[talk-ch] vandalism under the pretense of "simplifying"

RB tanrub at gmail.com
Fri Jul 1 08:43:44 CEST 2022


This discussion has in a large way gone both out of topic thematically and
somewhat sour in the atmospherics. Let me take the time to reply to the
various contributors in a lengthy but hopefully clear way.

I will firstly expose what constitutes a constructive edit as well as some
foreword about what I understand as a healthy discussion. Secondly, the
main arguments against forest micromapping as well as the result of my
reflection on these arguments will be outlined. One of these arguments will
most likely change the way I map forests and should probably be used as a
good practice for the wiki. Thirdly, I will refocus on the main topic and
expose the problem of the destructive edits. Finally, I will propose a
metaphor of my own way to see the whole thing.

*Constructive edits and constructive discussions on OSM*
A constructive OSM edit is an edit that brings the data closer to the
physical world on the base of sources compliant with the licence. Bringing
more accuracy with "illegal" sources isn't a constructive edit just as
mapping false features. In addition "spam" exists when either untrue or
unnecessary data is added, such as an additional node in a straight way for
instance. I will elaborate more on this particular point later as I feel
that there is ground for improvement on my side on this particular topic.

Regarding the atmospherics, this discussion has triggered expressions such
as "give up" or "frustrating", which I deplore. I assume that all the
participants of the mailing list and most OSM contributors act and speak in
good faith. That being said, prescriptive orders or semi "threats" from
individuals taking upon themselves to speak in the name of the community
sound very much like a mexican standoff. I am afraid to say that those who
confuse a genuine OSM contributor with a Western criminal are in the wrong
film. I like Sergio Leone movies but we are amongst volunteer contributors
using their free time for the sake of an open source project. I have never
destroyed good data, duplicated data nor used non compliant sources. The
tone of the discussion seems therefore at times inappropriate considering
the context. To which I confess I tend to escalate further, which isn't
smart.

Let's review some of the arguments against forest micromapping. I will
regroup them but I hope to cover all of them.

*The "data overload" argument*
This is an argument to protect the server's load but I believe that no one
should map for the server. However, there is a gray zone, on which I will
elaborate on the "node density exceeding the imagery" paragraph.

*The psychological / motivational arguments*
No one knows why other people do stuff. Why do some users map dog waste
baskets and others brothels, police stations or individual trees? No one
knows. I suspect that a large part of the reason is unknown to the users
themselves. However, making hypotheses on someone's motivation on a public
mailing list isn't only irrelevant, it inevitably leads the discussion
towards straw men and toxic atmospherics. The topic is certainly
interesting for an academic paper in psychology but irrelevant to this
discussion and most of the time toxic.

*The "wrong mapping according to the newer imagery" argument*
It'is the life of most OSM data : it gets improved with new sources. Most
buildings can be redrawn with newer imagery. 3D people might redraw them
using photogrammetry. The continental shift is, in a very small way,
putting a time limit on all the data. All these things are true and at no
time am I trying to suggest that all the detailed forest should be
considered as sacred or definitive. I hope that it is clear: as with the
rest of OSM, forests should be corrected and improved with newer sources.
It isn't specific to the forest, though.

*The "node density exceeding the imagery resolution" argument*
It probably comes from Sarah and it did trigger some thinking on my side. *I
suspect that, as I did some of these edits mechanically while thinking
about something else, I didn't always consider the level of zoom in JOSM.
It is likely that we find ourselves close to the "involuntary spam" gray
zone as the same level of accuracy regarding the imagery might be
achievable with less nodes in some cases. I will certainly pay attention to
that in the future.*

*The "changing nature of the forest" argument*
Very true, the vegetation changes. The same applies to all data. Buildings
are constructed and demolished. Businesses open and close but are mapped
anyway. While I don't like to speak about motivation, I suspect that it
might be interesting to study the evolution of vegetation over time with
vector data in the future.

*The "unclear limit of the forest" argument*
It is a fact that there is no clear way to decide what constitutes the
exact limit of a forest. But again, the same applies to all the buildings
(I suspect that many are mapped according to the roof and not the walls)
and almost all OSM data. Rivers
<https://www.openstreetmap.org/relation/945049> are often mapped with both
an (imprecise ) area as well as an (imprecise) way. Towns get mapped with
both a node for the center of the town as well as an area for the
residential area of the administrative boundary. The same logic could be
applied to trees (area for the leaf cover and node for the trunc, which
would be more precise that the center of a village or a large river),
although I have no intention of doing so. This topic, as almost all
features, is the consequence of the mapper assessment, the mapper time
dedication, hopefully some general good practices as well as the sources
used.

I hope that it summarized in an honest and comprehensive way the arguments
of the discussion. It does as I have understood them at least.

*The destructive edit problem*
Now to go back to the initial topic. It certainly wouldn't be a problem to
simplify what I have called "involuntary spam" provided something like that
is actually there. Removing nodes exceeding the image resolution is
desirable and I am not talking about that when I discuss this particular
case of (probably well intentioned) vandalism. Using automated tools to
"simplify" creates forest where there are no forests, sharp angles where
the forest is smooth. This is what I call a destructive edit (it regresses
the data away from the physical world) and constitutes the initial concern
triggering the discussion. It is one thing to map according to one's own
vision of what OSM should and shouldn't be. It is a complete other thing to
actively degrade the existing data according to this subjective vision. I
still don't know how to address that problem as the discussion stayed most
of the time out of topic.

*The metaphor*
The reason why this discussion has also been frustrating from my point of
view is the following.
It's like a person voluntarily cleaning a street and removing litter to
improve the neighborhood. This person might be a weirdo as she uses a
toothbrush instead of a broom to clean particular places in the street. Now
comes another person dumping garbage on the street. The original volunteer
addresses the problem to the municipal council and the discussion focuses
only on how bad it is to use a toothbrush instead of a broom. Some members
of the council also argue that it is customary to clean the street from
east to west and not the other way around while another municipal
councillor issues threats against the volunteer cleaner.

As all metaphors, It is debatable but it does at the very least outlines my
own subjective perception of the discussion.

Let me reiterate my hopes that everybody feels understood as well as the
hopes that some will sincerely consider my arguments as I considered theirs.

Regards.

Le jeu. 30 juin 2022 à 18:30, RB <tanrub at gmail.com> a écrit :

> I am sorry but I don't agree with their arguments and I have
> addressed them before. Let me restate.
>
> Sarah makes the argument that the forest changes. Of course it does. I
> think that precise mapping in this case is a feature not a bug as it will
> in the future allow us to study the evolution of the vegetation.
>
> Mark makes the argument that some of the precise forests are "wrong" and
> again, in some cases, he is right. I am afraid it applies to the complete
> set of OSM data. At some point, it should be deleted and corrected with
> better sources.
>
> Both these statements are valid but not as a counter argument to the
> precise mapping in general.
>
> I do however agree that at times, the mapping, below let's say one meter,
> doesn't make sense as the source doesn't have this accuracy. It is probably
> the only objective point in the discussion and I will pay attention to the
> zoom levels in JOSM.
>
> "4) On the appropriate pages in the Wiki, we insert the points made by
> Sarah and Marc and lay down that it is not appropriate to map forest
> limits with more than one node every 10 m (or so). This in order to prevent
> that mappers in a “gamification-mood” come to the idea to artificially
> create an excessive number of nodes with the main purpose of improving
> their ranking in the contributor’s statistics (@Pan: Maybe you did notice
> that you are currently ranked #1 in Switzerland’s OSMstats?)."
>
> This is again offending.
>
> 1. It is not my motivation
> 2. The motivation behind people mapping is totally irrelevant (why do
> people map camping places instead of water streams?
> 3. If you look at this page, you will see that I have mapped a lot of
> other features. A lot.
>
>
>
>
>
> Le jeu. 30 juin 2022 à 18:12, Michael Flamm <michael.flamm at micoda.ch> a
> écrit :
>
>> Hello RB/Pan,
>>
>> To be honest, debating with you Is a bit frustrating because you seem to
>> obviously ignore the strong points made by Sarah & Marc, and furthermore
>> you don’t answer the precise questions that were directed to you.
>>
>> As I see it, Sarah and Marc have demonstrated that your over-use of nodes
>> is simply wrong from a conceptual point of view. Thus, we are not anymore
>> in the realm of “tastes and colors”, as you seem to understand it. Frankly,
>> I hope you will acknowledge that because, if not, this thread will not come
>> to a constructive conclusion.
>>
>> My “Swiss Compromise”-Proposal to get there would be:
>>
>> 1) The community acknowledges your effort in detailing landuse=forest
>> mapping on a quite impressive area.
>>
>> 2) On your side, you acknowledge that you created landuse cover objects
>> with an excessive number of nodes and you adapt your future mapping
>> activity in accordance.
>>
>> 3) The  community does not start a massive effort to simplify the objects
>> you created, because there are lots of other things where mapping energy
>> can be directed at, bringing much more added value to the map.
>>
>> 4) On the appropriate pages in the Wiki, we insert the points made by
>> Sarah and Marc and lay down that it is not appropriate to map forest limits
>> with more than one node every 10 m (or so). This in order to prevent that
>> mappers in a “gamification-mood” come to the idea to artificially create an
>> excessive number of nodes with the main purpose of improving their ranking
>> in the contributor’s statistics (@Pan: Maybe you did notice that you are
>> currently ranked #1 in Switzerland’s OSMstats?).
>>
>> What do you think about these proposals?
>>
>> Michael
>> _______________________________________________
>> talk-ch mailing list
>> talk-ch at openstreetmap.ch
>> http://lists.openstreetmap.ch/mailman/listinfo/talk-ch
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.ch/pipermail/talk-ch/attachments/20220701/a4bcb9ab/attachment-0001.htm>


More information about the talk-ch mailing list