IWMW 2008 Mashup Workshop

29
mashups: more than maps mike ellis, solutions architect, Eduserv http://channels.ourmedia.org/files/bill-hillary-mashup.jpg http://flickr.com/photos/psd/492139935/

description

Distributed computing - where data is consumed from external Web sites, sometimes 'mashed', or displayed in some other way on your own site, has become a powerful way of providing functionality, and requires little or no financial outlay or technical understanding. This workshop will look at some of the services available and examine some of the ways that they can be combined or otherwise used on your site and for prototype development.

Transcript of IWMW 2008 Mashup Workshop

Page 1: IWMW 2008 Mashup Workshop

mashups: more than maps

mike ellis,

solutions architect, Eduserv

http://channels.ourmedia.org/files/bill-hillary-mashup.jpghttp://flickr.com/photos/psd/492139935/

Page 2: IWMW 2008 Mashup Workshop

2

so..

• here’s what we’ll be doing:

• Introductions

• Mashup overview

• The “mashup pattern”

• Deconstructing some examples

• Techniques and tools

• Problems

• Solutions

• Build?

Page 3: IWMW 2008 Mashup Workshop

3

today

• people have a variety of backgrounds, skills and requirements

• the entire point is that you go away with what you need from today...

• ...so please ask questions at any time

• or (preferably) talk to me over a pint...

• ...or do it online - details at the end.

Page 4: IWMW 2008 Mashup Workshop

4

me, me, me

• I am..Mike Ellis

• I was Head of Web for NMSI for 7 years

• ..and now work in the “Professional Services Group” for Eduserv (I am a consultant who helps people realise their content vision using appropriate technologies)

• I’m fascinated with the real use of technology: I’m a geek but hopefully a “real-world” one...

•...in other words, the things I’m going to talk about should be useful (!)

Page 5: IWMW 2008 Mashup Workshop

5

about Eduserv

• Eduserv are a not for profit IT company

• We work to “realise the benefits of ICT for learners and researchers”

• We provide web dev, hosting, consultancy, ATHENS, CHEST, and have a Foundation who plough money back into the HE community

Page 6: IWMW 2008 Mashup Workshop

6

you...

• 85% of you are “web coders” or more technical than this...

• 42% of you have built at least one mashup before...

• 85% of you want to have the technical knowhow to build mashups

• ..but let’s have a quick round the room

Page 7: IWMW 2008 Mashup Workshop

7

mashup overview

• “more than one source”

• consumption of data and services

• mashup environment = paradigm shift:

• the “web of data”

• machines talking to each other

• RSS

• this is (or can be) a black-box environment

Page 8: IWMW 2008 Mashup Workshop

8

mashup overview

• more importantly, this is about approach

• services are “out there” and available

• data is out there and available (or should be)

• (and when it’s not, we can still hack)

• in general: “we are smarter than me”

Page 9: IWMW 2008 Mashup Workshop

9

in one sentence

“mashups let you do more with less”

(because other people have done a lot of the work already)

Page 10: IWMW 2008 Mashup Workshop

10

• identify the need from your users

• find data and service sources

• look to see who/what else is out there – and see if you can borrow or copy. View source!

• decide on an acceptable SLA, and work on contingencies

• determine how you’re actually going to do it based on these decisions

project pattern

Page 11: IWMW 2008 Mashup Workshop

11

deconstruction

• as with much tech, you can identify patterns

• and because of the “black box” approach, these can be generalised

• so let’s have a look at a couple of examples

Page 12: IWMW 2008 Mashup Workshop

12

mapping

“Museums in Paris with free admission”

http://www.archi-nova.net/paris/tips.html

Page 13: IWMW 2008 Mashup Workshop

13

• institution provides the data (museum name, location, review, description, etc)

• mapping service provides a way to display these locations on a map (plus some additional stuff)

• data and service is mashed and displayed on your site

mapping

data service

your site

Page 14: IWMW 2008 Mashup Workshop

14

mapping

1. Page loads the API

2. Call includes your key

3. Initialize() is called

4. Your data is included here

Page 15: IWMW 2008 Mashup Workshop

15

timeline

From UK Natural History Museum “dinosaur directory”

http://simile.mit.edu/timeline/examples/

Page 16: IWMW 2008 Mashup Workshop

16

• institution provides the data (exhibit date, description, etc)

• timeline service provides a way to display these dates

• data and service is mashed and displayed on your site

timeline

data service

your site

Page 17: IWMW 2008 Mashup Workshop

17

timeline

1. Page loads the API

2. Options for display

3. Your data

4. Display it

Page 18: IWMW 2008 Mashup Workshop

18

“RSS to Image”

From me (!)

http://electronicmuseum.org.uk/experiments/rss-to-image/

Page 19: IWMW 2008 Mashup Workshop

19

• services provide data RSS (title, description, link..)

• service provides means to analyse this content

• my code provides a way to mash data and display

RSS (news)

Flickr search API

display

Term Extractor

data service

your site

“RSS to Image”

Page 20: IWMW 2008 Mashup Workshop

20

why mashup?

• developing these things from scratch would be:

• expensive

• complicated

• bespoke

• ...and therefore silly...

• it’s a quick way to prototype and hack

• you get a shared resource pool for any problems

Page 21: IWMW 2008 Mashup Workshop

21

case study “design pattern”

• determine what you want to do

• identify service (www.programmableweb.com)

• sign up for API key (specific to URL root)

• identify your data source(s)

• identify any data services

• shape data sources to fit requirement

• drop these into the code

• shape the prototype

Page 22: IWMW 2008 Mashup Workshop

22

techniques and tools

• there are three major “mashup building” tools:

• Yahoo! Pipes (http://pipes.yahoo.com)

• Google Mashup Editor (http://editor.googlemashups.com)

• Microsoft PopFly (http://www.popfly.com)

• these do similar things but are for different audiences

• all require you to “think flowchart”

• we’ll look at Y!Pipes in more detail in a moment...

Page 23: IWMW 2008 Mashup Workshop

23

Yahoo! Pipes

• drag and drop

• “patchbay”-like

• RSS-focus

techniques and tools

Page 24: IWMW 2008 Mashup Workshop

24

Google Mashup Editor

• code, pseudo-code, html

• google gadget output

techniques and tools

Page 25: IWMW 2008 Mashup Workshop

25

MS PopFly

• nice but confusing interface

• requires Silverlight

• webpages, too

techniques and tools

Page 26: IWMW 2008 Mashup Workshop

26

problems!• you are relying on 3rd parties

• often no SLA – is this ok for you?

• what if the service, system or data disappeared?

• accessibility: a changing landscape...

• how do you measure goodness...?

• where can you get data from, and how long will it take to clean it up? (don’t underestimate this!)

Page 27: IWMW 2008 Mashup Workshop

27

solutions!

• think about the true mission-critical nature of your mashup. Is it REALLY likely that google will go down more than your in-house server?

• work out contingencies: alternative services or ways of displaying your stuff

• never, ever use a service that swallows your data!

• look at www.programmableweb.com for ideas, sources of data and services. Or just google it.

• encourage everyone to produce machine-readable data as a matter of course. It isn’t hard, really!

Page 28: IWMW 2008 Mashup Workshop

28

build?

• is there time...?

• if so, let’s build!

Page 29: IWMW 2008 Mashup Workshop

29

thanks!

thanks very much for coming along. I hope it helped more than baffled..

you can talk to me anytime – you should have my details by now, or go here:

google talk: [email protected]

twitter: http://twitter.com/dmje

blog: http://electronicmuseum.org.uk or http://blog.eduserv-psg.net/