[Talk-us] MassGIS import: "condition" tag

Serge Wroclawski emacsen at gmail.com
Sat Jul 16 04:44:44 BST 2011


On Fri, Jul 15, 2011 at 9:13 PM, Greg Troxel <gdt at ir.bbn.com> wrote:
>

> I usually think that OSM needs to adopt terms/standards of existing
> professional communities rather than make up our own.   I am not clear
> on road condition standards, but clearly various government road
> maintenance bodies must be.

I disagree with you, and here's why (in no particular order):

1. OSM tags are de-faco international standards

One of the benefits of the OSM dataset is that tags are meaningful
across borders. A tag means the same in multiple countries. If you
start insisting on regionally defined tags, you get rid of this
benefit, which means you make OSM no more useful than local datasets.

2. Those proposing these tagging "standards" have, as of yet, not
proposed any mechanism for proper boundary handling.

If people wanted to move away from the OSM tagging system to some
national tagging system, we'd need very comprehensive tool support for
renderers, routing engines, etc. But as of yet, I've not seen any of
these US centric proposals come with tools that would actually handle
these problems in a comprehensive way. So they'd break what we have,
and not add anything back.

3. The OSM tagging system allows for tags to exist simultaneously.

If some external organization's data is being put in OSM, their labels
can be turned into OSM tags via a prefix. So if there's a need to
store the previous dataset's condition tag, it can be stored as
usdot:condition.

4. If the OSM tags aren't sufficient, fix them

I the tags aren't useful, then it's up to the community to fix the tags

5. Having "official" tags like requires that OSMers follow strict,
external requirements

This is where I think the rubber meets the road in these discussions.
Most of the folks recommending these kinds of strict tagging
definitions come from the professional GIS community. There, strict
labeling guidelines exist. But those are enforced (as much as they are
actually enforced) by external incentives (ie people pay them).

The OSM community has and I feel should remain a grass roots project,
with the largest contribution from individuals. If you aren't paying
people, then you need to motivate them some other way, and you need to
keep the barrier of entry low. Strict, complex guidelines don't do
either.

6. It gives the impression that the external datasets are better

After years in OSM, seeing other datasets which cost millions of
dollars to produce and comparing the data myself with on the ground
surveying, I've come to realize that being official doesn't speak to
actual data quality. If you look at data in the UK, for example,
analysis is showing that OSM is actually more accurate them Ordinance
Survey data. In DC, I found official datasets with incorrect
information, and was able to correct it in OSM.

If we rely on external entities to provide data, we give the
impression that externally provided data is better, and it isn't.

7. It discourages mappers from mapping

For the reasons outlined in 5, 6, external datasets with strict tags
hinder mapping and encourage imports. This is not good for the
project. I won't elaborate on this because the topic has been
discussed ad nauseum on the lists in the past.

- Serge



More information about the Talk-us mailing list