[Talk-ca] Forests/Land Use, was: Canvec reverts

Paul Ramsey pramsey at cleverelephant.ca
Thu Sep 1 20:22:57 UTC 2016


On Thu, Sep 1, 2016 at 11:27 AM, Adam Martin <s.adam.martin at gmail.com>
wrote:

> That is the key here. Deleting information without replacing it with
> something more accurate is inherently destructive. There must be some
> thought as to what will be put back or one is essentially ripping the map
> up simply because you don't like how something looks or how it closely it
> follows a given rule.
>
I'm not sure I agree. "Better than nothing" I guess is the principle, but
when what is there (not nothing) gets in the way of improving other
features, then it's not better than nothing. And what if what's there is,
from an information point of view, basically nothing?

Like the forests polygons that basically do nothing to delineate where
forests actually are (or residential polygons with same issue?) "Go map all
forests" is not actionable. Hell, even "clean up all forests in just the
area you care about" isn't. There's too much. So instead, I leave
demonstrably wrong "forests" in place.

[image: Inline image 1]

I can't even salve my conscience that they at least improve the rendering a
little at an aesthetic (if not informational) level, since they were
partially loaded in the region, and actually make it look worse.

[image: Inline image 2]

Anyways, I stick to my general feeling (un-acted upon) that more is not
better, and the map would be easier to work with without the big,
unhelpful, land cover polygons.

P.



> That would be like finding parking aisles tagged as drive throughs and
> deleting them as incorrect, instead of simply correcting the tags.
>
> On Sep 1, 2016 3:30 PM, "john whelan" <jwhelan0112 at gmail.com> wrote:
>
>> And as someone who has deleted quite a few things in OSM I would agree
>> with that statement.  When I didn't have a better replacement available
>> then I prefer not to delete unless I have done a ground level inspection
>> and there really isn't anything there.
>>
>> I think my favourite was a mapper who was demonstrating 3D software with
>> OSM.  They dropped in a group of multiple level buildings into an area I
>> was mapping in Africa.  They didn't consider what they did was wrong, it
>> was only Africa.
>>
>> Cheerio John
>>
>> On 1 Sep 2016 1:26 pm, "Begin Daniel" <jfd553 at hotmail.com> wrote:
>>
>>> *P: OSM is very much an "add only" project, since the social
>>> consequences of incorrectly deleting things seem so high.*
>>>
>>>
>>>
>>> What I do perceive in the current thread is that deleting something not
>>> perfect without replacing it with something better hurts, not that it is
>>> not acceptable to delete something.
>>>
>>>
>>>
>>> Daniel
>>>
>>>
>>>
>>> *From:* Paul Ramsey [mailto:pramsey at cleverelephant.ca]
>>> *Sent:* Thursday, 1 September, 2016 13:05
>>> *To:* Begin Daniel
>>> *Cc:* Talk-CA OpenStreetMap
>>> *Subject:* Re: [Talk-ca] Forests/Land Use, was: Canvec reverts
>>>
>>>
>>>
>>>
>>>
>>> On Thu, Sep 1, 2016 at 8:49 AM, Begin Daniel <jfd553 at hotmail.com> wrote:
>>>
>>> What is very cool with OSM is that you can edit the data. Urban polygon
>>> is wrong? Modify it! The definition is obscure in the Wiki? Change it! But
>>> yes, the learning curve is often steep, and you may need to discuss with
>>> someone else…
>>>
>>>
>>>
>>> "Just fix it" is not quite the answer. The point the original poster
>>> made, which I concur with, is that the very existence of these shapes makes
>>> working with the "important" data difficult. In terms of forest and land
>>> use polygons, every vertex I move there is a vertex I'm not going to move
>>> on something "important".  (And the vertex density of the forests/land use
>>> are another reason that working around/with them is painful and
>>> energy-sapping.)
>>>
>>>
>>>
>>> As discussed in the other thread, the shear volume of Canada means I'm
>>> never in 1M years going to "fix" the forests. As it stands, I mostly ignore
>>> them. Too many vertexes to move, for too little net benefit, so there's
>>> forests running through the new subdivisions of Prince George. At least the
>>> roads are there and hopefully correctly named now.
>>>
>>>
>>>
>>>  (I would, however, love to just delete the urban "land use" polygons,
>>> but who know if that's "allowed" or not. Absent a strong personality like
>>> the person who caused this thread, it seems like OSM is very much an "add
>>> only" project, since the social consequences of incorrectly deleting things
>>> seem so high. Nobody wants to be "that guy".)
>>>
>>>
>>>
>>> ATB,
>>>
>>>
>>> P
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> *From:* Paul Ramsey [mailto:pramsey at cleverelephant.ca]
>>> *Sent:* Thursday, 1 September, 2016 11:17
>>> *To:* Talk-CA OpenStreetMap
>>> *Subject:* [Talk-ca] Forests/Land Use, was: Canvec reverts
>>>
>>>
>>>
>>> I'm "glad" to see someone else w/ this issue. It's glancingly related to
>>> the canvec import issue, since the land use polygons are a source of some
>>> of the issues the reverter is complaining about (malformed multipolygons /
>>> boundary overlaps).
>>>
>>>
>>>
>>> In my own work in my old home town of Prince George, I've constantly
>>> wanted to just plain delete the "urban area" land use polygon (which
>>> doesn't seem to correspond in any way to the actual urban area of the
>>> present) and the forest polygons (which have the same problem).
>>>
>>>
>>>
>>> Unlike buildings and roads and water, land use is pretty sloppy: where
>>> does the "urban area" end? Is this a "forest" or just a bunch of trees?
>>> Since anyone making a real multi-scale map will fine some other source of
>>> land-use (like classified landsat) and since people trying to map at
>>> high-res are finding the forests add little value and much impedance, why
>>> don't we ... burn down all the forests (and the urban areas too)?
>>>
>>>
>>>
>>> P
>>>
>>>
>>>
>>> On Thu, Sep 1, 2016 at 6:54 AM, Loïc Haméon <hameonl at gmail.com> wrote:
>>>
>>>
>>> On a final note, though, I certainly would approve of any effort to
>>> reduce the size of the upload chunks and the assorted polygons. For new
>>> mappers like me, those create daunting challenges when trying to make
>>> incremental improvements to an area. Shortly after joining the OSM
>>> community I was back in my home town of Saint-Félicien, in a fairly remote
>>> region that hasn't had tons of local mapping done. Some of the inhabited
>>> areas I aimed to improve were covered by Canvec forest multipolygons, and I
>>> ended up giving up on them until I could get some more experience as I
>>> absolutely did not understand what the hell was going on....
>>>
>>>
>>>
>>> _______________________________________________
>>> Talk-ca mailing list
>>> Talk-ca at openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/talk-ca
>>>
>>>
>> _______________________________________________
>> Talk-ca mailing list
>> Talk-ca at openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk-ca
>>
>>
> _______________________________________________
> Talk-ca mailing list
> Talk-ca at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-ca
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-ca/attachments/20160901/79646011/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: screenshot_341.jpg
Type: image/jpeg
Size: 34980 bytes
Desc: not available
URL: <http://lists.openstreetmap.org/pipermail/talk-ca/attachments/20160901/79646011/attachment-0002.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: screenshot_340.jpg
Type: image/jpeg
Size: 42862 bytes
Desc: not available
URL: <http://lists.openstreetmap.org/pipermail/talk-ca/attachments/20160901/79646011/attachment-0003.jpg>


More information about the Talk-ca mailing list