Email list hosting service & mailing list manager


Link resolver comparison results Diane Westerfield 30 Nov 2012 00:25 UTC

Hi folks,

I received 11 responses and have condensed/edited/anonymized them. At the end you'll find my own, which is lengthy since I had the advantage of considering the many pros and cons after reading other thoughts. Thanks to the folks who responded, and I hope this helps some of you. If any vendors read this, please take it as constructive criticism.

I have also looked at the top 40 liberal arts colleges as ranked by U.S. News & World Report, since my school is in this group, and tried to determine what link resolver product was being used by each college's library. Here's how the numbers shake out in this group: Serials Solutions: 25, SFX: 4, Serials Solutions + SFX: 1, EBSCO: 3, WebBridge: 1, Gold Rush: 1, Undetermined: 5.

Meanwhile, checking some large & notable schools: Harvard, Yale, University of Chicago, Northwestern, Stanford, University of Illinois Urbana-Champaign all have SFX.

The very unscientific conclusion I am coming to is that SFX is better suited for large libraries with well-staffed Systems departments. Serials Solutions is a popular choice, but cost is an issue for some schools. Gold Rush might be a good choice for smaller schools who need a reasonably-priced, less-sophisticated product. EBSCO appears to be a work in progress, but the company has a lot of resources at its disposal and improvements are being worked on. If you already have a number of products from one vendor, adding their link resolver might make your life easier because the systems will tend to agree with each other better. One respondent who uses WebBridge along with other Innovative products reports that it works great.

I wanted to point out that no link resolver can ever be 100% perfect because journal holdings are a moving target - especially with aggregator databases where titles are constantly being added and removed. Moreover, vendor data is often inaccurate, incomplete, or out-of-date. If a vendor supplies bad title lists, *all* link resolvers will have that same bad data.

I will be conducting further research -- perhaps hitting people up with a formal survey -- so stay tuned!

A note about distinctions between Link Resolvers as OpenURL resolvers and as KBs (Knowledge Bases) that can display public A-Z lists. To me, these functions go hand-in-hand, so I'm not separating these out conceptually.

=====================================================================================================================

EBSCO A-Z/LinkSource (4 respondents)

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

Respondent currently uses EBSCO A-Z; previously used Serials Solutions:

Respondent misses Serials Solutions because EBSCO A-Z has the following problems: EBSCO sometimes overwrites custom collections. Public searches often miss titles or incorrectly display title results (EBSCO doesn't follow OCLC rules on variant titles). Local custom holdings can't be edited easily, requiring labor-intensive file downloads, Excel, and uploads for the most simple edits. Can't download specific custom collections by themselves to make the editing process easier. The files are complex and require a learning curve when editing in Excel, making it difficult or impossible to delegate this work.

Respondent has contacted EBSCO with these problems and has been assured they're being worked on.

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

Respondent currently uses EBSCO LinkSource; previously used Gold Rush:

LinkSource is still pretty new and respondent can't offer the most detailed review. Institution used Gold Rush and liked it, but LinkSource came along with a package deal. Gold Rush staff always responsive and good with communication. EBSCO customer service has been prompt but respondent feels there is a "certain level of disconnect" because it's a larger organization. LinkSource appears to have a learning curve compared to Gold Rush.

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

Respondent currently uses EBSCO A-Z and LinkSource, previously used SFX:

SFX has a steep learning curve; it's really designed for Systems Librarians. Knowledge of UNIX required to customize anything. User interface is clean and patrons liked it. Excellent tech support, good discussion board (although sometimes too techy). With SFX, you could do a lot of tweaking of titles and running reports.

EBSCO Pros: doesn't require working with UNIX, administrator interface is straight-forward. Customization of colors, fonts, etc. is easy. Offers LCSH (subject headings). Overlap analysis. Tech support has been getting better over time and is available by phone.

Cons: Respondent didn't know how to apply proxy settings at first, causing access issues for their campus, but EBSCO tech support helped them solve the problem. When downloading data, you receive one large file regardless of how much or how little you need. Some aspects of public display customization are frustrating, such as not being able to hide the "books" button if you don't include those in your LR. One report didn't work for some months, but it has since been resolved.

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

Respondent currently uses EBSCO LinkSource:

Respondent's institution recently got LinkSource and there are problems, including: links displaying for embargoed journals, open access links not working, confusion when there is availability in more than one database, and issues with local holdings.

Respondent is working with EBSCO to fix these problems.

=====================================================================================================================

Serials Solutions 360 Link (4 respondents)

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

Respondent who currently uses Serials Solutions and previously used a link resolver developed by University of NC Greensboro, now owned by WT Cox:

Respondent prefers Serials Solutions over previous link resolver (which had issues as a product and with customer service/communication).

360 Link pros: much easier and less time consuming to work with. 360 Link allows for easy inclusion of print holdings; easy administration of staff-end permissions. Easy changes of single or multiple updates to coverage, links, and display of titles. Serials Solutions sends out weekly updates of changes; you can request to get updates of particular resources. Customer service is usually quick and responsive. Free & government resources available and easily turned on. Overlap analysis feature useful for considering new purchases. Integrates seamlessly between library website and vendor databases. Offers the one-click option. Has option to change the order of or eliminate duplicates from display. Works with Google Scholar.

Cons: respondent wishes they could afford more of the 360 family of products. Updates take a day to display publicly.

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

Respondent who currently uses Serials Solutions and previously used SFX:

Respondent prefers Serials Solutions because the staff interface is more intuitive and easier for non-technical staff to use. Customer service is easier to reach, more responsive, and available by phone. Respondent's library uses other Serials Solutions products and this makes life easier because the systems are more likely to agree about holdings data. Respondent is satisfied with accuracy of holdings data and responsiveness to correcting mistakes.

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

Respondent who currently uses Serials Solutions 360 Link and previously used SFX and OCLC link resolver via Sirsi:

Respondent reports that OCLC link resolver was not a good product, lacking the ability to make changes and KB was not updated frequently enough. This product may no longer exist, or it may have evolved into something else.

Serials Solutions pros: Admin module has been significantly improved to allow for easy updates of wording, display, etc. that library staff can do themselves (no vendor action needed). Serials Solutions keeps up the KB well.  Customer service is good and the openURL linker has no problem. Respondent reports issues to Serials Solutions and they are dealt with quickly, but respondent feels that many errors come from poor vendor data. 360 Link is the method of access to many of their resources.

Cons: the price is high for the type and size of the institution, but respondent wants to keep it since it is the main method of reaching some of their e-resources. Occasionally there are problems related to titles and ISSNs but Serials Solutions is willing to add alternate titles to make searching smoother.

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

Respondent who currently uses Serials Solutions, previously used SFX:

Respondent recalls that SFX is more flexible than Serials Solutions in the way you could set up menus. Respondent has heard that SFX's KB is not as good as Serials Solutions, but that might be a local issue.

=====================================================================================================================

SFX (ExLibris) (1 respondent)

Respondent who currently uses SFX and previously used Serials Solutions:

SFX difficult to manage because of reliance on Perl in regards to URL and date ranges (not a problem with Serials Solutions). SFX KB not as comprehensive as Serials Solutions, especially regarding title changes. SFX related objects function not useful. Respondent has had to add local targets, objects and portfolios which SFX won't load into the global KB, even though these titles are not confined to this particular institution.

Workflow: e-journal and e-book metadata harvested every night, however the workflow is complex and prioritized based on usage, paid versus free, whether in a database. At a previous institution, this respondent reports they used Serials Solutions and csv files loaded into the ILS, the only added metadata were subject headings based on LCSH and the database name.

======================================================================================================================

WebBridge for Millennium ILS  (Innovative) (1 respondent)

Respondent has access to ESBCO LinkSource but they find it difficult to work with and don't like its handling of title changes (all title variants lumped into one title).

Pros: Respondent declares they are "wholly satisfied with WebBridge LR."  Works well with the Millennium ERM module. Allows easy editing of links and coverage, and of URL syntax. Print holdings can be added easily.  [Diane adds: Sounds like your KB has to be built from the ground up and maintained locally(?)]

Cons: as this is a less popular link resolver, database vendors may struggle to work with it.  Official documentation is not that great, but the ample user-supplied documentation is helpful.

Workflow: Update coverage for subscribed and print titles when changes occur, monthly re-loads of the free and full-text database coverage, also run (almost) daily updates of changes to our A to Z list and adjust the coverage accordingly.

=====================================================================================================================

OVID Link Resolver (Wolters Kluwer) (1 response)

Pros: Great when it works. Possibility of including e-books, but unsure if it's just Lippincott books.

Cons: A lot of the links break. Date coverage can be wrong and customers can't fix coverage on their end. Tech support not very responsive. Can't display print holdings.

This respondent looked at Serials Solutions and preferred it, but patrons didn't. May look at TDNet in the future; TDNet might help this library with their authentication issues.

====================================================================================================================

Gold Rush (Colorado Alliance of Research Libraries) (1 respondent - Diane Westerfield)

Respondent who currently uses Gold Rush and previously used SFX:

Pros: Gold Rush is free to Colorado Alliance members and reputedly "reasonably priced" to non-members. Staff-side interface is easy to use and understand; unlike SFX, no programming or UNIX navigation required. Gold Rush staff responds quickly to error reports. Updates are easy to make and reflected immediately on the public side (had to wait a day with SFX). Overlap analysis is available. Print holdings may be entered as local holdings and they display well. URLs are stored and reported "as is"; SFX required a slow-going Perl script to derive URLs. Gold Rush staff are happy to take your title lists. I was told in the past by SFX staff that they would accept our considerable menagerie of rare/local periodical titles, but they never loaded the titles into the global KB; I ended up making local portfolios. Gold Rush adds LCSH subject headings to journal titles and you can add your own, potentially making subject searching a lot more useful than link resolvers with very general subject headings. Offers One Click option. Often has free webinars. Works with Google Scholar. Has an XML gateway which the more Systems-inclined librarian can take advantage of.

Cons: Gold Rush relies heavily on ISSNs for its operations. It does not include the thousands of no-ISSN titles from LexisNexis Academic Universe in its KB, and it will also miss the dozens of no-ISSN titles in EBSCOHOST databases. There seems to be a lack of separation between local and global KB, such that the occasional strange alternate title entered by another school can contaminate your search results. Title lists don't always load properly, requiring manual edits. Sometimes edits you make seem to disappear later on. The basic look of Gold Rush is clunky (I believe they're working on this). The public interface is customizable but if your CSS isn't that strong, you'll have a hard time making it look slick. As a less popular link resolver, some vendors may struggle with implementing it; we have only experienced this problem with Wiley. Sometimes the Gold Rush public interface throws large Cold Fusion error messages, which look unprofessional. The staff side of Gold Rush has no way to indicate when a database was last updated, so if you are trying to maintain ejournal holdings in your catalog, you may be wondering when you should run another load of a particular e-resource. Gold Rush doesn't handle title changes well, nor does it impose much quality control on titles (you can see main titles for records where diacritics have been transformed into blank spaces). We seem to have had book linking with SFX, which no longer works with Gold Rush (this may be due to my lack of implementation or a misunderstanding). I can't add a special Sherpa/Romeo target to Gold Rush to the same extent for which I was able to implement it in SFX

One difference between Gold Rush and SFX, which you may or may not like, is that SFX populates (or used to, anyway) targets with all the little one-off documents they have like constitutions, speeches, single reports, single conference proceedings, and that sort of thing. EBSCOHOST databases are liberally sprinkled with such items. Gold Rush usually doesn't include one-off documents, because they naturally lack ISSNs. Much to my dismay, I realized long after submitting official statistics from SFX that when I ran the report analysis in SFX to remove duplicates, it was counting many extra thousands of "unique titles". All those one-off items were included. I wasn't experienced enough at the time to realize what was going on. So our official e-journal stats were over-inflated and the real number (reported from Gold Rush data and deduped in Excel) is now significantly lower. Awkward. Nobody searches a link resolver trying to find Woodrow Wilson's First Inaugural Address, do they? But on the other hand, these things are included in EBSCO title lists. I guess if I were running one of the big commercial link resolvers, I would feel constrained to include these items in the KB for the sake of completeness.

My recommendation is that Gold Rush is suitable for libraries that don't need a more sophisticated link resolver and can't afford a high price tag. Gold Rush works perfectly fine with journals that have ISSNs. So if you're at a hospital library, for example, and the vast majority of your journals are modern, mainstream titles with ISSNs, Gold Rush would be great. I can imagine it would be fine for school, small public, and community college libraries as well. Rose Nelson at the Colorado Alliance is really easy to work with, she'll take good care of you.

Workflow: we had a very organized workflow with SFX. The old monthly workflow with SFX was to see which targets had updated portfolios, download those targets' portfolio data, run it through the SASK::SCRIPT, and load the resulting file into the Millennium ERM to create/update brief bibs and holdings. These updates were tracked on an Excel spreadsheet by SFX target. Sounds easy-peasy, right? It was actually a very involved process -- I'm glossing over all the many intervening steps -- and I usually had to do a lot of cleanup on Lexis Nexis Academic Universe titles on the tail end. And I had to do the monthly work in two pieces because one single load would take too long and possibly overwhelm the systems involved.

With Gold Rush it's been more of a catch-as-catch-can thing. As I mentioned previously, there's no systematic way to tell what databases have been updated, so you're kind of guessing, or waiting to get an email saying that something's been updated. It took me a long time to convert the ALTLOOKUP in Millennium from SFX ID numbers to Gold Rush IDs (which are just print ISSNs with a few added characters). One plus for Gold Rush is that the reports are very easy to run and make use of. I was able to write a simple Python program to convert Gold Rush reports into data for the ERM Coverage Load. My little bit of inexpert code runs blazingly fast compared to SASK::SCRIPT which had to hit up ExLibris servers to generate each journal's URL. What's problematic is that Gold Rush titles don't necessarily match up to SFX titles, plus there are all those titles without ISSNs. It's been hard trying to get Gold Rush to synch up with what we already had established in Millennium.  You would probably have this issue transitioning between any link resolvers.

I am also tasked with updating the print serial holdings twice a year in the link resolver, plus manual updates as necessary. The big updates have proved more difficult with Gold Rush because title lists don't seem to load properly. But with SFX too, and I think with any resolver, getting your print serial holdings can be a long, painful process.  Email me off-list if you want to hear more.

I should mention that we generally do not buy full MARC records for e-journals. Instead, we use e-journal data to create brief bibs and checkins via the Millennium ERM Coverage Load. Our two catalogers will work on full bibs for e-journals during their downtimes between higher priority projects. They're only cataloging full bibs for journals in stable resources like JSTOR and Project Muse, not titles that are only in aggregators (like EBSCOHOST and Lexis Nexis). Note that the Millennium ERM Coverage Load does not overlay full bibs with brief ones, so there's no danger of the catalogers losing their work.

=============================================================================================================

Diane Westerfield, Electronic Resources & Serials Librarian
Tutt Library, Colorado College
diane.westerfield@coloradocollege.edu
(719) 389-6661
(719) 389-6082 (fax)

***********************************************
* You are subscribed to the SERIALST listserv (Serials in Libraries discussion forum)
* For additional information, see  the SERIALST Scope, Purpose and Usage Guidelines <http://www.uvm.edu/~bmaclenn/serialst.html>
***********************************************