can have as many subject headings as you want (at least in library cataloging) so we won't have to limit ourselves to choosing a primary subject or genre, just a primary collection.

EDIT: Sorry for the library cataloging lingo, folks. LCSH=Library of Congress Subject Headings

thistlechick · Post by **thistlechick** » February 14th, 2006, 3:04 pm

I would like to see a combined system where we have a set list of genres from which to choose and are assigned by the MC's when cataloging the item... but also includes the option for other free-text keywords ("tags" in the current lingo) that can also be added...

So, I see two issues here in terms of classification:

Subject - what the item is about
Genre - what type of item it is

For example, the item might be non-fiction (a genre) that is about philosophy (a subject)

Genre's can also be more specific such as: Political fiction

LibraryLady · Post by **LibraryLady** » February 14th, 2006, 3:10 pm

Yes, when I said collection I meant what you are saying with genre. Genre is probably the better term, more user friendly, less library-ish.

So are you thinking we'll have a static list of subjects in an item's record and then make the entire record searchable by free-text keyword? That makes sense, I guess. It would be rather complicated to create subject headings that link to lists of other works in that subject heading. Maybe what we can try to do for starters is use uniform subject headings without links and then eventually add the links.

Arjun · Post by **Arjun** » February 14th, 2006, 8:35 pm

Hi Everyone:

I think that it is never too late to add subject headings that describe the printed works in standard cataloging systems. I have looked at the LibiVox catalog and wondered what standard was used in setting this one up? I have looked into the sourceforge site as Hugh had suggested but did not find anything there when I clicked on the download button.

I was perhaps thinking of using a combination of MARC and Dublin Core, but I hope I am not being presumptuous here. As open-source catalogs become more available to search on the internet, it could perhaps be better to think long-term. A well-planned catalog would allow you to retrieve many related items in one search.

LibraryLady, you are right on the money about a uniform subject headings approach but local labels can also be used.

As for Peter?s idea: With XML metatags, you can assign multiple values to a title (a title could be described as novel, romance, etc.) and URLs can serve as links to Wikipedia, Gutenberg catalog, and other catalog entries.

Hugh: you can also use top-level as well as sub-categories not to mention foreign equivalents, now that we have multilingual collections.

Kristen: the catalog on http://www.mediatinker.com/librivox/catalogBack.html has important information about the media itself, so that can also be combined into a larger catalog.

Actually, if you look at Hugh?s posting from November, that pretty much outlines what could be done while the collection is still relatively little. It may be much more complicated to convert the records later.

Here is a question though: are the recordings treated as monographs or performances? The experience of reading text with an internal (i.e., your own) voice is very different from the experience of getting the same through an external (reader's) voice. This is where the inter/subjective aspects of interpretation (hearing vs. reading) would differ dramatically.

Well, I hope I am not getting on anyone's nerves here. Cheers

-- Arjun