[postgis-users] Interest in open-source road data

[Michael Graff]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

{ Moderator:  If this discussion takes too much bandwidth, please let
  me know and I'll create a list for it. }

"Thomas, Cord" <cthomas at rand.org> writes:

> 1 - Are there any open source community-based data gathering/improving
> projects out there that this could be modelled from or folded into.

That's part of why I asked the question.  The only open source data I
know of that is useful in a GPS/routing world is tiger, and that's
only just usable.

> 2 - Access - what kind of delivery mechanisms would there be - 
>   a - could you select subsets of data

Ideally.

>   b - what server(s) would be responsible for delivering this data

I can probably put a server up on an OC12 colo facility without much
issue.

>   c - what data formats would be available - postgresql dumps, shapefiles,
> ascii, etc.

To be decided.

> 3 - Implementation
>   a - you could use some sort of common open source model as in what has
> been done with linux- someone maintains the metadata (aka kernel - a
> linus-person), while individuals maintain the integrity of types of data
> (one person maintains interstate highway data, etc.), individual geographic
> regions (one person maintains each state)

Oh this would have to be a hierarchy.

>   b - a need for an agreed upon data conflict resolution model (if an
> interstate moves, what happens to the local roads, and the other features,
> like water).
>   c - what happens when TIGER data is updated - where do we maintain the
> link between TIGER and TIGER-OS

That's something I've been thinking about, yes.

I think the short answer is that is if this sort of thing takes off,
the tiger data would be a subset of it, rather than the other way
around, so it may be a one-time import.

> 4 - Quality Assurance / Control
> Similar to the module compilation for linux, your data would be submitted
> for change by the public, and reviewed by the responsible person. 
>   - useful to have supporting documentation for each change - how did use
> come about to change the data
>   - useful to know if we are changing single sources more than once (are
> different people visiting a particular shape more than once) and if so, are
> subsequent visits using more accurate info
>   - is it proposed that spatial information could also be modified.  while
> TIGEr may have a scale-accuracy of 1:24,000 or whatever it is, would the
> goal be to eventually improve TIGER data using local orthophoto-derived
> landbase information, GPS surveys, etc.?

Quality needs to be high, but with an open source project, hopefully
that Just Happens.  A lot of data can be gathered without driving
roads -- for instance, number of lanes can often be seen from aerial
photographs and/or locally published map data.

> All these comments/questions aside, i think having a TIGER-derived database
> available from which one could select data (all roads from state of X, or
> all data from this town and all towns surrounding it) and have it made
> available on an FTP server would be very very useful.  As it is, grabbing
> data from TIGER and manipulating it everytime for small projects can eat up
> considerable time.

That may be a place to start.  I'd probably statically generate the
data for each region and only update it when needed.

- --Michael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (NetBSD)
Comment: See http://www.flame.org/~explorer/pgp for my keys

iD8DBQE9yFXYl6Nz7kJWYWYRAt0bAJ9omGlSKos6dePrNU05M2rR0+wPrACdHyst
1xOpQDLXXjoGTkztJ06Qk6M=
=5EWt
-----END PGP SIGNATURE-----