00:00:00
Insights 2015: By Ed Parsons - Topic "Making SDIs Work" - Recorded at GeoQuest2015
00:00:06
My name is Ed Parsons. I'm Google's
geospatial technologist.
00:00:09
I think the the principal, the
concept behind Spatial Data Infrastructures
00:00:13
is a good one
00:00:14
Clearly no one would argue that
sharing
00:00:17
geospatial information with a broadest community possible
00:00:21
is a good idea. I think part of the
problem
00:00:25
is historic. The SDI's have largely
been developed by
00:00:29
traditional data publishers, mapping
agencies
00:00:32
Cadastral agencies, used to serving a
00:00:36
professional market in a particular way.
00:00:39
And, I think as a result of that in the
SDI's that you see around the world at
00:00:43
local, regional
00:00:45
national, even at a global scale, have
00:00:49
developed on technology platforms that requires specialised
00:00:53
and just do geospatial. And, as a result of that I think they have missed
00:00:57
many the opportunities, that the broader web,
00:01:00
which is more standardised, more widely
used, brings.
00:01:04
And as a result, I think often the content
00:01:07
that is published by SDI's, remains
invisible to the broader community web users.
00:01:13
I think usability
00:01:15
on the web ultimately comes down to a
question accessibility.
00:01:19
How accessible could you make your
information. And there's a
00:01:22
technology element to that.
00:01:24
And that's about being a good web
citizen, making sure that you publish
00:01:28
your content on the web,
00:01:29
that is accessible via a web browsers via a URL or a URI.
00:01:33
And then obviously, there are the policy issues as well.
00:01:38
And that's more towards Open Data
policies,
00:01:41
making sure licensing is is brought and is
easily accessible.
00:01:45
And on both of those aspects, I think
SDI's have gone quite a long way.
00:01:50
But, its 80% of to the job done. There's another
00:01:53
20% that is about making the data
00:01:57
web accessible, being good web citizens.
And I think still probably some work to
00:02:02
do around
00:02:03
licensing, with the recognition that the
end uses
00:02:06
aren't necessary going to be someone that
you know. Lot of the
00:02:10
the success of the modern App Market,
00:02:14
is built on the APIs and programming
platforms,
00:02:19
where a developer has quite a loose
connection to the platform.
00:02:23
Developer may not be aware of whose building
00:02:29
applications and what their
00:02:30
end user of those applications are doing.
And I think that's something that many
00:02:35
traditional mapping agencies and
government agencies are
00:02:39
uncertain about. They want to know
who their end user is, and sometimes
00:02:43
for valid reasons. But that's kind of a
counter to the way the apps are
00:02:48
develop these days..... Security is a complex issue
00:02:51
I and not specifically going to talk about Google.
00:02:55
Technology on the web is about, you
choosing to share information
00:02:59
with how you want to share it with,
00:03:02
and not sharing it with any one else. you And that I think now a days means
00:03:06
that
00:03:07
most of the services that you use, data will
be
00:03:11
encrypted, when it is REST, as we say as a service somewhere.
00:03:15
And it will be included in transit, say
when is moving from one server to
00:03:18
another server,
00:03:19
it will be encrypted. And modern encryption is very very strong.
00:03:23
Its almost impossible to break
00:03:26
modern encryption, that's widely
used.
00:03:30
In any said a reasonable period of time,
00:03:33
if the data is encrypted, and the
00:03:36
organisation is hosting your data, they are doing agood job. You should be
00:03:41
reasonably sure that your data is
secure. But I think
00:03:45
in terms of SDI, may be there is still a fundamental question
00:03:48
the people need to answer is
00:03:50
do I really want to share my data? And
by share
00:03:53
actually does mean its available to
anyone, to reuse it,
00:03:58
remix it,
00:03:59
combine it with something else. Because
once your data is
00:04:03
out on the internet, you're sharing, it's a bit like Mercury. It will flow in
00:04:08
whichever direction it chooses to flow.
00:04:10
And once its out there, you can't really stop
that.
00:04:13
So in case if I've chosen to share data, it will be shared.
00:04:17
If I choose not to share data, its encrypted,
no really gets access to it.
00:04:22
I think in many ways you
could argue that, the consumer
00:04:26
market is trying to do simple things. In many cases the data that traditional
00:04:31
SDIs are dealing with is much more
complicated,
00:04:34
have much richer data models and so on. But I think,
00:04:37
what we can take from the consumer
market is is a very
00:04:40
strong focus on the user. They are very
00:04:43
focused on what are the actual user needs
for a particular application.
00:04:48
How do we measure those? How do we see how well we meeting those needs?
00:04:53
How can we monitor if those needs are
changing? And I think those people that are
00:04:57
developing Apps today, be it on the web
or in mobile devices
00:05:01
have that very strong focus. In that you see
00:05:04
releases of mapping apps and mobile phones happening
00:05:09
every two or three weeks or every
month with
00:05:12
developments, changes and most of those
developments and changes are based on
00:05:17
their feedback of
00:05:18
monitoring how people are using applications, what's working, what's not working.
00:05:22
That user focus I think is
really powerful, really valuable.
00:05:26
And it's something that developers of SDIs, focused maybe more
00:05:31
government users could take on board.
I mean this questions is an easy one
00:05:36
...
00:05:36
interesting one. I mean, everyone has constraints. Even,
00:05:41
large commercial companies have to pick
and choose
00:05:45
how they invest their time and capital in building new services, new
00:05:48
applications.
00:05:49
There is probably more flexibility if
you have a community users thats
00:05:54
paying license fees or you're supporting
00:05:58
your services by advertising. Which is what Google does.
00:06:01
So, perhaps there's a bit more flexibility
there.
00:06:05
But I think, there's a more fundamental
00:06:08
question. Which is, what do we expect
in infrastructure to provide?
00:06:12
And often we throw too much functionality,
00:06:16
too much complexity in to an infrastructure.
00:06:19
In my talk at GeoQuest 2015, someone
asked a question about
00:06:23
the balance between, what goes into the infrastructure
00:06:28
and what goes into the applications that sit above that. If you look at, you know the
00:06:31
the really popular apps on your
mobile phone - the Google Maps, the
00:06:35
Facebook..
00:06:36
Instagram, their applications sit on top of a relatively
00:06:41
basic Internet infrastructure. Behind
all those apps,
00:06:45
there are web service and traffic that is
moving over HTTP.
00:06:49
You don't see that! That's invisible to you. And compared to the complexity of
00:06:54
application, that that traffic is quite simple.
00:06:57
Perhaps in SDI I terms, we've thrown too much into the infrastructure.
00:07:02
We cam maybe, back-off a little bit and say, this is in a very simple level just
00:07:06
about
00:07:06
sharing data on a website somewhere, and we let more of that complexity to happen in the
00:07:11
applications.
00:07:12
Which are by their nature, more focused
on vertical markets around particular user needs.
00:07:18
I think so, it's only Web GIS that solved
00:07:22
problems. So big issues in terms of publishing data.
00:07:26
And..
00:07:28
in some ways maybe, you can think of your Web GIS is the obvious application
00:07:32
that sits on top of an SDI.
00:07:34
And don't bring, web mapping into SDI itself, but think of it that sits
00:07:40
on top of an SDI.
00:07:41
And the SDIs is much more at this very
basic level, saying
00:07:45
at this particular URI, you will always
find information about this particular
00:07:50
land parcel.
00:07:50
And that means, that data is always accessible to people,
00:07:54
is consistently accessible, and then its
00:07:58
in a format that an application
developer, another agency can build
00:08:03
on top, with that reassurance that the
underlying infrastructure is not going
00:08:06
to change.
00:08:07
I think first of all be good web citizens. And as I said that's about
00:08:12
publishing all your content by default
on the web,
00:08:15
using as simple tools as possible you know. Using HTML,
00:08:19
using SVG, using the language is
that the rest of the web
00:08:25
use.
00:08:28
And looking at the emerging fields of
Semantic Web and Linked
00:08:34
data.
00:08:34
Because above and beyond publishing your data, to the next step
00:08:38
is to publish your data in a semantically rich
00:08:42
way, that makes it easier for people then
to build applications. Say that
00:08:46
if you talk about a highway, and someone else talks about a motorway, we
00:08:51
know you were talking about the same
thing.
00:08:53
That's the next step. And I think the third step
00:08:56
is to build an ecosystem that works, support that ecosystem
00:09:02
Say, I know that I've built my infrastructure that people won't necessarily come and use it,
00:09:07
unless I advertise, I support, I market
00:09:11
what I've built. You know there is a famous film quote today "If you build
00:09:15
it they will come!"
00:09:17
But actually that doesn't work! If you
build it, nobody comes unless you tell
00:09:20
them.
00:09:21
So be a good web citizen, invest and look
at the Semantic Web and then
00:09:26
foster the ecosystem around the
infrastructure you have built.
I think ..
00:09:31
Geo Portals are great, and they meet the needs of a particular community of
00:09:35
users.
00:09:35
But reality is, if you really want to meet
00:09:40
needs of some specialist vertical markets,
00:09:43
or particularly easy communities, most
government agencies
00:09:48
indeed many commercial organisations,
don't have the bandwidth and don't have
00:09:52
the expertise.
00:09:53
Those people that are smart build an
infrastructure,
00:09:57
the allows an ecosystem to exist around
that,
00:10:00
and allows application developers to go down a particular route, plug
00:10:05
into that infrastructure and build an
application that is
00:10:07
very focused on particular using needs a
particular market requirements.