[OAM-talk] Fw: RE: Re: [Geodata] [crschmidt at metacarta.com: OAM Prototype]

[crschmidt at metacarta.com]

This is a response to an email from Jo Walsh, who may not be on the OAM-talk
list. Please direct your responses appropriately.

Jo, 

I've removed the email from Antonio at IGN (for his privacy), and 
forwarded my response to the OAM-talk list, as I believe it is generally
informative of my viewpoints on the project.

On Sun, Dec 02, 2007 at 05:10:25AM -0800, Jo Walsh wrote:
> i dropped your OAM notice to Antonio at IGN - he's heading up the team
> that is getting 50cm imagery for the whole of Spain for (NC) re-use
> next year - he responds with interest and with licensing questions
> and also performance questions - [[PNOA ortophotos covering all Spain
> implies 4 Tb in ecw format and a hugh cache to have a efficient WMS-C,
> let's say at least more than 1Tb.]] 

We're not yet prepared to host that level of imagery. It's in the plan,
but there's a lot of work to be done in building up the toolset and
pipeline before it will be ready. That said, one of the very near term
tasks for OAM is implementing the ability to load data into the map from
a remote WMS. There are some technical limitations here, but they're on
my todo list to fix.

(For context, the original statement of the question was:

> - Under what conditions are data published in AOM? Or in other words,
> there is some mechanism to allow each agent to give information about
> the usage conditions of their data?
)

The question as to whether there will be different restrictions
available for different sources is that yes, I have a plan to do that.
Specifically, I expect that OAM will, at some point in the future,
provide the ability to create your own 'view' of the data, choosing
things based on resolution, license, age, quality, etc. Once you create
your own view, you would then be able to use this view in your own
applications (either as a WMS, or a cached WMS, or something different).
However, this does not exist yet, and it will take time to build. The
most important thing from my opinion is to make it possible for users of
OAM to use the data with *the fewest number of restrictions placed on
it* as possible: which means that if the data is originally public
domain, you should have the technical ability to use it under no
restrictions, whereas if you're willing to accept attribution, you can
do that, etc. Building one coherent view of the world doesn't mean that
it should be the *only* view of the world, and the ability to choose
which layers you're interested in is a key difference between a service
like OAM and a service like Google Maps.

I'd like to state that I would encourage *anyone* donating imagery to do
so under a license that is *not* limited by anything other than
attribution under the current licensing environment. There is some work
being done to resolve this situation, but (so long as I have a voice in
the project) the main OAM map will *not* contain any NC licensed data.
The terms of the CC license with regard to non-commercial licensing are
simply too broad (or at the very least, poorly established) such that
each person considers the meaning of NC different -- and none of them
allow me to, for example, use the imagery on my website, which is
partially ad supported. Similarly, the terms of use with regards to -SA
have what may be painful ramifications for mapping: A CC-By-SA license
may mean that, for example, the data can't be used to create information
for wikipedia, becuase CC-By-SA is incompatible with the GFDL.

In general, I believe that there are only two usable licenses for
geodata available at the moment:
 
 * CC-Attribution license
 * Something along the lines of the "WFTPL":
   http://sam.zoy.org/wtfpl/COPYING
   "0. You just DO WHAT THE FUCK YOU WANT TO."

Anything else is likely to suffer the law of unintended consequences, as
users of the data who you would like to encourage find the licensing
terms overly restrictive for their particular use case. 

Additionally, there was a quesiton posed about OAM being a 'view only
application', and I can state that although this is the case currently,
I have already received questions about exporting data for use offline,
and the concepts for that have been percolating in my head. It is my
expectation that there will be many datasources for which the holder of
the data is unwilling to have the source data downloaded, but is willing
to share the resulting imagery or reprocessed imagery resulting from
that. I'm not sure if this expectations is unfounded or not. In any
case, I have an expectation that there will, at some point in the
future, be a way to build up a multi resolution database and export it
as a file usable in 'local' gis applications. 

Lastly, there was a question posed about whether OAM is really
going to be able to perform well. I'll state that currently, the machine
that OAM is running on is sitting with lots of idle CPU, but there are
other hardware limitations -- specifically, disk and network -- which we
are working to resolve. I had no expectations that OAM would be 'open
for business' as quickly as it was, and it has suffered some growing
pains as a result of being designed for rapid development rather than
for deployment. However, it is my belief that in the long term, by a
combination of the incredible work put forth by telascience in causing
more resources to magically appear and Murphy's law, we will be able to
create a service which will scale to the users it gets.   

Thanks for the information, and looking forward to seeing more.

Regards,
-- 
Christopher Schmidt
Web Developer