Get notified about Mike Atherton

Sign up to a email when Mike Atherton publishes a new video

I'm not a developer I am a I've worked
in user experience for about 15 years
I've always believed Ennis had now
become fashionable that development is
very much an integral design of the part
of the design process can we have a
quick show of hands who is a software
developer Software Architect okay who
would call themselves the user
experience designer who would call
one and a half two and a half this is a
talk about information architecture it's
called beyond the polar bear and it's
familiar with this book this is a book
called information I approached you for
the world wide web it came out about 15
years ago it's very much seen as kind of
the Bible for ia and and it got that
practice kicked off in designing for the
web but a lot has happened in that last
15 years Google happened and social
media happened and smartphones happened
and IP TV happened and so the roots that
we take to access content these days us
are splintered they're more fragmented
and more kind of wibbly-wobbly than they
ever were before today we're going to
look at how the BBC who I did a bit of
work for a couple years ago now met that
challenge head-on to make websites that
better represent their core business
today and better deliver on this mission
that they've had since 1922 to inform
educate and entertain so for starters
let's look at the problem you see the
problem is that information isn't really
very neat and it's something that that
polar bear book teaches of this the
information is very hard to categorize
in many ways taxonomy taxonomical site
design hasn't really moved on from its
roots in library science where you had
to sort of fit one physical book on one
physical shelf and in the last few years
you know that we've done things
like tagging to sort of expose different
facets of data and thus unlock more user
journeys but still typically a you your
information architect is limited to
saying that that one thing is sort of
more or less specific than another thing
and so we do exercises like card sorts
the boxes and they will make site maps
like this in the hope that people can
start etre a homepage and work their way
down through categories and
subcategories and they're all going to
make sense and then we get frustrated in
user testing sessions when we find that
people aren't actually coming to the
homepage at all but deep linking
directly to the content from Google and
there's another problem at the BBC no
one really knew whether it was one site
or many sites in theory it's one website
not really a very connected cohesive
experience and then there's a massive
amount of content to consider we know
what we all know the BBC right they do
you know they put out 1500 programs
every day across eight national TV
channels ten national radio channels and
over 40 regional radio step channels so
difficult to find a way to
comprehensively represent that massive
output online and as a result the BBC
ended up building a web presence in just
about the most expensive way possible
these siloed teams working in isolation
hang cranking out these these different
sites different CMS's all kinds of
different processes and it led to all
kinds of problems the clusters of pages
that just weren't very joined up or
repeated content or things that weren't
linked to or things pages that were
mothballed whatever that means or taken
away from the web completely resulting
than nasty four or four errors and then
while I was there there was winds of
change they made a commitment to reduce
the number of ha ha no to half the the
number of top-level directories on their
site which is kind of a weird concept
for something that's supposed to be one
website and they also reduced their
operational spend they had to to slim
down as in response to government
criticism of BBC spending so this forced
the need for a new and more efficient
way to publish some content now sadly
these days information architecture is a
bit of an uncool term it's it's getting
backwards we're trying to reclaim it but
it's a subsumed of course into the much
wider sexier better marketed field of
user experience design
now UX is rapidly growing up and it's
rapidly sort of understanding its place
in the world and I go to UX talks where
people talk about business model
generation talk about psychology but
actually just a few years ago and
certainly at the time that this work was
going on UX was synonymous with the
interface layer it was synonymous with
with with interaction design if you like
a focus on presentation and interaction
and these projects which I'm sure as
developers you relate to start with
these beautifully conceived Photoshop
mock-ups which may or may not have
anything to do with reality the
perception was to some degree still is
that these kind of faked mock-ups are
cheaper and easier to build than getting
anybody getting their hands dirty with
cold certainly anybody who might call
themselves a designer doing there
and so these things get kind of
incubated in a back vacuum no one gets
consulted and these mock-ups get created
committing a lot of product decisions
committed creating a lot of metadata
decisions up front before really anyone
concerned technical development has been
consulted about the answer the possible
new certainly before anything gets put
in the house
a real user so this whole web thing
needed a bit of a rethink and some rebel
forces striking from a hidden base one
their first victory against the evils of
this siloed thinking they recognized the
power of the Semantic Web and how it
could be used to build bridges between
content extend user journeys and stitch
all of the BBC's offering into a single
extensible framework another thing about
ia usually when deny an information
architecture is doing their sourcing and
making their boxes to put stuff in there
thinking about a little bit how or what
categories to put things in you know
we're like librarians still finding that
one shell to put that one book on and so
you get things like boxes for news or
boxes for video things like pictures
here alongside things which aren't
really the same sort of thing like World
Cup now there's a couple of problems
with with with this approach one is that
as we say information doesn't really fit
into boxes so neatly like that it's
actually the connections between things
are richer and more graph-like and the
other thing is that when we think about
our favorite subjects we're not thinking
about these kind of containers at all
we're not really interested in a video
about and therefore were really more
interested in real things real world
things to deviate from the BBC a little
bit I'm a bit of a Disney parks fan so
let's just run through this little
exercise so what Disney World is a real
place it has a real location in Orlando
within it it has some parks of which the
Magic Kingdom is one Epcot is another I
could go on within the parks or a number
of lands each park has a weenie did you
know this the the the big architectural
thing in the center of each park is the
castle in the Magic Kingdom it's the big
ball in Epcot that helps people navigate
back to the center of the park
each of the lands has a number of
attractions those attractions have a
creator and Imagineer and may also be
based on a bit of prior art like a movie
or a book or something like that
the resort's also have hotels which
confusingly Disney also call resorts and
those resorts have restaurants as do the
parks have restaurants i to the lands
have restaurants and those restaurants
those meals are associated with a
particular character who might make an
appearance for that meal and that
character themselves might also be built
based on a bit of prior art so you can
start to see this kind of graph like
connection between these real world
things which wouldn't naturally fit into
a hierarchical site design so that
brings us on to domain driven design
which is something that's far more
familiar to many of you I'm sure this
book became the new Bible at the BBC in
looking to represent a subject online we
moved away from thinking about the
interface to first thinking about the
things the concepts within that subject
far way before we have a look at web
pages or anything like that
so domain driven design promotes it
proposes not a taxonomic structure but
more of an ontological one and a an
approach to modeling subjects that
matches users mental models which
naturally we call a domain model it's
really just a way to identify the
important things within a subject in the
way that we use it and the way that
those things connect together domain
model links down the first step for
building new projects at the BBC and
it's seen the move away from this
process that starts with kind of
wireframes and mock-ups of user
interfaces - one that starts with a
logical model of the subject that we're
covering and working upwards from there
something like this we've done for the
Champions League soccer soccer
tournament and you can see I mean I
don't know anything about football but I
can follow this but that we have a
competition that competition is broken
in two rounds each round has a number of
matches those matches are held in a
hopefully unless Inge was playing
there's some goal scored the matches are
two teams pitted against each other
those teams have players the goals of
sports will score by a player you can
start to see how these things connect
together and as I say I mean I don't I
don't know anything about football that
doesn't really matter because I can go
to a domain expert I can go to somebody
who doesn't need to know anything about
web pages or user experience or
information architecture and certainly
not about software development they just
need to know about their subject so I'll
go and I will ask them lots of questions
in a and and get an understanding of
their world get them to unpack any
complexity in their terminology find out
exactly what terminology they use and
then I'll go and talk to some end users
find out what they care about what what
language do they use and not on the not
on the field but but to describe the
subject I like to say that experts help
you map the territory but users mark the
points of interest so we know that that
can be challenging sometimes especially
if you've done any information
architecture because while people might
appreciate the different social
differences between things they can such
they can struggle to articulate what
those differences are I like the the
novel well I like weathering highs but
what do I mean by that good I like the
novel or maybe I like a movie adaptation
that I saw maybe I liked but Kate Bush
songs if it is the novel then do I mean
this or do I mean any of these things
which are all different things but
they're all still weathering Heights
than all so you see you start to see how
you need to break down this complexity
and perhaps in this case talk about
different additions because people could
be there people can understand that
there is a difference between something
like the Beatles album Rubber Soul and
all of these other things but they can
struggle to articulate what they mean by
that but defining the damn things is
vital if you want to be able to point at
them to take the Wuthering Heights
example now I expect that most people
want to go copy of the novel are
thinking in general terms they're not
necessarily and yet if you go to Amazon
they have really no concept of watering
Heights in general terms to them two
different editions of weathering Heights
are as different as a copy of weathering
Heights and a copy of Pride and
Prejudice so when you google weathering
Heights you get a bunch of different
Amazon results all kind of splintered
all over the place it kind of fragments
there Google juice splashes that Google
juice I guess you'd say so it's bad for
SEO and then it would be to have
something that was much more about the
general concept of weathering Heights
and could have some other things that
are hung off of that so we tend to think
when we're doing some of this modeling
at this basic level which people are
searching for things that people think
about things and try to represent a
canonical version of that concept this
is probably sounding very analogous to a
lot of them of course the good news is
that we don't forget it right first time
this domain driven process is heavy on
iteration right down to a very
fundamental level but with that shared
model and a shared understanding of your
subject domain
along with the shared lexicon to
describe it
you get that consistent user experience
based baked in from the ground up so
domain model modeling is a big deal I'm
sure many of you come across it before
if you haven't there's a good book on it
out there in the lobby
it's it's it's it's heavy going we take
a quite light approach to it if you like
but it gives you some idea of the
process that the BBC has put into
practice but let's look at this practice
now first with how the BBC programs team
met and realized a long-held ambition to
provide a permanent pay a page on the
web a persistent location for every
single one of those 1500 programs that
they've put out every single day now TV
is a bit of a messy business actually it
was never really be designed to be put
into much order sometimes one-off TV
shows end up being the pilot for a
longer series sometimes schedulers will
decide that like a two-part story is
like a miniseries within a series as is
the case with Silent Witness here a
general sense of a show like Silent
Witness or the Glee or Mad Men or
Breaking Bad is what the BBC would call
a brand and it's different to a
particular season or series of that show
and of course with what I say of course
but because I know how this works but
believe chiapas so it even comes in a
different a number of versions
apparently deaf people like to watch TV
at 3 o'clock in the morning in my
country and and so different versions of
the same episode will go out sometimes
with subtitles or sign language and
things like that Sherlock was a show
that was popular I still still popular
when I first came out I only had three
episodes so what does that make it does
that make it season one does that is
would happen if they made any more they
also had an associated TV channel BBC
one it went that it aired on if they
made any more Sherlock which of course
they could then on and did what what
would those extra episodes be would
you know a number of episodes to
should we say this is Sherlock series 2
or should they be tacked on to series 1
there was there wasn't really an answer
at the time that Sherlock came out and
then how does that affect what now
Sherlock is and now we've got another
level of abstraction saying we have the
Sherlock brand and within that we have a
number of series and within that we have
a number of episodes and in fact this is
the way that it has gone and there's now
series 3 coming soon very exciting
it could also switch to another channel
conceivably it could switch from uc1
juicy to a lot of a lot of shows on the
BBC they've started out on on some of
the lesser channels that with the BBC
importance it is the less the viewers so
things started on BBC three sometimes if
they're really successful they might
move to c2 if they're really successful
tonight we do see one
so it's problems like this that the
programmes team have to make sense of
when they're defining some kind of model
so they went to schedulers they said
okay what's it all about oh there we go
and they came up with this this was that
this is the first cut rough domain model
for BBC programmes and in it we see that
there is a brand like Doctor Who
GOx who has a number of series which
again depends when you start counting
1963 2005 within each series there's a
number of episodes each of those
episodes has a number of versions every
one of those versions is either gone out
on air broadcast on air at some point
always have been available on demand at
some point on the catch-up service
iPlayer they collectively they go out on
something we sort of called services
which was an attempt to rationalize TV
channel and radio network and things
like that
we call them programs but they have a
format like drama and a genre like
but with that model in place they could
populate it with business data from
around the BBC information that
schedulers already had to hand and came
up with something resembling webpages
for the first time only after all that
modeling work had been done but because
those things and those relationships had
already been put in place they'd already
been defined in a data model actually
outputting a ton of linked webpages
became relatively straightforward and so
for the first time in history a
broadcaster had represented its entire
output persistently and scale of lis
online the approach generated one single
page for each thing each thing in this
case being the program one page for a
program with one URL to locate that
look for a second or something like
Wikipedia many many thousands of single
pages each of them very focused on a
very single topic having a single URL
makes it easy to point at that thing
blog about and tweet about and generally
share well internally it means that the
content teams have a single point of
reference to put all their multimedia
niceties to enhance the page and it
focuses all the although of search
activity or the Google juice squarely on
a very single location making it easy
for people to to find that particular
episode which brings me on to URL design
URI design an important though often
overlooked part of user experience
design I think the beaver use some
guiding principles in making nice web
addresses they should be persistent in
other words you know always reliably
there they should be hackable so you
know you can chop them back at this
flashes and still get some meaningful
content and they should ideally be human
readable because you want something to
put on the side of a bus or something to
read out over the radio now balancing
those needs can be a very tough tough
and as this hierarchy of needs indicates
the most important one of all to get
right is persistent after all as Sir Tim
says cool your eyes don't change your
web address is er is an implicit
contract with everyone who's bookmarked
or linked to your site and if you break
it you're breaking the web but how do
you maintain that kind of persistence in
that messy world of TV with content
that's all subject to change what do we
need to look at well I think we need to
get it get rid of everything that
changes in the URL everything that's
subject to change this is a URL but we
could have used for one of those
episodes of Sherlock how many people
have this kind of thing at the end of
their URLs why nobody cares about a
technology stack if you decide to
completely rewrite it in PHP or whatever
are you going to keep those extensions
they're not meaningful and then that
so that leaves us with something like
that which is fine kinder I say in that
in that Sherlock example it was a little
bit ambiguous we didn't really know
there was a series one until there was a
series two and sometimes when shows I
should have sold over overseas sold to
America they get repackaged American
networks don't really want a three
Episode series that goes nowhere they
want they want something that's a little
bit more a little bit more complete in
itself and so things get can get
concatenated and made into different
series so the notion of a series sort of
that leaves us with something like this
which is fine but as I say many brands
have made the leap from one TV network
to another some have even made the leap
from radio to TV so the idea of a of a
particular TV station channel
owning a brand now and forever is a
that leaves it with something like this
which is nice but let's say we didn't
use Sherlock let's say we use something
called the breakfast show or drive time
can we honestly say that between now and
the end of time there will never be
another show with this name and what are
we going to do if there is so I'm afraid
it has to go
which leases with this which is
so in fact the correct URL is this one
not very pretty is it but it leaves
about just about the only thing that we
know for sure which is that this is a
program and that's it it's not pretty
and it leaves a URL that's only partly
human readable but we've made damn sure
that it's going to be persistent and if
we want something to put on the side of
a bus we can use nice 301 redirects to
have vanity URLs that will help take us
back to the canonical URL for that
episode so on we go so the program site
launched not with a bang but with a
whimper giving quietly giving people
URLs to link to when referring to any
any BBC program any episode of any
program from the worthy and educational
ones that make good use of the BBC's
public funding to the ones that people
actually watch know no longer would only
those very popular web brands have a
popular program brands have a web
presence now did it set the world alight
actually not really at first it was a
huge feat of information architecture
software engineering but you know as
they say content is king and when these
pages first came out they didn't really
have a lot of content on them they the
programme metadata a few broadcast dates
that kind of thing it was very much the
walking skeleton that needed filling out
but pretty soon the the program
production teams realized that they now
had a million points of life that they
could illuminate more they had these
very defined locations which they could
enhance with extra content first of all
by adding a copy of the program itself
onto that page which is pretty cool
to the iPlayer catch-up service and
things like production stills and video
clips and contributors and actors and
things like that
oh and track listings which is an
interesting one this is an episode of
Mad Men that has Bob Dylan track on it I
mean it's great soundtrack from 60s
music but Bob Dylan is a thing and he
also not only exists in our programs
domain but he also exists in our music
things with Bob Dylan is a boundary
object we'll just unlocked a new user
journey to take somebody from that cool
song they heard on Mad Men - all about
Bob Dylan pretty amazing huh so program
pages became the standard for
representing program content on the site
and gone are those days of those teams
turning out those kind of hand-cranked
siloed microsites you know the ones with
the you know downloadable wallpaper and
that were full of glory while the TV
show that they were talking about was
hot and then just kind of left to wither
on the vine when it was not instead
there's this completely scalable
ecosystem that comprehensively
represents all these different programs
going out to air but still people do
quite like those funny little micro
sites don't they so the content teams
have been given tools to make these
branded support pages which to all
intents and purposes do the same job
they're just as rich as anything ever
was but instead of being siloed they're
stitched into this wider fabric there
even believe it or not looking to model
things inside the content of a program
in like narrative events plots events
and characters and their journeys with a
view to recontextualizing stories beyond
the confines of a 30 or 60 minute
episode of a show so
modeling programs is really only the
start the story where if you've got a
business and a Content stack that's
largely built on having this massive
content archive and you can point to all
your programs and even to individual
bits of your programs then you've just
unlocked a massive amount of content
that you can use to make more products
like this one
the BBC food site was doing or I it's
been around for a long while never
really that great it's a recipe database
so you know you might see something on
TV that you want to cook and you can go
and look up the recipe online well I
don't know about you but if if I want to
find the recipe for apple crumble then
there's only one place that I would go
first and that's Google and I'd probably
get a ton of recipes back and I pick one
from the first couple of results in the
in the recipe sites market if you like
SEO is therefore hugely important and
frankly the SEO the BBC food site wasn't
that great oh it had the apple crumble
recipes of course 280 odd of them again
splintering the splashing the Google
juice all over the place not really
focusing it efforts on a single location
so how then could SEO be improved and
more value from that content be a lot
well surprise surprise we went back to
what people think about and in this case
the things that people are searching for
here's our happy stock photo couple and
in my modern sensitive view of the world
men think about Nigella Lawson and women
think about chocolate
when he thinks about chocolate he thinks
things about Nigella is thinking about
chocolate cake it gets no worse than
this I promise
meanwhile she's thinking of a a show
that she saw this morning on which there
was some recipe talks about and so they
decide that they're gonna make a
chocolate cake so great minds thinking
alike they're let's put some labels on
that we have a program we have an
ingredient of chocolate chocolate we
have a dish this idea of a chocolate
cake we have a chef Nigella and we have
a particular recipe
let's neat nap up a bit and look at a
very simple domain model for BBC food at
its heart is the recipe it's written by
a chef and it has some ingredients we've
got some little interesting
round-tripping going on here if your
recipe is for something like chicken
stock or corn beef or something then
once you've made it you can then use it
as an ingredient another recipe because
this is the BBC that recipe has come
from a program and then there's the dish
which we'll get back to in a second but
think about recipes for a minute there
are interesting sources of data aren't
they a lot of lot of information in
there but usually it's kind of just
baked in if your pardon the partners
there's this very static on exploitable
text but through applying these domain
modeling techniques the food team were
evil ever able to make each ingredient
and even some techniques into a
pointable thing and so you know they
were able to move the site from sources
to resources if you will that was the
job and he said and you can see so so
that each of these things the lobster
the boss of the champagne even the
technique of removing meat from cooked
lobster there are all now pointable
things and we can we can pull together
pages on everything that's made with
lobster or indeed everything that's made
with champagne we can have a page about
this technique
and and then reference all the recipes
that it's used in so by modeling around
the structure of a recipe they unlocked
new ways of exploring food and indeed
the subject of cookery show me all of
Nigella's cupcake recipes show me what I
can do with this zucchini that was the
other job most important was the concept
of the dish and this is again it's
another example of having that kind of
canonical work a mental image of
spaghetti bolognaise if you will but
it's distinct from any individual recipe
for spaghetti diagnosed at which there
might be many I never knew this about
recipes when I when I first worked on
this but they're very rights protected
the chef's put them online they
published them but after a while they
want that recipe back and that again if
you would if you were linked into all
those recipes individually you get a lot
of broken links a lot of 404 is that
kind of thing but if you've got this
canonical spaghetti bolognaise concept
then that holds firm even though
this focus on unlocking many new and
into woven routes to concept along with
a big push on SEO friendly language it
massively boosted the internal link
density of the site and the
opportunities for people to link
externally to the site and as a result
it doubled traffic BBC food now is about
20-25 percent of the UK's directed
recipes market at this at the time that
they made this change was a hundred and
fifty thousand uniques every week from
search engines alone but let's look at
those users for a second and think about
something else
in their first year they saw 20 percent
rise month in month on access from
mobile devices makes sense right this is
a recipe site people are getting iPads
for Christmas and smartphones and stuff
they want something that can take into
the kitchen 75 70 percent of traffic to
Google more than any other source if you
know how percentages work but but
inevitably far more than the BBC food
site itself so some essentially people
are coming directly from Google so the
question is which is the real BBC food
homepage yeah in my world of user
experience design designers go crazy for
these kind of pages right these are the
ones I show off to the clients home
pages you gotta have a good homepage
right everyone likes the home page
well increasingly that could have been
the home page because more and more
traffic was coming from mobile but
really if you believe the traffic
well this is the home page so that old
rhetoric that the home page is your most
important page is just plainly not true
if you think of it in terms of effort
70% of your effort should be spent on
making a good recipe page you would
think page and only 30% on your actual
home page because no one cares about
so finally let's look at what's perhaps
the standard barrier bearer for this
domain driven approach at VBC a site
whose subject could be could be said to
inherit and evade domain model and a
site that needs more data than even the
mighty BBC can provide the BBC wildlife
finder is a showcase for Natural History
footage something that the BBC has this
fantastic reputation for for being the
best of the world but far more than
trying to be a simple video
Gallery the product aims to
recontextualize these clips present them
in a new and useful way that tries to
educational informative and entertaining
clips all that works I've been pulled
from thousands of hours of archived
footage and the learning and this is the
bit I love but the learning is not just
in the clips themselves but it's in
showing how the species and adaptations
and behaviors neckl zones of the of the
natural world are connected what for
example does the polar bear have in
common with the bar now anyone I'm sure
there are many things but you weren't
gonna pick polygamists were you where
one boy polar bear gets many girl polar
bears but not vice versa
what else is polygamous will they
hippopotamus why do we know about
grassland it said adapted to swimming
it's hurt or herbivore and it's on the
vulnerable list all these links are
actually teaching me new facts these are
not just random related links like we
used to get on the right hand side of
the page this is this is about showing
me how this piece of context it fits
into a wider universe as you can imagine
the domain model is rich it's based in
part on the Linnaean taxonomy of that
guy who lived over there by a furball
into a classification and it's a huge
subject right this isn't even a really
good representation and while the BBC
has this Natural History Archive
stretching back over 50 years that only
provides the video clips there aren't
teams of BBC elves working away in the
content mines to write descriptions for
every creature in the animal kingdom ah
well there are actually but they do it
on Wikipedia
yeah you see the wildlife team are fans
of using the whole web as their content
management system you need a description
for an animal or pull it in from
just some Wikipedia entries suck yeah so
edit them on Wikipedia that way not only
do you benefit but everyone else
benefits as well and it's the point of
Wikipedia right but even that wasn't
enough the clips in while Wildlife
finder just defined as being segments of
programs and so they live in the
program's domain but there's other data
that we needed to pull in from other
sources conservation state is we don't
got any of that stuff but we can pull it
in from the animal diversity web
University of Michigan and some
distribution data from the worldwide
life well i fund and this gets pulled in
a data level which hints at the power of
the linked data cloud which is a subject
for another day but essentially treating
you the whole of your where the web is
your database from which you can draw
bits of pieces of content to plug the
gaps in your own offering most of this
stuff is free well I'd say all this
finder are our canonical things again
and they can be used like Lego bricks to
build scooped cool stuff on top the big
addition that they made to this product
was curation addressing one of the
criticisms of this approach which is in
that in many ways it's kind of too
democratic you need a little bit of
editorial voice a human touch if you
will though the ability to feature
certain bits of content curation here is
a layer on top of these resources it
collects together some handpicks
examples and uses them to tell a
particular story collections add new
context and timeliness and promotion
it's an alternative view through the
eyes of storytellers or our filmmakers
or different events that are going on in
the world and that kind of curation
really enhances this experience but does
it without undermining the
the overall structure so the success of
these projects changed the way the BBC
now makes websites from music to the
solar system to the Olympics to the
World Cup to history and increasingly
all of knowledge and learning that
domain-driven design means that
everything not just links to but
actually seamlessly integrates with
everything else so what can we learn
from this well as user experience
designers are people who are working
alongside user experiences einar's
understand and accept and get them to
deeper than presentation it reaches down
through all the layers of the software
stack and UX design has really asked to
get their hands dirty now not just
software development are not just
prototyping business logic SEO or point
ability document design URI design all
of these things have a very profound
effect on the experiences that users
will have on the product we need to
think about the web as a whole and how
our content fits within it the world
wide web didn't evolve by chance it was
invented by a guy who said that its
universality was its most important
concept we're supposed to use this space
to share information not build ourselves
these live at these ivory towers so make
everything that you do find a bull and
share a bull and searchable and
pointable it's all about the remix and
go bottom-up everywhere don't start from
wireframes or user interfaces but from
this mental model of the subject that
you're trying to cover put most of your
effort not into your homepage is a
top-level page but your bottom level
pages into your actual core content
we say we should design for our least
able user first actually your least
stable user is the Google pop there
that's that's who probably your most
common user as well the person who needs
to make sense of your page once all the
CSS gobiins has been stripped away
most traffic will come from deep lengths
so think of Google as your front door
markup / mock-up I mean these days I
don't need to say too much of this I
mean this is this is some when I first
started given this - a couple years ago
the idea that a designer should get into
making things in HTML and CSS and not
photoshop was heresy but we've got got
warmed up to the idea now and even if
they can't do it then lend them your
skills because because we need to be
working within the native fabric of the
media designing Chrome and an even
responsive design I mean is our huge
leaps forward but ultimately what we're
about is designing the interplay of
content and you can't do that from a
static mock-up how much more effective
would your prototypes be if they were
powered by real content and put in front
of real users real fast
so yeah the web has changed quite a lot
since the polar bear book came out it's
gone back to its roots away from silos
and towards the shared network of
information it's shifting from dumb
documents to smart things of course it's
still about giving people the shortest
route to content and making it easier
for people to find what they're looking
found it but it's no longer about this
series of private libraries it's about
it's no longer limited to interaction
design visual design we were actually
now in a web of data and we have the
power to obtain the roots and branches
of knowledge itself there's rich social
graphs to mine our appetites for
infographics and data journalism and
growing free repositories of our free
data to exploit so consider the whole
web your canvas and make your content
mesh seamlessly into it design for a
world where Google is your homepage
Wikipedia is your CMS and humans and
all right there's a couple of questions
in first one is how do you handle the
ddd terms across languages and cultures
that's a great question
I don't have a good answer for it
because we haven't had to deal with it
but one thing that we are about is when
we're defining concepts we're defining
you our eyes to do to to to point to
those concepts and a single URI of off
of which could then then be hung
different language abstractions if you
like there were there are got though
difference of conceptual differences a
little harder to deal with but
differences in in languages could be
could be do something you say
disambiguated at URI level I guess and
related to the same question from the
same user how do you capture and
document the times yeah that's that's
something I can answer more easily I
mean that's that's that's a
old-fashioned user experience research
process we talk to a lot of people we
listen to what they're saying when
they're when they come out with strange
terminology we ask them what they mean
by that and we find these terms and we
take that expert view and that expert
terminology to users and we've sort of
Road testers if you like so that we can
see which terms have relevance to users
which terms are the same which concepts
are the same but have different
terminology which you know which which
concepts are not at all interesting to
end users and the idea then is to is to
produce to come up with a best fit model
that is authoritative enough to appeal
to experts but accessible enough to be
useful to users and then this one final
question that was actually also on my
mind is set with a Twinkie smiling and
it says did you lick wildlife animals to
recipes may ask something there because
there's actually a question there their
had long discussions about boundary
objects exam yeah something that should
be the same in one comes context and
isn't in another I mean also we had we I
think at the time that we're doing a lot
while I find that the first Narnia
movies were coming out and there was a
big discussion about whether our lions
page should include fictional lions as
well or not at the end of the day you
just have to make a call about these
things and and it's about understanding
the mental model and it's not always
about being you know absolutely
technically correct and the other thing
that I showed that I don't talk about
here but is that is that just because
it's in your model doesn't mean you have
to expose it to the to the website
straight away you can choose to make
that call later on all right that's what
I had of questions is there anyone want
to take a question here at the end no
well then please remember to vote and
thank you to Mike