I think that is what was being asked for, but I suppose I'm trying to  
get to what the underlying driver is - and assuming that users wikld  
like it wondering if there are reasons beyond familiarity.

If the classification does capture some unique aspect of 'likeness'  
that's fine I suppose - although I wonder why? But isn't  
classification just another possible facet to browse?

I suppose I'm just thinking out loud, but I think we need to  
understand the way people use browse to navigate resources if we are  
to successfully bring the concept of collection browsing to our  
navigation tools. David suggests that we should think of a shelf  
browse as a type of 'show me more like this' which is definitely one  
reason to browse - but is it the only reason?

Owen



On 30 Sep 2008, at 21:52, "Tim Shearer" <[log in to unmask]> wrote:

> Owen,
>
> Unless I'm misunderstanding, what's being asked for is a  
> visualization tool for the *classification*.  Faceted browsing by  
> subject is dandy, but is not at all the same thing (though arguments  
> can be made that the lines are blurring).  Books that sit next to  
> each other in a classification (DC or LC, or whatever) may not share  
> a majority of subject terms.  That collocation via classification is  
> yet another (and occasionally more useful) way of saying that this  
> item is like that item.  One that is not necessarily trapped in any  
> other way than call number.
>
> -t
>
> On Tue, 30 Sep 2008, Stephens, Owen wrote:
>
>> I'd second Steve's comments - replicating an inherently limited  
>> physical
>> browse system seems an odd thing to do in the virtual world. I would
>> have thought that the 'faceted browse' function we are now seeing
>> appearing in library systems (of course, the Endeca implementation  
>> is a
>> leader here) is potentially the virtual equivalent of 'browsing the
>> shelves', but hopefully without the limitations that the physical
>> environment brings?
>>
>> Is it the UI rather than the functionality that is lacking here?  
>> Perhaps
>> we need to look more carefully at the 'browsing' experience. Thinking
>> about examples outside the library world, I personally like the
>> 'coverflow' browse in iTunes, but I'm able to sort tracks by several
>> criteria and still see a coverflow view. I have to admit that in  
>> general
>> I prefer the 'album' order when using coverflow, because otherwise it
>> doesn't make sense (to me that is). It would be interesting to look  
>> at
>> what an 'artistflow' might look like, or a 'genreflow'.
>>
>> However, as far as I know I can't actually replicate the experience  
>> that
>> I would have with my (now in boxes somewhere) physical CD  
>> collection -
>> why was divided by genre, then sorted by artist surname (ok, I  
>> admit it,
>> I'm a librarian through and through)
>>
>> Perhaps a better understanding of the 'browse' experience is needed?
>>
>> Some questions - when we browse:
>>
>> When and why do people browse rather than search?
>> How do people make decisions about useful items as they browse?
>> Browsing stacks suggests that items have been 'ordered' - is there
>> something about this that appeals? Does it convey 'authority' in some
>> way that the 'any order you want' doesn't?
>>
>> Owen
>>
>> Owen Stephens
>> Assistant Director: eStrategy and Information Resources
>> Central Library
>> Imperial College London
>> South Kensington Campus
>> London
>> SW7 2AZ
>>
>> t: +44 (0)20 7594 8829
>> e: [log in to unmask]
>>> -----Original Message-----
>>> From: Code for Libraries [mailto:[log in to unmask]] On Behalf
>> Of
>>> Steve Meyer
>>> Sent: 29 September 2008 21:45
>>> To: [log in to unmask]
>>> Subject: Re: [CODE4LIB] creating call number browse
>>>
>>> one counter argument that i would make to this is that we  
>>> consistently
>>> hear from faculty that they absolutely adore browsing the
>> stacks--there
>>> is something that they have learned to love about the experience
>>> regardless of whether they understand that it is made possible by  
>>> the
>>> work of catalogers assigning call numbers and then using them for
>>> ordering the stacks.
>>>
>>> at uw-madison we have a faculty lecture series where we invite
>>> professors to talk about their use of library materials and their
>>> research and one historian said outright, the one thing that is
>> missing
>>> in the online environment is the experience of browsing the  
>>> stacks. he
>>> seemed to understand that with all the mass digitization efforts, we
>>> could be on the edge of accomplishing it.
>>>
>>> that said, i agree that we should do what you say also, just that we
>>> should not throw the baby out w/ the bath water. if faculty somehow
>>> understand that browsing the stacks is a good experience then we can
>>> use
>>> it as a metaphor in the online environment. in an unofficial  
>>> project i
>>> have experimented w/ primitive interface tests using both subject
>>> heading 'more like this' and a link to a stack browse based on a  
>>> call
>>> number sort:
>>>
>>> http://j2ee-dev.library.wisc.edu/sanecat/item.html?resourceId=951506
>>>
>>> (please, ignore the sloppy import problems, i just didn't care that
>>> much
>>> for the interface test)
>>>
>>> as for the original question, this has about a million records and
>>> 900,000 w/ item numbers and a simple btree index in the database  
>>> sorts
>>> at an acceptable speed for a development test.
>>>
>>> -sm
>>>
>>> Walker, David wrote:
>>>>> a decent UI is probably going to be a bigger job
>>>>
>>>> I've always felt that the call number browse was a really useful
>>> option, but the most disastrously implemented feature in most ILS
>>> catalog interfaces.
>>>>
>>>> I think the problem is that we're focusing on the task -- browsing
>>> the shelf -- as opposed to the *goal*, which is, I think, simply to
>>> show users books that are related to the one they are looking at.
>>>>
>>>> If you treat it like that (here are books that are related to this
>>> book) and dispense with the notion of call numbers and shelves in  
>>> the
>>> interface (even if what you're doing behind the scenes is in fact a
>>> call number browse) then I think you can arrive at a much simpler  
>>> and
>>> straight-forward UI for users.  I would treat it little different  
>>> than
>>> Amazon's recommendations feature, for example.
>>>>
>>>> --Dave
>>>>
>>>> ==================
>>>> David Walker
>>>> Library Web Services Manager
>>>> California State University
>>>> http://xerxes.calstate.edu
>>>> ________________________________________
>>>> From: Code for Libraries [[log in to unmask]] On Behalf Of
>>> Stephens, Owen [[log in to unmask]]
>>>> Sent: Wednesday, September 17, 2008 9:17 AM
>>>> To: [log in to unmask]
>>>> Subject: Re: [CODE4LIB] creating call number browse
>>>>
>>>> I'm not sure, but my guess would be that the example you give isn't
>>>> really a 'browse index' function, but rather creates a search  
>>>> result
>>> set
>>>> and presents it in a specific way (i.e. via cover images) sorted by
>>> call
>>>> number (by the look of it, it has an ID of the bib record as input,
>>> and
>>>> it displays this book and 10 before it, and 10 after it, in call
>>> number
>>>> order.
>>>>
>>>> Whether this is how bibliocommons achieves it or not is perhaps
>>> besides
>>>> the point - this is how I think I would approach it. I'm winging it
>>>> here, but if I was doing some quick and very dirty here:
>>>>
>>>> A simple db table with fields:
>>>>
>>>> Database ID (numeric counter auto-increment)
>>>> Bib record ID
>>>> URIs to book covers (or more likely the relevant information to
>>> create
>>>> the URIs such as ISBN)
>>>> Call number
>>>>
>>>> To start, get a report from your ILS with this info in it, sorted  
>>>> by
>>>> Call Number. To populate the table, import your data (sorted in  
>>>> Call
>>>> Number order). The Database ID will be created on import,
>>> automatically
>>>> in call number order (there are other, almost certainly better,  
>>>> ways
>>> of
>>>> handling this, but this is simple I think)
>>>>
>>>> To create your shelf browse given a Bib ID select that record and
>> get
>>>> the database ID. Then requery selecting all records which have
>>> database
>>>> IDs +-10 of the one you have just retrieved.
>>>>
>>>> Output results in appropriate format (e.g. html) using book cover
>>> URIs
>>>> to display the images.
>>>>
>>>> Obviously with this approach, you'd need to recreate your data  
>>>> table
>>>> regularly to keep it up to date (resetting your Database ID if you
>>>> want).
>>>>
>>>> Well - just how I'd do it if I wanted something up and running
>>> quickly.
>>>> As Andy notes, a decent UI is probably going to be a bigger job ;)
>>>>
>>>> Owen
>>>>
>>>> Owen Stephens
>>>> Assistant Director: eStrategy and Information Resources
>>>> Central Library
>>>> Imperial College London
>>>> South Kensington Campus
>>>> London
>>>> SW7 2AZ
>>>>
>>>> t: +44 (0)20 7594 8829
>>>> e: [log in to unmask]
>>>>
>>>>> -----Original Message-----
>>>>> From: Code for Libraries [mailto:[log in to unmask]] On
>> Behalf
>>>> Of
>>>>> Emily Lynema
>>>>> Sent: 17 September 2008 16:46
>>>>> To: [log in to unmask]
>>>>> Subject: [CODE4LIB] creating call number browse
>>>>>
>>>>> Hey all,
>>>>>
>>>>> I would love to tackle the issue of creating a really cool call
>>> number
>>>>> browse tool that utilizes book covers, etc. However, I'd like to  
>>>>> do
>>>>> this
>>>>> outside of my ILS/OPAC. What I don't know is whether there are any
>>>>> indexing / SQL / query techniques that could be used to browse
>>> forward
>>>>> and backword in an index like this.
>>>>>
>>>>> Has anyone else worked on developing a tool like this outside of
>> the
>>>>> OPAC? I guess I would be perfectly happy even if it was  
>>>>> something I
>>>>> could build directly on top of the ILS database and its indexes  
>>>>> (we
>>>> use
>>>>> SirsiDynix Unicorn).
>>>>>
>>>>> I wanted to throw a feeler out there before trying to dream up  
>>>>> some
>>>>> wild
>>>>> scheme on my own.
>>>>>
>>>>> -emily
>>>>>
>>>>> P.S. The version of BiblioCommons released at Oakville Public
>>> Library
>>>>> has a sweet call number browse function accessible from the full
>>>> record
>>>>> page. I would love to know know how that was accomplished.
>>>>>
>>>>> http://opl.bibliocommons.com/item/show/1413841_mars
>>>>>
>>>>> --
>>>>> Emily Lynema
>>>>> Systems Librarian for Digital Projects
>>>>> Information Technology, NCSU Libraries
>>>>> 919-513-8031
>>>>> [log in to unmask]
>>>
>>> --
>>> Stephen Meyer
>>> Library Application Developer
>>> UW-Madison Libraries
>>> 312F Memorial Library
>>> 728 State St.
>>> Madison, WI 53706
>>>
>>> [log in to unmask]
>>> 608-265-2844 (ph)
>>>
>>>
>>> "Just don't let the human factor fail to be a factor at all."
>>> - Andrew Bird, "Tables and Chairs"
>>