[Geowanking] a modest proposal for google maps
Chris Holmes
cholmes at openplans.org
Thu Jul 21 09:31:25 PDT 2005
Great thoughts guys, this is in line with much of my thinking of late as
well. I do wish they would start up their foundation, which has been
pending for a long time, as it could encourage the geospatial web with
the clear intention to 'be good', and not simply avoid being evil. In
line with their stated intentions for the foundation it could be a
great way to leverage their resources, both human and storage/hardware,
to do good in a way that really no one else can. They could encourage
radical experiments for good, like trying out open street map ideas by
opening up their satellite photos to have the derived works be licensed
under open geodata licenses. If the tools are easy enough wikipedia is
certainly pointing to the fact that it's _possible_ that user
contributed and corrected data can rival and sometimes even beat the
traditional ways of doing things.
They could also leverage their data capabilities to host geographic data
for developing countries. And indeed important environmental data and
the like that affects us all. I'm in Zambia right now, looking at the
potential for open source software to make geographic data more
available. And by far the biggest problem is sustainability of any
solution. I could spend 6 months and gather a bunch of data and set up
a great GeoServer/MapServer/GeoNetwork/MapBuilder/UDIG suite. I even
have a server that was bought awhile by some donor to help this. But
there's no hosting, and there's no one to maintain it after I leave.
Donor funding is generally more geared towards flashy things, the
initial set up, but not the grunt work of maintaining it. The
government has a lot of spatial data they'd be willing to share, and
indeed many donors come along and create geographic data for thier
projects. But the primary output is not to create geographic
information, it's to install radio towers for school learning in the
ideal place, for example. Many would be happy to share their data, and
would be even more happy if it was easy to get data from others.
Instead of each donor agency and country having to maintain a server, a
large central site could take care of that. Combine WMS/WFS with
offline access, mailing cds/dvds with selected data sets and open
source GIS software, and informed decision making and government
transparency could greatly increase, and time spent finding data and
replication of work could decrease.
Combine this with open street map ideas, and you could get accurate road
information in areas that just aren't profitable enough for Navtech and
tele atlas. And if it is built on open source software then national
mapping agencies can take over responsibility when they're ready, and
tailor the solution to meet their particular needs. But they could
start sharing mapping data _now_ if a good, easy to use server was
available. I want to pursue this when I get back to the states, but
the geographic data is so large that it'll require more resources than
just a couple fast computers and a decent connection.
Ok, I'll stop rambling about my dreams for spatial data in africa. I
agree that the aggregration of hack layers is key. There needs to be a
way to get data _out_ of the google hacks, so you don't just have a
picture of the data, but the 'source code' to the map.
One thing I remember from Google's ipo stuff was a bit about short term
vs. long term: 'If opportunities arise that might cause us to sacrifice
short term results but are in the best long term interest of our
shareholders, we will take those opportunities.' I hope that it's more
than just losing money paying data providers to take all the best hacks
for themselves, but instead realizing that they could play a huge role
in nurturing a thriving geospatial web and position themselves for
'search', against everyone's data, instead of trying to making 'their'
spatial web the best.
The keyhole stuff is really annoying, I only recently looked and
realized how close KML and 'network links' are to GML and WFS/WMS,
without ever explicitly saying it so that you might be able to use the
same tools and perhaps coexist. That one really feels like embrace and
extend, since there's no way they came up with KML without looking at
GML.
But I remain optimistic, creating the maps api in response to the
hacking was a large step in the right direction, and perhaps keyhole
just missed the 'don't be evil' memo. I just wish they would be more
transparent with their plans for the future - I don't understand why so
much secrecy and so many Non-Disclosure Agreements are needed if your
intention is to do good.. I'd like them to see themselves (and to be
seen) as collaborators with a wider community trying to do good,
instead of an autistic 600 lb gorilla convinced its way is always the
best. And the role of OS geo hackers should to change in response,
instead of wringing our hands that a gorilla showed up in our sandbox,
we just need to train him to be the friendly giant he wants to be, and
get him do some of our heavy lifting.
Chris
Quoting Mike Liebhold <mnl at well.com>:
> Thanks for your excellent thoughts, Anselm.
>
> I, for one, am profoundly disappointed that just as the Mapping Hacks
> book is released, a milestone in geo hacking, Google has stolen the
> energy and enthusiasm for grass roots mapping. Instead of trying out
> some amazing open hacks, the Kiddiez are all scraping google
> javascripts.
>
> I'm beginning to think that Google might do best, to embrace open
> mapping and release themselves from servitude to their vendors,
> Navteq,
> Teleatlas, et. al. google is constrained, -and- constraining open
> development by their dependence on these outside private geodata
> suppliers.
>
> Maybe Google can be convinced to 'do no evil'
>
> how? by agregating, and combining user created google map hack layers
> so
> that the Google user hacks and data are more important than the
> Navteq,
> and teleatlas base layers. Google could continue to ride the growing
> wave of map hacking energy by enoouraging value of -agregated- user
> hacks 'n apps, so they could become less dependent on private data.
> google could crawl for google map hacks, and build tools to combine
> the
> user data -minus proprietary base layers.
>
> Given, Navteq and Teleatlas data IS more accurate and 'prettier' than
> TIGER, and Open Streetmap data, But as Rich Gibson suggests, it
> would
> be cool if they added a Tiger layer and allowed free-er access to
> that.
> I'd go further and suggest that just as Google actively supports the
> Mozilla foundation they should actively support openstreetmap, and
> World Wind WMS/WFS/GML and other related projects too.
>
> Somehow, google has to realize that there's lots to gain by
> abandoning
> Navteq, and Teleatlas dependencies, and instead actively supporting
> export google tiles into WMS servers and Worldwind, actively import
> and
> agregate Openstreetmap and other user created data projects into
> google
> maps, and and abandon Keyhole's proprietary tactics driving KML as
> 'a
> standard' and instead actively support OGC standards especially
> supporting native WMS/WFS/GML projects too.
>
> If google really is right on the edge between open mapping and
> vendors,
> it shouldn't take much to pull them over. We should do what we can to
> win Google over instead of leaving open mappers futiley trying to
> compete, with the 'bulldozer in the sandbox'
>
> - mike
>
>
> Anselm Hook wrote:
> > There are still good opportunities for the open source community:
> >
> > 1) Volatility. There seems to be a trend towards pursuing a
> > 'volatility edge'. Services can compete by capturing events that
> are
> > more and more up to the moment. Google was in itself a large
> > improvement on AltaVista for example. One can imagine open source
> > projects that capture the transient flow of human activity as it
> > happens. There are privacy concerns of course but a lot of the
> most
> > interesting and most relevant information is that which is most
> > timely. Jim Fournier deserves some credit for this encapsulation
> -
> > but it seems like there is almost a law at work here - a data
> version
> > of Moores law; that the value of data decreases exponentially with
> > time.
> >
> > Since Google tends to pursue the larger and more static data
> > aggregation - there does seem to be room for innovation here.
> >
> > 2) Insider knowledge. Some people have argued that Google tends
> to
> > study systems from the 'outside'; in an almost autistic manner -
> not
> > realizing that they are a part of the same system themselves.
> >
> > This is probably just a legacy of their initial attack on the
> problem;
> > clearly there are a lot of bright minds there and they're just as
> > likely to build a really great community service and collect
> > information from that as to do brute force aggregation and
> analysis.
> > They can afford to take all paths simultaneously and select the
> best
> > of all outcomes, or can be second to the table and still dominate.
> >
> > At the same time however they have yet to favor social knowledge.
> > Dodgeball and their blogging application (whatever its name is I
> > forget) are not contributing to their other projects; and seem to
> be
> > kind of lost in the muddle at the moment. For example the
> blogging
> > application doesn't help refine google search, and the dodgeball
> > application doesn't help refine google maps. One can easily
> imagine
> > the kind of say marketing analysis that a blogging community could
> > yield for example, or the kind of highly refined hotspot list a
> > location game could generate.
> >
> > I've actually found delicious outperforming google for some simple
> > single word queries. For example I was searching for 'Taekwondo'
> and
> > found some instruction on the kata's faster and better on Del than
> > with Google. So that is at least one example where a slightly
> > intelligent social tool running on basically toy hardware by
> > comparison physically outperformed a massively huge brute force
> > aggregation and analysis strategy. It is instructive.
> >
> > There was a classic Brittanica versus Wikipedia moment at the
> Where
> > 2.0 conference last month. During the 'Driving the Mean Streets'
> > presentation by Robert Denaro of NavTeq the comment was made: "We
> have
> > invested 100k man hours in the last year to collect street data
> and
> > this is a significant cost that we have to recoup". This is point
> > where open source developers can leverage a better understanding.
> If
> > NavTek had invented the Web instead you might imagine the same
> > executive saying 'How are we going to create all of the web pages
> on
> > the internet for everybody to read? Can you imagine the cost!'.
> >
> > There may be some momentary opportunity here therefore for open
> source
> > developers to work more closely within the highly social
> communities
> > that they typically inhabit and understand so well.
> >
> > Also it is speculatively arguable that _real_ social knowledge and
> > real social awareness is to some degree 'anti-commercial'. It
> cannot
> > be commercialized and leveraged because by definition is is that
> > knowledge which a community uses to speak about itself. It can be
> > highly critical of commercial interests and or can lose its value
> when
> > not spoken by a genuine and respected participant in that
> community.
> >
> > 3) Domain Interests. Google Maps specifically doesn't yet speak
> to
> > domain interests such as archeology or geology. There is a
> wonderful
> > book called "In Search of Ancient Oregon" (
> > http://bookswelike.net/isbn/088192590X ) that I would love to
> > translate into an application; to actually dynamically explore the
> > movement of the continental plates such as that of North America
> from
> > Triassic to present.
> >
> > A geo-annotated 'guns, germs and steel' also would be wonderful to
> > explore the ebb and flow of native cultures around the world - to
> > develop an intuition about these events.
> >
> > The grass-roots social cartography layer that I think most of us
> are
> > excited by can be considered just a domain interest. Clearly
> there
> > are a lot of domain interests; and Google Maps only accelerates
> those
> > interests rather than inhibits them.
> >
> > I have other thoughts re this but have to get back to hacking in
> any
> case.
> >
> > - a
> >
> > On 7/11/05, Brian Lalor <blalor at bravo5.org> wrote:
> >
> >>On Jun 15, 2005, at 3:52 PM, Mike Liebhold wrote:
> >>
> >>
> >>>>By the way, I guess I missed why big business is ruining open
> >>>>source GIS (sic).
> >>>
> >>>This is a delicate moment, just before the boom. It's possible
> that
> >>>enthusiam for google maps could dilute attention from lots of
> other
> >>>important grassroots and open mapping projects just gaining
> >>>critical momentum, that could really benefit from the kind user
> >>>contributed code and hacks that people are doing with google
> maps.
> >>
> >>I would argue that this is a good chance for those other projects
> to
> >>learn from Google and make their products as easy to integrate as
> >>Google has done! I don't look forward to the day when Google
> starts
> >>embedding ads and whatnot into their maps, but I love that I can
> >>*easily* (easily being the key word here) visualize all kinds of
> >>information in a browser. I'd be just as happy if I could do it
> >>*easily* with Tiger data. Google has provided a dead-simple API
> and
> >>done so SO much of the hard work for me.
> >>
> >>The definition of "easily" is of course dependent on the person,
> but
> >>for someone with a limited knowledgebase of geospatial
> mathematics,
> >>it's just so durn cool that I can stick an icon on a map at my
> >>address with 5 lines of JavaScript code.
> >>
> >>--
> >> __ ____
> >> / / / __/ Brian Lalor "If you still have gas, you're
> >>not lost."
> >> / _ \/__ \ blalor at bravo5.org -- Jacques Strappe
> >>/_.__/____/ http://bravo5.org/
> >>
> >>
> >>_______________________________________________
> >>Geowanking mailing list
> >>Geowanking at lists.burri.to
> >>http://lists.burri.to/mailman/listinfo/geowanking
> >>
> >
> > _______________________________________________
> > Geowanking mailing list
> > Geowanking at lists.burri.to
> > http://lists.burri.to/mailman/listinfo/geowanking
> >
> >
> _______________________________________________
> Geowanking mailing list
> Geowanking at lists.burri.to
> http://lists.burri.to/mailman/listinfo/geowanking
>
----------------------------------------------------------
This mail sent through IMP: https://webmail.limegroup.com/
More information about the Geowanking
mailing list