�ݺ�ߣ

Bodleian Digital Library Systems and Services (BDLSS)
Bodleian Libraries
University of Oxford
Primom: A Mobile Interface for
Primo
Masud Khokhar
Applications Developer
Bodleian Libraries
University of Oxford
masud.khokhar@bodleian.ox.ac.uk
@mkhokhar
October 4, 2014
Andrew Isherwood
ICT Development
Information Services
Aberystwyth University
aai@aber.ac.uk
@andrewish

Primom: A Mobile Interface for Primo October 4, 2014
Page 2
University of Oxford Libraries
 University of Oxford
 More than 100 libraries
 Hold collectively approx. 11.1 million catalogued book stock
 470000+ ebooks
 50000+ electronic journals
 460+ databases
 66591 patrons (University and Library)
 The Bodleian group is the second largest library in the UK after
the British Library
 Ex Libris products in Oxford
 Primo, Primo Central, Aleph, MetaLib, bX, ARC and SFX.

Oxford’s Primo (SOLO)
 Two environments (Production and Test with plan to have a
Dev soon)
 Primo topology
 One BE server, two load balanced FE servers, two SE servers, one
NFS server, and one DB server.
 All virtualized except for the DB server
 Approx. 8M records.
 Multiple sources
 Aleph, MetaLib, Fedora and more in future
Page 3

Page 4

Why develop for mobile?
 More and more users are using their smart phones/tablets to
Page 5
access library resources
 Exponential growth in the users
 Approximately 20% growth every 2 years in the UK for the last 4
years.
 Attitude towards web based resources
 People expect a mobile/tablet ready version
 Disappointed when everything looks the same as desktop version
 Don’t care about webapp vs native as long as the experience is fast
and responsive
 Not always true (we love our iPhones)

Primo’s mobile interface?
 Out of the box ------------ 
 CSS based ------------ 
 Doesn’t play well with local customizations ------------ 
 Network traffic issues ------------ 
 Not tested enough ----------- 
Page 6

OOTB Ex Libris CSS
Page 7
 For some customers, works quite nicely.

OOTB Ex Libris CSS
Page 8
 For Oxford, not so nicely...

What about custom developments?
Page 9

Network Traffic
 Surely interfaces can be improved, you just need to work on
your CSS
 What about network traffic though?
 A quick analysis on Firebug reveals:
 Opening http://solo.bodleian.ox.ac.uk takes 57 requests
 Doing a search for “oxford” takes a further 66 requests
 Network latency can be a major issue
 Almost all instances of Primo launch 25+ network requests for
home page and 35+ network requests for brief search
Page 10

Network based issues
 Asynchronous vs synchronous calls?
 Your mobile search interface should not hang/die if:
 Your PDS is down (SSO check)
 Google or Amazon images is down
 Some third party tool you use is down
 etc, etc
 Similarly, not everything is required to be loaded after a search.
 Load only what is required.
 What about caching?
Page 11

Common feelings..
 We don’t have enough resources to invest in mobile even
though we would like to have a mobile webapp, an iPhone/iPad
app, and an Android app.
 And then after 10 days, you face an angry reader using Kindle Fire.
 How will we ever maintain so many developments, probably
best to develop just a webapp.
 Or if you can afford it, ask an external company to develop native
apps (which has unfortunately its own set of problems)
 We want to develop a webapp, what’s next?
 How to develop it (responsive vs framework)?
 How to link it with Primo functionality?
 How to maintain it? (growth and requirement changes)
Page 12

Blockers
 Responsive layouts vs frameworks
 HTML5/CSS3 (e.g.Twitter bootstrap)
 Frameworks (50+, we will talk about jQuery mobile and Sencha
Touch 2)
 Primo/LMS API Documentation
 Almost non-existent till developers showed keen interest
 Still quite poor
 Confusing – X-services, Web services, RESTful APIs, Plug-ins, what
to use when and how?
 Some good work being done by Barak Hecht but more is needed
 a lot more...
Page 13

Primo mobile development in Oxford
 Oxford started working on this informally about a year ago
 We face similar issues, not enough resources, not enough time
 We tried looking into iOS development but stopped
 Steep learning curve
 Only impacts Apple users
 Selected jQuery mobile to experiment with
 Low learning curve as we work with jQuery anyway
 Fairly stable framework
 Does quite a lot of things nicely
Page 14

First attempt – jQuery Mobile
Page 15

Page 16

Page 17

Page 18

Page 19

Collaboration – Oxford and Aberystwyth
 Primo-Discussion-List questions about APIs lead to a
collaborative effort
 Quick to start on collaborative working
 Some issues persisted
 Lack of quality API documentation
 Busy schedules
 New issues emerged
 Multi-lingual needs
 Complexity of Oxford against Simplicity of Aberystwyth
 The idea to develop a webapp that all Ex Libris customers can
use – everything controlled by setting files, etc.
Page 20

Page 21

 Did anything change?
 A lot in the background
 Extreme control over every parameter
 We soon realised that there are limits to what we can do with
limited screen space
 Facets?
 Page numbers?
 Search details?
 Number of results per page?
 Resulted in a re-think of the whole idea
Page 22

Oxford – Since the re-think
 Change in framework
 From jQuery mobile to Sencha Touch 2
 Pure MVC, far more responsive, the possibility of generating native
apps (with native device APIs), ease of using HTML5 local storage,
etc
 And it looks nicer (full support for theming with Sass and
Compass)
 Considering Ex Libris was asked a lot of CSS based questions, may
be its time to look at Sass or Less (reduced CSS size)?
 Change in development methodology
 Agile (Functionality based sprints)
Page 23

Page 24

Page 25

Page 26

Page 27

Page 28

What’s missing?
 My Account (Loans, Holds, Requests, Fines, etc)
 E-Shelf
 Facets (Can we afford them in a mobile app?)
 Modern browser and platform detection
 Feature detection
 Modernizr.js
 Balancing act (between functionality, screen sizes, processing
power, network traffic, memory, etc).
Page 29

Does it stack up?
 Reminder: Primo OOTB for SOLO:
 Opening http://solo.bodleian.ox.ac.uk takes 57 requests
 Doing a search for “oxford” takes a further 66 requests
 With our jQuery mobile prototype, opening up the main page
takes 13 requests and 76KB of data.
 A search for “oxford” takes a further 2-10 requests (depending
on how many different icons are being loaded)
Page 30

Does it stack up?
 With our ST2 prototype, opening up the main page takes 11
requests for 223KB of data
 A search for “oxford” takes a further 2-7 requests (depending
on how many different icons are being loaded)
 All further requests take 1-2 network requests
 Because of aggressive local storage based caching
 You do the math 
 And if it matters, Facebook’s biggest mistake was that “we
burned two years not working on mobile”. (Forbes, 9/11/2012)
Page 31

Thank you
 Questions?
Page 32

�ݺ�ߣ

Primom - A mobile interface for Primo - IGeLU 2012

More Related Content

Primom - A mobile interface for Primo - IGeLU 2012