[Talk-us-massachusetts] Why Barnstable is such a mess, and other MADness (Wayne Emerson, Jr.)

[Greg Troxel]

Yury Yatsynovich <yury.yatsynovich at gmail.com> writes:

> Not sure. Western Mass doesn't have counties as admin_level 6 because, as
> the note says, that "there is effectively no county-level government here".

That is bogus, and I consider the deletion vandalism.  It may be true
that the county government is empty, but counties are still a thing, and
people know what county they are in, and it matters.  You have to put
down what county you live in on all sorts of government forms, including
searching for health insurance on the connector.  When you cross town
lines the signs say what county you are in, at least when changing
counties.  The Army cares about counties and has a program of getting
information about disaster status per county, via the Military Auxiliary
Radio System.  So I think Franklin and Hampshire county still matter and
should have a level6 (because that's how we spell county in osm :-) on
the map.

And agreed that the named villages of many towns do not have separate
governments.  Nor do wards and precincts, which is the other thing
people say belong in 9 and 10.

So if county deletionism prevails, we should not have level9.  But
really we should have counties.

> My suggestion would be to include addr:suburb for the cases when two or
> more identical addresses exist within a city

I strongly object to addr:suburb.  There is no established usage of
anything named a suburb in Massaschusetts at all.  It does not appear in
addresses.  We should not be trying to grab random notions from OSM that
map to the way addresses are in other parts of the world and use them as
part of fitting a database model to the osm model.   So far, I haven't
seen an address that is more than just a word in the town slot -- partly
because everything in the US expects names to be

  house-number road-name
  town-name, state-name  zip-code

so addresses that don't fit  that pattern are too awkward and aren't
assigned.


>>>> 2. While checking the ???Street Name Mismatches??? is there anything we
> should do on the google spreadsheet if it seems like the MAD name is wrong?
>
> I was recording what is written on the street sign (column "What is written
> on street signs?") and, if OSM is wrong, correcting the street names in
> OSM. The cases when MAD is wrong can be corrected automatically before the
> import

If MAD is wrong it would be helpful to talk to the town officials that
send data to MAD, separately from the process of mapping known MAD
erroneous names to the correct names (for variances in spelling and
spacing and other minor stuff where it's really really obvious that it's
the same name, munged).

> Indeed, it also happens that different spelling can be seen on different
> street signs along the same street within the same town. In such cases I
> add a note to the Google spreadsheet that both names are possible.
> It would be great, if we had lists with official street names to
> resolve/automatically correct such spelling/punctuation discrepancies.

MAD is an official list of names; it came from the towns.  It's just
that official lists have errors.

>>>> 4. Were the street names in MAD originally in all caps or should we
> notify the local assessor???s office if the capitalization is wrong in the
> google spreadsheet? i.e.: "J a McDermott Circle" or "Oconnor Drive"
>
> Wrong capitalization of names in the Google Spreadsheet (like "Oconnor
> Drive") is a result of my code's' imperfection -- I've been recording such
> cases and adding to the dictionary in the code to improve it. But, true,
> absence of, say, " ' "- sign (shouldn't it be "O'CONNOR DRIVE" in the
> capital case instead of "OCONNOR DRIVE"?) might be worth reporting. But,
> again, how do we know that the official street name is not "OCONNOR DRIVE"?

I think this is tough, and it's a really interesting question if street
names are upper case only or if they are mixed.  Pretty obviously MAD
thinks they are upper.   I don't know what the towns think when they
assign them.


>>>> 5. I have also noticed other weirdness such as where the Humarock
>>>> section of Scituate uses Marshfield???s Zip code. Do Zip codes matter for
>>>> the MAD import?

zip codes are a post office thing that organizes delivering mail, and no
more.  It is perfectly ok for a street in one town to have the zip code
of an adjoinging town.  But the town name should match the town it is
in.

> So far I'm keeping zip codes in the files for import, but not using them
> for any matching and, as wiki says, we are probably not allowed to import
> zip codes.

I don't follow "not allowed".  We have permission from MassGIS - are you
saying that you think MassGIS doesn't really have rights to distribute
zip code information?  Or that OSM shouldn't have zip codes?  Or ???