How big is Summon right now?

I just noticed that the number of items Summon says we have within and outside of our collection.

Within our collection there are are:

65,022,106 items
Including:

• 51,749,144 journal articles
• 11,824,532 newspaper articles
• 629,475 trade publication articles
• 408,957 books
• 159,911 audio recordings
• 143,292 book reviews

With a language breakdown of

• English (63,431,926)
• French (166,902)
• German (126,925)
• Spanish (60,797)
• Japanese (45,541)
• Chinese (39,077)

When you extend the search outside of our collection you get
495,415,785 items

Including:

• 112,587,009 journal articles
• 361,490,384 newspaper articles
• 3,884,816 patents
• 3,551,703, theses
• 851, 262 books
  

With a language breakdown of:

• English (473,599,062)
• Japanese (5,571,128)
• Spanish (5,247,447)
• German (3,243,938)
• French (2,791,481)
• Portuguese (1,685,196)

That's a propos of nothing really except that it just struck me how big it is.

Google Scholar Widget with EZproxy

I've updated the Google Scholar widget we use in LibGuides and other places so that it automatically uses EZproxy. This means that the 'Find It @ JCU' link will appear regardless of whether the user has set their Scholar preferences to show JCU's openurl link resolver.

I tried this in the past but found that form inputs wouldn't pass through from the form through EZproxy to Scholar. But Helen Hooper raised the issue again so I took another look and found OCLC has instructions for allowing a form to use the ezproxy prefix in a form action.

Even better is that because the master code is stored in a LibGuides content box, and our LibGuides link to it AND because the gadget code I distribute is called from LibGuides I've managed to update all instances of the code with one edit. Chuffed.

Web Usability with (almost) no resources

Just starting a review of JCU Library's web presence. I'm adopting the Steve Krug methodology of user involvement early, often, regular, in small numbers.

I've been researching intently on different approaches and Krug's pragmatism is very appealling. I'm not a web usability consultant with a team and $100,000 contract (I've worked in that situation in the past - but as the client). I'm just a wee library tech head with wide ranging responsibilities and no dedicated staff. I can draw on my colleagues to help but they too have wide ranging responsibilities and little time to dedicate to other projects.

Simple points Krug makes:

• Usability testing can be done in house (it's not rocket surgery^tm)
  
• Any testing is better than no testing
• 3 users will tell you 80% of what 20 users will tell you
• Don't sweat on demographics/user selection:
  • Big problems will be big problems for everyone
  • There is no such thing as a typical user
  • Two male first year law students from a non-English speaking background could be as different to each other as they are from any other demographic slice
    
• One user test early is worth one hundred late
  
• Never stop testing

He convincingly argues why user testing has primacy over:

• Analytics: can tell you a great deal about what people are doing on your site, but not why they are doing it e.g. are they spending a lot of time on one page because it's exactly what they are looking for or because it's completely confusing and takes a long time to digest?
• Focus Groups: great for sampling users opinions and feelings in the abstract, but not great for determining whether your site actually works
• User Surveys: asking people to remember what they did, when, and how often will result in a cleaned up, fuzzy version of what actually happened. Nothing beats observing in real time.

This is a new approach for us which will be evaluated after the implementation of the resulting redesign. If successful it offers a new way to approach our web presence that incorporates:

• A continuous improvement model
• A user centred approach
• An evidence-based approach

We stand to learn a lot more about our users than we already know. This is a chance to remove our preconceptions from the equation, to get user input when we have conflicting opinions. I've often wondered when marking arbitrary decisions about web content arrangement whether I'm seeing the big picture or if I'm just sitting too close to the screen.

Step One is defining the site's main purposes. Stay tuned for what we come up with and more about Step Two...

Format, Philosophy Of

I've been working with Clair Meade on how we've mapped formats in Horizon to Summon and have learnt things that make me ponder.

In advanced search Summon lists a bunch of formats you can limit by that our collections do not use. If a user selects one of these they would get zero hits, not because we had nothing that matched their requirement, just that we hadn't applied the same format name.

The format names are open to interpretation, overlap and ambiguity – for example ‘Government Document’ and ‘Report’ and ‘Web Resource’ could all be applied to the same document, but we can only map one to an item.

I’ve realised that many of the formats we haven’t used are listed because they are used in the ‘Beyond your library’s collection’, but as that’s not the default search target I smell misled clients. I've suggested via the Summon Clients list that the ‘Beyond’ box be above the format dropdown and the dropdown be populated with available formats live depending on whether ‘Beyond’ is checked.

What is a Format?

On the Summon Clients list here was some discussion about whether an ebook is a separate format or just a book that’s available fulltext online, and that's the approach Summon takes to journal articles (no distinction made between e and print). Obviously anything can be digitised so every traditional format will have ‘online’ equivalents.


I think the basic problem is one of philosophical approach. We (‘we’ being librarians overseeing the profession in the evolution from physical to virtual information containers) have accidentally blended two approaches:

1. The format of the information content (a picture, dictionary, moving image with sound)


2. The format of delivery (for a picture that could be poster, painting, 800x600px gif; for a dictionary that could a book, a web site (OEDOnline), or an ebook (from say Credo Reference); and for moving image with sound that could be Super8 film, videocassette, DVD, streamed media, or a download of an MP4 file)

I guess I’m wondering how many of our users will be happy to know that limiting to ‘journal article’ will exclude web resources, government documents, and reports that have the same ‘content size and structure’ as a journal article.

I worry that we use one field to describe two very different facets of a piece of information.

Has anyone else pondered this and decided on an approach?

A Mystery Explained: The Cookie Problem with Link Resolvers

A long standing issue we've had with our link resolvers (SFX, and now 360 Link, do the same thing) is the situation where the 1-Click banner displays but instead of the article appearing below you get a message about settting cookies and following the login link in the error message doesn't resolve the issue. A recent question on the Summon Users discussion list provide an explanation of what is happening.

Apparently some publishers (Wiley, Springer and Lexis for example) have a setting in their web publishing systems that does not allow a cookie to be set if their page is called as part of a frameset. The 1-Click banner is a frameset calling the publisher's page. Technically it's an issue with the P3P header setting. P3P = Platform for Privacy Preferences. The consensus appears to be that IE and Safari are affected.

If you are ever confronted with this situation simply click on the 'Article not displaying? Completely lost? Need help?' link which will display the Find It menu showing all sources of full text. then click on 'Show Article' for your publisher of choice, displaying the full text without frames. Oddly enough once you've done this 1-Click works for all other articles from problem publishers for the duration of your session.

The long term solution is for problem publishers to change their settings. Many institutions and Serials Solutions itself have lobbied publishers to do this, so far with little success.

Summon beta soft launched at JCU

We soft-launched Summon, branded as One Search, late last week. Still poring over issues and reporting them to Serials Solutions, and learning a lot more about the product than we did, and scratching our heads about other things.

At the user group post-VALA a lot of other institutions reported resistance from liaison librarians, something I haven't had to deal with here, although some are quick to point out failings I don't see that as a bad thing, you can't address issues you're not away of. One reaction that I think could be wrongly interpreted as resistance, is the realisation that Summon is not universal in anything it does and is not meant to be the be all and end all of advanced research (at least not yet). I see it as a big step towards getting new researchers over the hump of where and how to start searching, and not a bad way of getting an overview of what's available in a subject area (the way some academics use Google). For more focused advanced research that requires comprehensiveness there are other tools, but by the time you need them hopefully you more comfortable with search and retrieval and have a relationship with your liaison librarian.

Each question I send to the client center opens my eyes a little more and reminds me of the 'cultural' differences between Ex Libris and Serials Solutions (sort of like McDonalds vs Burger King) I reported on back when reviewing our AARLIN participation. EL were a lot more open, focussed on client contribution and consensus, reporting issues widely, SS keep that sort of communication narrow channel between you and them. Both are responsive and there are pluses to both models and SS are widening those channels with reps responding on mailing lists which is good to see.

Things I've learned:

• Summon doesn't use Ulrich's to identify peer-reviewed journals (yet), so using that filter will exclude some journals that are peer-reviewed.
• If Summon doesn't get a response in 4 seconds from your catalogue it will have a link to the catalogue labelled 'Check availability' rather than the branch and loan status.
• Records disappear and come back again, I imagine because of re-ingestion of existing records
• Our library news item announcing it only inspired two pieces of feedback, clearly marketing is our next hurdle

I'm getting positive secondhand feedback from the liaison librarians in Cairns, so I live in hope. And SS are currently ingesting our institutional repository records (ResearchOnline@JCU).

Still quiet on the discovery layer front!

Access 2007 and network shares

Simplest of tips - if you've recently had Access upgraded to 2007 and try and open a database on a network share with macros you get a bunch of security prompts. Rather than go through the process of allowing them each time.

Click on the office logo
Click on Access Options
Click on Trust Center
Click on Trust Center Settings
Click on Trusted Locations
Tick the box next to Allow Trusted Locations on my Network (not recommended)
Click on Add New Locations
Select drive (directory optional as is the 'include subfolders' tick box
OK your way out of there!

This came up because our HDMS databases are on a network share

Twitter augmented Panel Discussion Day 3 Morning Plenary

This morning's panel was decidedly more wide roaming than yesterdays. With the chair tapping into the zeitgeist via Twitter. Lots of talk about user demand for multimedia and a brave attempt to relate what digital natives wanted/preferred.

One tweep said that her/his children searched for things in this order:
Youtube
Google Images
Google

Lots of real and virtual chat about access to ebooks in Australia and New Zealand and a little handwringing from the public library sector. Rights, metadata, full text and the near death of indexing and abstracting services unless they can be one-click to full text.

Teula Morgan pulled the blanket off the elephant in the room when she asked do libraries even needed to be involved if people were getting what they needed without our interference. No-one seemed willing to answer.

All in all a pretty interesting session, and confirming that itch in the back of my brain that sometimes libraries are a solution looking for a problem.

Day 2 arvo: APIs, eResearch and the Twitter vs Tagging discussion

Paul Bonnington from Monash did an overview of the explosion in data collected, collated, and visualised using automated methods, and some of the lessons he'd learned while trying to preserve that data for wider use than the original purpose of the research. Organisationally the area he works for sits at the nexus between IT, Library, Research Office and Academics and he underlined the importance of partnerships, you can't do this on your own.

He stressed that a sense of ownership of the data by the researchers was crucial to its preservation, and the importance putting the data archiving tool in place at the time of collection.

He cited a couple key articles and books but I was only quick enough to jot down The Fourth Paradigm http://research.microsoft.com/en-us/collaboration/fourthparadigm/ for a tweet, and I can't find his presentation online or on the CD yet.

As interesting as this area is I don't think it's going to be our direct responsibility to carry out the ANDS mission at JCU although we will have a role in supporting it and linking the data to the publications in our institutional repository that result from it.

Roy Tennant roller skated at a rapid pace over the possibilities of APIs to vastly improve services by augmenting them with information from disparate resources and he offered a feast of links to lists of APIs and real world examples. Between this and the Linked Data stuff I'm starting to think I should start removing myself from the operational tasks that currently eat up so much of my time, and doing more of the 'cool stuff' that Heather envisioned when she first created the job of 'library technologies coordinator'. Sigh. Anyway for my future reference here are two of Roys links to lists of APIs:
http://techessence.info/apis/
http://www.programmableweb.com/apis

We really need more librarians with these sort of skills because I think it's difficult to see the possibilities if you're not getting your hands dirty looking at this stuff. That said, to badly misquote Jacques Attali from memory 'We tend to do what the technology allows us to do, rather than developing the technology that allows us to do what we actually want'.

And I lost my Twitter virginity today. (If your a tweep search for the #VALA2010 tag or post VALA I'm @cockerilla). I mention this because the final session was Top Trends with a star studded panel that included, but was not limited to, Roy Tennant, Marshall Breeding, Tom Tague and Karen Calhoun.

This Geek & Poke cartoon appears courtesy of a Creative Commons Attribution-NoDerivs 2.0 License.

There was lively discussion, but the only 'trend' discussed was social tagging and whether libraries (and other cultural institutions) wanted to control, censor, or normalise it and if so to what extent and how.
In the land of Tweet the natives (and there were lots) were simmering about how the discussion was getting bogged down in the minutae of just one trend. I finally saw a use for Twitter - how would the discussion have evolved if the feed was displayed to the room, or even just to the panel or chair?
Apparently the VALA organisers wanted to display the feed but the venue couldn't do it. I'm not sure whether that sort of immediate, but quiet, feedback to a panel from the audience would improve this sort of event, and whether the feed would become the centre of attention rather than the speaker, but I'm kind of interested in finding out. No doubt chaos might regularly be the result until protocols evolved to make it workable. I can't help thinking that it might end up looking like 'The Word' from the Colbert Report, where the commentator says one thing while the text and graphics provide humorous counterpoint.

Marshall Breeding at VALA2010

Marshall covered a lot of ground. Some of the bits that stood out for me were:

• The Abby video that introduced VALA2010 could have replaced his presentation
• Web 2.0 is old hat, it should be part of the core of our apps not an add on
• We can't be device nazis - and discriminate against users with various devices (iPhones, iPads et al)
• Kuali OLE, Ex Libris URM, WorldCat cloud LMS are the revolutionaries in LMS's (as opposed to the evolutionaries
• Do you want to go open source if it's just replicating the functionality you have in your current system?
• APIs as an alternative to opening up proprietary systems
• Open Source tend to have more limited APIs than the proprietary ones
• Most library web sites are a list of links to disparate resources (as I say all the time but sometimes I wonder if anyone listens)
• Marshall's ideal looks a little like Summon...
• Deep indexing (particularly of book materials - Hathi Trust
• Mobile is coming, but have a unified approach there too
• Cloud computing is inevitable and efficient
• Economic restraint is not an excuse for slowing tech innovation, it's a reason to accelerate it
• Outlook for next five years
  
• most still using evolved systems
• more next gen LMS
• Resource discovery matures
• Mobile mainstreaming
  
• Transition from local to cloud

In response to a question from the floor said that the problem for Linked Data is the business case for IP owners to make content Linked Data (by definition freely accessible)

VALA 2010 on Twitter

Ok Twits, here it is:

http://twitter.com/search?q=%23vala2010

If you like that sort of thing

OpenCalais

I've seen so many articles about the semantic web it's a pleasure seeing someone actually doing it.

OpenCalais is a free (for under 50,000 requests/day) service analyses unstructured data and returns structured metadata, really, really quickly. And that metadata links to related 'packets' of 'Linked Data'. That's pretty meaningless without seeing it in action. But try pasting an article into http://viewer.opencalais.com/ and see how it breaks the data down into social tags, entities, people and events down the left hand side.
Sure it's a roughasgoodenough approach, but it still impressive, and has allowed many organisations to free up human resources and speed up throughput (and improve accuracy) when processing large wads of information. Or you could simple use it to add tags to you blog posts...

As Tom Tague says, "Match up a geek with a domain specialist and see what happens".

Tim Berners-Lee talks about Linkded Data on TED:


The following are just links and terms for me to follow up on when I have time but...

http://opencalais.com/
http://viewer.opencalais.com/
OpenCalais Marmoset
OpenCalais Tagaroo
http://semanticproxy.com/
Linked Data W3C standard http://linkeddata.org/
OpenPublish (based on Drupal) http://www.opensourceopenminds.com/openpublish
Powerhouse Museum (kEMU, REST API,
OpenCalais Gallery of parsers

VALA - Plenary

In Sunny Melbourne - just out of the plenary session delivered by Karen Calhoun (OCLC) who gave me a warm fuzzy feeling by iterating principles I've emphasised in the Library ICT draft plan.

The VALA 2010 intro video gave us all a (nervous) laugh:



Karen did give us some good new lines to use in trying to convince our colleagues that good intentions, commitment and sticking to an old blueprint for services is not going to cut it for our users.

"Take a fresh approach (...) give them that they want"

"Find it on the web, use it in the library."

"It's not enough to do your best, you have to know what to do and do your best"

and a Proust quote:

"The real act of discovery is not in finding new lands, but in seeing with new eyes"

She really pushed for the idea of outreaching collaboration and loosely coupled synching systems - REM: I post a link to the full presentation when I can find it!
Here it is courtesy of Karen: http://www.slideshare.net/amarintha/the-3109781

Historical Browser Statistics Visualisation

See this nifty little visualisation of the distribution of browser use from January 2002 to August 2009.

If you're still using IE6 you've probably seen a few web sites have pop up warnings about ceasing to support your browser - that's about 14% of the browsers out there. You'll have noticed the CMS does display a little weirdly in IE6. And I've just realised I'm using every browser they currently have stats for except Safari.

Drowning Not Waving

In my long list of things to look at more closely that I never find the time to is Google Wave.

A few QULOC-ICT members were playing with it but we lost steam.

If you have a Wave account and want a playmate I'm up for it. If you want a Wave account I've got 25 invites, just ask.

If I did find time I think I'd read Google Wave: Collaboration Reworked, Melissa L. Rethlefsen's article in Library Journal and I think I'd pay particular attention to these links she provides at the bottom:

• The Complete Guide to Google Wave


• Google Wave Samples Gallery


• Google Wave: You Need To Pay Attention to This


• How To Add Twitter to Google Wave


• 7 Things You Should Know About Google Wave


• What Problems Does Google Wave Solve?

Sun not setting on Horizon?

Interesting news on HORIZON-L from COSUGI chair Chris Hauser. SirsiDynix are resurrecting the Horizon Enhancement database, where the user groups can feed into development priorities for Horizon. So it's not dead yet... and this comes on the back of some discussion that SD are definitely not saying that Horizon is EOL. One wag commented that he wished SirsiDynix killed Horizon a decade ago considering that since Horizon 8 was dumped there's been more development of Horizon then ever before....

And, coincidentally, Marshall Breeding's Perceptions 2009 survey of library automation has been released. Oddly it shows that Horizon customers are more satisfied with SirsiDynix than Unicorn/Symphony customers. Weird.

As always with Marshall's survey the free text comments are very interesting. I found myself nodding in agreement with one Horizon user's remark about not being impressed with what was currently available in the ILMS market and that they were happy to sit with Horizon and look at other methods of presenting their collections while waiting for what the future brought for both print and electronic collection management. Then I suddenly realised I could well be reading my own comment!

Summon @ JCU or something else?

A lot has been going on behind the scenes with our Summon implementation. Serials Solutions report we are on track for a mid-February release.

Through the discussion list I've made contact with two other Horizon using 'Summoners', Dave Pattern at Hudddersfield and Emily Lynema from North Carolina State, whose brains I'm hoping to mercilessly pick in developing a regime of extracting deletions, additions and changes to Horizon records to update Summon.

As I've done all I can until SS get back to us with something to tweak the only outstanding task is naming the service. It's been briefly discussed at Library Management Committee.


So what are our options?

1. Call it Summon (my quick hunt of launched Summon sites shows about 80% of institutions have done this) .
   Pro: people moving between institutions will know what it is.
   Con: To the uninitiated Summon doesn't mean anything (but neither did Google)
   
2. Call it X Search (the name we currently have for our federated search tool)
   Pro: institutional continuity.
   Con: Meaningless to new members of our community
   
3. Call it something else.
   Pro: we could choose something self-explanatory.
   Con: What?

I personally have reservations about naming things that have little or nothing to with the function or service the thing provided. I think every university catalogue should be called 'the catalogue', every intranet should be called 'the intranet'. But I fall down when I think of a meaningful name for 'federated searching' or 'discovery layer' or 'web scale discovery tool'.

As a working title I've gone with One Search because Serials Solutions need to stick something in place of 'name' and it's easily changed.

In the workroom we've had a couple of brainstorming sessions with an emphasis on silliness, the results included:

• JCUgle (pronounced J Koogle)
  
• Thing that you must search if you dont want to fail
• Get a HD
• Spare Brain
• Better Than Google
• Everything@JCU
  

As you can see I am desperate need of better suggestions! Marks will be deducted for:

• Anything prefixed by a lower case letter (I'm over iThis, eThat, mTheother, xAnything)
• Joined up words with mixed case (they are so NinetiesSpeak)
• Trendy words and catchphrases (they don't stay fully sick for long, dude)

Bonus points for:

• Brevity
• Meaning

Please post your suggestions urgently!

Break, what break?

Outside of work, many people say to me 'Things must be quiet at work, with the students away'. Wrong. For many library staff it's the time when they have to catch up on all the projects they identified during the year that they couldn't spend time on because they're spending time helping students and liasing with faculties. For example the liaison librarians have been powering through the creation of liguides to replace the old subject guides.

For me it's the time of year we're I can do major system changes with a much reduced chance of inconveniencing large numbers of clients.

One project was to incorporate ezproxy prefixes on all the subscription resources listed in our electronic reserve collection which we hope will significantly reduce the numbers of off site users having access issues. When we first implemented ezproxy we adopted a model whereby we expected users to engage ezproxy before looking for resources. In response to the 2007 client survey we have been systematically embedding the ezproxy route to resources so that the client no longer has to think about engaging ezproxy.

Part of that process was virtualising ezproxy to ensure stability, previously it had lived on an old PC in ITR.

We've built it into our ejournal portal, federated search and link resolver, as well as our database listings, libguides and the hyperlinks in our catalogue. With the conversion of Reserve Online I believe we have plugged the last hole through which off site students can be trapped by being outside our domain - the links to Reserve Online embedded in LearnJCU (our LMS; Blackboard). Using the ezproxy building block for Blackboard has the added advantage automatically authenticating to ezproxy so that user logged into Blackboard is no longer confronted with the authentication screen for ezproxy.

We could now consider removing the 'Remote Access' button from it's prominence on our home page.

Not to say that we are never going to have a remote access issue again. Some problems we are aware of:

• Emailed Table Of Contents services bypass ezproxy causing problems for users
• Some local network environments, particularly in workplaces, have network restrictions that prevent ezproxy access
• Odd combinations of user environments and ISPs have caused problems that we haven't been able to diagnose, not least because they seem intermittent
• Some sites use javascript to create hyperlinks and ezproxy can't convert these to 'hold' the ezproxy session with them
• People on campus email links to people who are off campus expecting them to work

The workaround for most of these problems is part user education and part web programming. I've hacked together some javascript that takes a native URL and 'ezproxyfies' it, and we've been collecting examples of what other institutions do to come up with something we can provide to the JCU community to ensure their links work on and off campus - but just providing that tool isn't going to solve the problems unless clients are aware of why they are having the problem and that the tool will solve it.