[Eril-l] KB discrepancies

Thu Oct 8 13:47:21 PDT 2015

Hi Steve, 

A group of us in Canada who are responsible for consortia package lists for inclusion in link resolver services have noticed discrepancies in the lists among all of the KB/link resolver services even though each service provider receives the exact same list from the consortia. 

Reasons for possible discrepancies that we've deduced include: 

1. Update scheduling - how often does the KB/link resolver service update their lists? and how soon after getting the update. Also how do they update the target? With an overlay for new data or complete overhaul of the data? 

2. KB/link resolver service may add value by making edits to the original list - to include title changes not accounted for in the original list. Other services may simply add the data as received without any further editing. 
I think in your situation regarding PsycBooks from EBSCO - there is possibly a problem with the SFX target. It probably needs to be updated to account for the new addition of APA handbooks that were added earlier this year. 
My institution loads MARC records for PsycBooks (because there are no serials in that collection, so we don't use the KB for that) and we have over 4000 MARC records in our catalogue for PsycBooks. So 2000+ seems like a really low number to me for PsycBooks. 

A check on another KB service has 3600+ titles for the same database - also all books (no serials) - so I'd suggest contacting SFX to update their target. 

Sandra. 

----- Original Message -----

From: "Steve Oberg" <steve.oberg at wheaton.edu> 
To: eril-l at lists.eril-l.org 
Sent: Thursday, 8 October, 2015 13:08:03 
Subject: Re: [Eril-l] KB discrepancies 

Jill’s idea of KBART comparisons is spot on, and it’s something I’ve worked quite a bit on. Yes, this can be done for both SFX and OCLC. As helpful as KBART is, it’s not the complete answer, though. As an example of what I mean, you might think, ok, both accept KBART, so why not update one KB with the KBART output of the other? This simply doesn’t work, for a variety of reasons. 

Some of those reasons have already been articulated in other replies. KB vendors simply do not count things in the same way across the board. And the ways in which those KBs are wired to do what they do makes a big difference in terms of how they are created and maintained. For example, in the OCLC realm, the OCLC number is extremely important. Not true in the SFX realm (although technically one can assign an OCLC number to a title). 

As noted before, I’ve done a lot of work to compare/contrast — even just at the collection level — these two different KBs. They are actually quite different, more so than you might suspect. I completely agree in theory to the principle of simplicity — just choose one and stick with it. But I’d say that this assumes we are comparing the same things in terms of content coverage, and basically, we are not. At least, I can say that for OCLC vs. SFX. 

I have also found vast disparity between how quickly one KB gets updated vs. another and cannot give a definitive answer to which vendor does a better job of managing their data. I will see that the method used by SFX is much more comfortable for me because it gives me a lot more control as needed, at a local level, than OCLC WMS KB can possibly do. 

But let me clarify that I’m not trying to point fingers or focus on one vendor vs. the other, but rather, trying to understand why KB discrepancies occur in general. Your responses thus far have been very helpful. I’m also interested in how others successfully manage to juggle the maintenance of more than one KB at a time, however success is defined. 

Steve 

Steve Oberg 
Assistant Professor of Library Science 
Electronic Resources and Serials 
Wheaton College (IL) 
+1 (630) 752-5852 

From: Eril-l on behalf of "Katherine T. Silton" 
Date: Thursday, October 8, 2015 at 2:50 PM 
To: "Donley, Leah", " eril-l at lists.eril-l.org " 
Subject: Re: [Eril-l] KB discrepancies 

I have previously had to maintain multiple KBs (EBSCO and Ser Sol) and I think that Leah’s explanation is a good one. Long-term, it’s definitely a lot of work to maintain two and I wouldn’t recommend it. 

From: Eril-l [ mailto:eril-l-bounces at lists.eril-l.org ] On Behalf Of Donley, Leah 
Sent: Thursday, October 8, 2015 2:53 PM 
To: eril-l at lists.eril-l.org 
Subject: Re: [Eril-l] KB discrepancies 

In addition to everything mentioned so far, I’ve worked with publishers and KB providers in the past to correct title lists for collections. I suspect that when this happens, the KB feeds aren’t always updated (or maybe not updated for all KB providers, not exactly sure how that works), because we’ve switched providers a few times and often a collection I’ve corrected with one provider will not be correct in the new provider’s KB. 

-Leah 

Leah Donley 

Information Specialist 

Research Library 

Brookhaven National Laboratory 

Upton, NY 

donley at bnl.gov 

From: Eril-l [ mailto:eril-l-bounces at lists.eril-l.org ] On Behalf Of Jill Emery 
Sent: Thursday, October 08, 2015 1:59 PM 
Cc: eril-l at lists.eril-l.org 
Subject: Re: [Eril-l] KB discrepancies 

is it possible to get the KBART feeds for both knowledgebases? That may explain the difference. 

Both companies endorse KBART so that may help get to the bottom of this situation. 

Here is the registry contact information for both providers: https://sites.google.com/site/kbartregistry/ 

On Thu, Oct 8, 2015 at 10:53 AM, Daniel Peterson < dcpcreations at gmail.com > wrote: 

I'm not sure how SFX and OCLC KBs work, but issues I can think of to take into account are: 

1. Do the title lists include only full text content, or full text and A&I titles 

2. If full text only, how comprehensive does that full text need to be ("selected full text" titles cause confusion with link resolvers) 

3. How are journal title changes tracked? Single entry for all title versions, or separate for each title/ISSN variant? 

4. Conference proceedings and monographic serials can be a mess 

Daniel 

On Thu, Oct 8, 2015 at 10:40 AM, Steve Oberg < steve.oberg at wheaton.edu > wrote: 

<blockquote>

I'm managing SFX as well as the OCLC WMS KB. I don't intend to open the can of worms of: why have more than one? So...If you are in a similar situation to mine, and have insights about the following common scenario, I'd love to hear your perspective. 

KB #1 has a particular collection. KB #2 has the same collection. Yet the number of individual titles within each varies wildly. Here's a specific example: SFX has a target for EBSCOHOST_PSYCBOOKS that has 2,183 portfolios. The same thing in OCLC WMS KB (called a collection, with title, EBSCOhost PsycBOOKS) has 4,021 titles in it. 

If the source for each KB is a vendor feed of some sort, isn't it logical to expect the numbers to be somewhat close? But they frequently aren't. I'm not sure there is a ready solution to the problem but I'm wanting to at least have a better grasp of the possible reasons for these discrepancies. 

Thanks, 

Steve 

_______________________________________________ 
Eril-l mailing list 
Eril-l at lists.eril-l.org 
http://lists.eril-l.org/listinfo.cgi/eril-l-eril-l.org 

-- 

Daniel Peterson 
dcpcreations at gmail.com 

_______________________________________________ 
Eril-l mailing list 
Eril-l at lists.eril-l.org 
http://lists.eril-l.org/listinfo.cgi/eril-l-eril-l.org 

NOTICE: This e-mail correspondence is subject to Public Records Law and may be disclosed to third parties. –– 
NOTICE: This e-mail correspondence is subject to Public Records Law and may be disclosed to third parties. –– 

</blockquote>

_______________________________________________ 
Eril-l mailing list 
Eril-l at lists.eril-l.org 
http://lists.eril-l.org/listinfo.cgi/eril-l-eril-l.org 

-- 
_________________________________________________________________________________ 
Sandra Wong, Electronic Resources Librarian 
Collections Management Office 
Simon Fraser University Library 
8888 University Drive 
Burnaby, BC Canada V5A 1S6 

Email: swongj at sfu.ca or lib-licensing at sfu.ca 
Phone: 778.782.4930 
Fax: 778.782.3023 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.eril-l.org/pipermail/eril-l-eril-l.org/attachments/20151008/84cbe84e/attachment.html>