On 21/11/2017 13:47, Darafei "Komяpa" Praliaskouski wrote:
I've posted a -dev mail about reusing nighttime of tile rendering
servers. Some likes on GitHub, some reviews from passer-by's, no
merge, nothing about "what to fix to get it merged". For a year.
Patience you say?
https://github.com/openstreetmap/mod_tile/pull/152
Whilst I'm not a contributor to the repository there, I do have some
familiarity with the code. What you seem to be doing is interpreting
the mod_tile repository as "part of the infrastructure of
OpenStreetMap.org", and you seem to be viewing OpenStreetMap.org as an
end-user Google Maps competitor, not as a "creating map data enabler".
I regularly use mod_tile on memory-limited machines and would be
concerned if I was suddenly not able to process as large data extracts
that I could previously. I don't see any thought given in what you're
proposing to what the knock-on effects of your change would be.
/map call is technically 40x slower than it should be, but issue is
being closed with "we are not complete idiots" comments. No action
taken wherever.
https://github.com/openstreetmap/operations/issues/135
The second line of your issue starts "This causes hatred when editing
something", which is not exactly helpful if you want an in-depth
investigation of a perceived performance problem. Despite this, the
conversation that follows covers in detail the status of the problem,
and a suggestion to you where you can help. Your contributions there
(at https://github.com/zerebubuth/openstreetmap-cgimap/issues/122 )
stopped after a day.
I've said elsewhere the developing things _around_ OpenStreetMap and
with OpenStreetMap data has a surprisingly low barrier to entry - you
just download the data and off you go; there's no API with Ts and Cs to
negotiate. However, _changing_ the way that the the project or the
existing osm.org infrastructure does something will necessarily require
a series of arguments to be made and people to be persuaded, and it
seems to me that you haven't successfully done that yet, just as Yuri
didn't with his approach to mechanical editing, which led indirectly to
the WeeklyOSM article and the thread that this one developed from.
Where there are competing requirements (and there are always competing
requirements) you can't always expect everyone to agree the your view of
the requirements is the "most valid" one - see for example
https://github.com/gravitystorm/openstreetmap-carto/issues/765 . I took
the hint from that to create something else with OSM data that was (for
my purposes) better; perhaps you could do the same?
Best Regards,
Andy
_______________________________________________
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk