[Imports] Import in progress: NYC DEP Watershed Recreation Areas

[Kevin Kenny]

I just realized on further inspection that 'protect_class' on the
protected areas is erroneously spelt out 'protection_class'. I think
that this proves that no matter how careful one is, at least one bug
slips past when a system 'goes live.'

I'll correct the typo in the other 39 townships, and correct it
manually in the 25 places it appears in this changeset. I'm going to
defer the correction in the upload until I'm resuming uploading. so as
to avoid complicating a revert should one prove necessary.


On Sat, Jun 4, 2016 at 11:00 PM, Kevin Kenny
<kevin.b.kenny+osm at gmail.com> wrote:
> I've now begun the import for the NYC DEP Watershed Recreation Areas.
>
> According to the plan, I've imported the units that lie in a single
> township out of the forty townships that contain these areas. (Town of
> Andes - the first of the list in alphabetical order). The changeset
> for this activity is https://www.openstreetmap.org/changeset/39807853
>
> JOSM uploaded the areas without errors or warnings.
>
> A quick visual check at least appears plausible. I notice several things:
>
> There are numerous cases where the parcel is clearly supposed to be at
> a fixed setback from a highway. Some of these do not align to the
> highway. In every case that I checked visually against Bing Aerial.
> New York City is right and TIGER is wrong. Virtually all the highways
> in the area are unreviewed TIGER ways.
>
> The alignment of the protected_area boundaries with the New York State
> lands is rather untidy. There has been considerably interagency
> coordination of these in recent years, and the current version of
> 'NYSDEP Lands' is in considerably better alignment with the New York
> City data. That misalignment is one of a number of reasons why I think
> that the 'NYSDEP Lands' file should be reimported, and not just have
> its tagging fixed by a mechanical edit. I'm still working on a plan
> for this. I think I may be close to a breakthrough on techniques for
> semiautomatic repair of the sort of mildly inconsistent cadastral data
> that we have here.
>
> According to the plan, I'll now stop for a week, so that if there are
> any screams of horror, there will be only a single changeset,
> consisting only of added objects, to be reverted.