I'd love to move into rational and studied discussion of corporate involvement
in OSM and the application of machine learning techniques.
It's easy to get caught up in rhetoric. I dislike "turbocharged" as much as I
dislike "exploitation". The entire application of machine learning is plagued
with overblown rhetoric, when after all, it is simply a statistical technique.
OpenStreetMap was founded on equal parts radical, reactionary rhetoric, and
JFDI. It's also easy to forget how much traditional map making rejected OSM --
that the craft of surveying is not something to be left to wild hooligans.
While at the same time the involvement of companies was a critical part of the
vision since 2004, from helping build software, selling GPS devices, hosting
servers, and contributing data. And certainly bringing new people into the
community -- no matter how people find OSM, I have almost universally seen a
magic gleam in the eye of people who take part, that forms the core of many
corporate initiatives in OSM. Just because my brain exploded with that vision
of OSM before I started having the supreme privilege to spend my working days
on it does not entitle me to some more exalted position.
I wonder if some of us have lost touch with that spirit, as OpenStreetMap has
succeeded so wildly. I was so absorbed the audacious vision of OSM in 2005, I
still am regularly shocked that anyone takes OSM seriously. Yes it is radical
in 2019 to reject corporations and machine learning. But I think we have a lot
more to offer than conservative rejection; rather we have a wildly successful,
collaborative, practical approach that puts humans in the fore of complex
technologies, as the world grapples with very complex times.
The reaction to Facebook's work really confuses me. Have critics of it actually
tried it? I found it a measured approach, where every edit needs to be examined
closely by a human and is checked for quality. The advantage of it, where I
tried it in a dense partially mapped urban settlement, is that it highlighted
missing streets very well, and made what would have been a maddening squinting
process a bit smoother and more enjoyable. I still felt satisfaction in what I
was doing. From talking with folks here in Kenya, there is genuine excitement
at these new techniques. They've experienced the challenges of creating the
map, and want to focus and build skills where their human abilities are most
valuable.
Now I am not saying that we accept anything without a critical examination.
Absolutely not! What worries me is that our criticisms are not informed. And
that there are valuable corporate contributions, and those that are not, and
the same goes for new technologies.
Yes, there are quality issues. Yes, there are issues of the experience of the
map and the community we built. Yes, there are serous issues of displacement
and alienation. What are these specifically, and what are the range of
responses we can explore together?
To take one example, Simon rightly points out that road geometry is only a
portion, and perhaps the easiest portion, of what needs mapping. And that
metrics to measure overall completeness sets real goals for us. How can we
rally and build community around this? So many of our tools are oriented to
greenfield mapping. What creative workflows, metrics, analysis and
visualizations of OSM data can bring the same thrill of creating the map from a
completely blank slate, to a stage of the map where the base geometry is there?
-Mikel
* Mikel Maron * +14152835207 @mikel s:mikelmaron
On Saturday, July 27, 2019, 01:43:59 PM GMT+3, Simon Poole <si...@poole.ch>
wrote:
Am 26.07.2019 um 19:30 schrieb Naveen Francis:
Including my ₹ 0.10 (Indian ten paisa) Echoes same thoughts of Brazilian
Real.
AI-assisted human mapping tools will be a good aid for the OSM community.
"Map faster, Map better".
40,00,000 kms to be mapped in India. 15 years of OSM mapped 18,00,000 kms.
The (rhetoric) question is, why is this the case?
Because the community in India is still very small relative to the population
size.
So from where will the additional contributors come from that will turn the
additional 4 million road geometries in to something really useful? There is a
real danger of the desire for "completeness" instead of quality resulting in
multiple TIGER 2.0s, and we are just now slowly working ourselves out of the
hole we dug (full of good intentions) with the original.
Note on the side: outside of raw total road length, a much more sensible
comparison would be completeness measures per road categories (which I suspect
is likely to look far less dramatic) and which might give more realistic goals
for the community.
Simon
thanks,
naveenpf
On Fri, Jul 26, 2019 at 4:42 AM Sérgio V. <svo...@hotmail.com> wrote:
Just adding my R$0,02 (Brazilian Real).
I guess soon the AI assisted Human mapping will happen, it may be a very good
help.
But I can't evaluate what's been publicized July 23, 2019 by
https://ai.facebook.com/blog/mapping-roads-through-deep-learning-and-weakly-supervised-training
"To browse our machine learning road predictions or start mapping with RapiD,
please visit mapwith.ai."
So at "Map faster, Map better" https://mapwith.ai/#14/6.13864/6.7698 , I
actually can't evaluate any result for roads at max zoom level 14, to see if
it's really better. I can just believe it can be.
- - - - - - - - - - - - - - - -
Sérgio - http://www.openstreetmap.org/user/smaprs
_______________________________________________
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk
_______________________________________________
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk
_______________________________________________
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk
_______________________________________________
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk
