[Talk-us] parcel boundaries and associated data in OSM

Serge Wroclawski emacsen at gmail.com
Fri Feb 15 18:36:55 GMT 2013


On Thu, Feb 14, 2013 at 3:29 PM, Brian Cavagnolo <bcavagnolo at gmail.com> wrote:

> We really want a nationwide consolidated, standard parcel database to
> build upon.

> This idea is [obviously] inspired by OSM.  And my immediate thought
> was, Fun!  Let's add parcel data to OSM!  How do we do that?

We've started an Import Committee to help with such questions. I need
to schedule the next meeting, but I invite you to join us and help
shape the conversation. I'm facilitating the committee but the
opinions I'll express below are my own.

> This
> inquiry has of course led to numerous more detailed questions, the
> most fundamental one, of course, being: Is parcel data welcome in OSM?

As you found out, this is a complex question that will depend on who you ask.

>  I've spent some time reading through the mailing list history.  In
> addition to gaining an appreciation for some of the issues regarding
> the management of parcel data, I promptly learned that this is a
> controversial question.  For each claim that a consensus exists
> against parcel data in OSM, a parcel data advocate seems to emerge.
> This leads to debate, which seems to focus on a specific set of issues
> that I have posed as specific questions below.  I've also dusted off
> and enriched the wiki page and associated talk page on the matter
> (http://wiki.openstreetmap.org/wiki/Parcel).  My hope is that people
> can respond to these questions and we can reach a clear consensus on
> {whether,what sort of,conditions under which} parcel data is welcome.
> And of course feel free to bring up any issues that are not
> represented in this list.

> Finally, even if you believe that parcel
> data does not belong in OSM, but that a nationwide open consolidated
> parcel database would be useful (and possible:) I'm super interested
> in this perspective.

I am of this view. Furthermore, I think that projects should Free
datasets intermixing this way, just as we do with topo data.

> Is parcel data useful to OSM?

This is actually a three part question.

1. Is the data useful?
2. Is the data useful to OSM users?
3. Does the data belong in the OSM core dataset?

In my opinion, the answer to question 1 is yes. The question to two
and three are more subtle. I think the data is of use to the OSM
project, but does not belong in the OSM dataset. What I mean by that
is that we have tools (renderers, geocoders, routers, etc.) which may
want to use parcel data. I think that such tools should be able to.
But I think the data belongs alongside the OSM dataset, rather than
part of it.

So to make this clear: I think the data is useful, but would be more
useful to OSM users if it's not part of the OSM crowdsourced dataset.

> Can parcel data possibly be kept up to date?

Parcel data (with very few exceptions) can't be manually surveyed by
amatuer mappers. Therefore it doesn't benefit from the OSM process of
survey, refinement, survey to provide additional detail and over-time
accuracy. Put in plain English "How can a regular person, with no
additional information, survey the area to find mistakes in the survey
data?" - the answer is that for the most part, they can't. Parcel data
is determined by a central authority.

So then if we had it in OSM's core crowdsourced database, we would
need a synchronization process. This is something many of us have
wanted, and worked on, for several years, and not come up with a
solution for.

But if we had the data as a database that could be integrated by
tools, then the data could be optionally rendered, used for geocoding,
used for routing, etc. in just the same way as OSM data, and OSM users
would get the benefit of current, up to date parcel data. That would
be a real win.


> Does parcel data meet the "on the ground" verifiability criteria?

I don't see how, but I'm open to being shown that I'm wrong.

> Can tools be adapted to accommodate parcel data density?

I don't understand this question.

To summarize, I think this is a great idea. I'm in total support. I'd
love to see the data available to OSM, but not part of OSM itself.

- Serge



More information about the Talk-us mailing list