[Tagging] Feature Proposal - Voting - boundary=forestry(_compartment) relations

[Paul Allen]

On Thu, 18 Feb 2021 at 13:28, David Marchal via Tagging <
tagging at openstreetmap.org> wrote:

The main issues that currently prevent reaching a consensus are, according
> to the vote comments, the following:
>

Part of the problem may be that it was difficult to figure out what you were
trying to do.  In one place you made a clear distinction between physical
(here be trees) and logical (here is the boundary of managed forestry, not
all of which might actually have trees on it right now) and I was happy
with that.
Elsewhere it appeared to read as though you wanted to replace all physical
tagging with your new boundary, and lots of people (including me) would
object to that.

Your proposal is also too long.  Move the bits explaining how people
currently use physical tags and why that tagging is not suited to mapping
managed forestry to a sub-page.  Otherwise it contributes to the
impression that your proposal completely replaces all of them.  Or
maybe you do intend to completely replace existing physical tagging,
in which case keep those bits in your proposal so people have more
reason to vote against it.

the proposal does not enoughly state how would the transition to the new
> scheme be done. I kept that out of the proposal, because I felt it was
> irrelevant, but it seems that this is necessary out of pedagogy. Given the
> complexity of the proposal and the wide impact it will have if approved,
> understanding it upon first reading is essential to allow voters who did
> not participate in preliminary debates to better understand what it is
> about, without having to dwelve in all the details; I'll then include some
> sort of transition digest, especially a table which will explain how to
> update the tagging in different situations, and a detailed argumentation of
> the design choices made in the proposal (why boundary=* and not landuse=*?
> why deprecating landuse=forest? why not a landuse=forestry?…);
>

And with that you make me worry that you really do intend to completely
replace existing current physical tagging.  If that is so, I object very
strongly
(and suspect a lot of others will also object strongly).  Wrapping a
boundary
around existing physical tags doesn't take much explaining; nor does
adding parcel boundaries.

As long as the logical tagging and physical tagging are orthogonal, mappers
can map what they see (trees, sometimes with clearings in them) and
other mappers can refine matters later by adding a boundary.

some voters are attached to landcover=trees. I was not able to understand
> this attachement during preliminary debates:
>

Again, you insisting on covering this at all in your proposal makes it
appear
that you wish to change the physical tags we use.  There will be a lot of
resistance to that idea, if that is what you really intend.  A proposal
that is
agnostic to the physical tagging would have far fewer objections.

As to why the attachment to landcover=trees, it is because it is
one way of dealing with problems of physically tagging the presence
of trees.  I think landuse=forest was originally intended to be a way
of indicating managed forestry but has come to mean both that and
also "here be trees" (it is also semantically incorrect and should
have been "landuse=forestry").  There is ambiguity in natural=wood:
does it mean woodland that has never been altered by the hand of man
or does it mean "living things that happen to be trees"?  Some people
see landcover=trees as doing exactly what it says on the tin: the land is
covered with trees, thereby avoiding the problems of the two other
ways of tagging an area of trees.

-- 
Paul
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20210218/89674ef4/attachment.htm>