Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Daniel Koć]

W dniu 26.06.2018 o 18:08, Sven Geggus pisze:

> This seems to completely kill my rendering performance.
> Will I need additional indexes?


We're currently investigating it:

https://github.com/openstreetmap/chef/issues/168#issuecomment-400232261


I see two potential sources of this problem at the moment:

https://github.com/gravitystorm/openstreetmap-carto/pull/2640
https://github.com/gravitystorm/openstreetmap-carto/pull/2874

Could you test which one seems to be hitting your performance?

-- 
"My method is uncertain/ It's a mess but it's working" [F. Apple]


___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Sven Geggus]

Daniel Koć  wrote:

> Today, v4.12.0 of the OpenStreetMap Carto stylesheet (the default
> stylesheet on the OSM website) has been released.

This seems to completely kill my rendering performance.

Will I need additional indexes?

> time render_single_tile.py -s osm.xml -o test-v4.11.0.png -u 
> /15/17090/11446.png
real0m10.586s
user0m0.940s
sys 0m0.304s

> time render_single_tile.py -s osm.xml -o test-v4.12.0.png -u 
> /15/17090/11446.png
real6m52.459s
user0m1.336s
sys 0m0.440s


I'm using Mapnik 3.0.12 and carto 0.18.2. Would I need to upgrade?

Regards

Sven

-- 
"Thinking of using NT for your critical apps?
  Isn't there enough suffering in the world?"
   (Advertisement of Sun Microsystems in Wall Street Journal)
/me is giggls@ircnet, http://sven.gegg.us/ on the Web

___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Daniel Koć]

W dniu 22.06.2018 o 17:16, Daniel Koć pisze:
> Dear all,
>
> Today, v4.12.0 of the OpenStreetMap Carto stylesheet (the default
> stylesheet on the OSM website) has been released. Once changes are
> deployed on the openstreetmap.org it will take couple of days before
> all tiles show the new rendering.

Hi again,

The OSM.org deployment is not happening yet, and one of the problems is
some performance issue on low zoom levels:

https://github.com/openstreetmap/chef/issues/168#issuecomment-400232261

Could anybody test what's the source of the problem? I expected rather
that it would speed up the low zoom rendering due to the smaller reads
from a database together with the update of indexes:

https://github.com/gravitystorm/openstreetmap-carto/pull/2874

-- 
"My method is uncertain/ It's a mess but it's working" [F. Apple]



___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[ika-chan! (OpenStreetMap)]

Hi all,

I wish to express my concern about how the update could make it harder for me 
to recolour the roads to meet local customs or requirements. This is in 
relation to my tutorial about how to make the stylesheet align with the 
customary road colours of the British Isles (not just the UK): 
https://wiki.openstreetmap.org/wiki/User:Ika-chan!/British_road_colours

Health permitting, I am currently in the process of creating a tutorial about 
setting up a server for private worlds.

Best,

ika-chan!

> On 22 Jun 2018, at 16:16, Daniel Koć  wrote:
> 
> Dear all,
> 
> Today, v4.12.0 of the OpenStreetMap Carto stylesheet (the default
> stylesheet on the OSM website) has been released. Once changes are
> deployed on the openstreetmap.org it will take couple of days before
> all tiles show the new rendering.
> 
> [snipped by the responder, due to length]


___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Christoph Hormann]

On Saturday 23 June 2018, Christoph Hormann wrote:
> [...]
>
> I still hope Lucas will write up a few details about the technical
> approach that would enable you to more easily re-implement this in a
> different style.

By the way Lucas has already done so, you can find it here:

https://github.com/sommerluk/roadpatternrendering

Apart from the 2d pattern rendering of lines this also nicely explains 
how the drawing order works in Mapnik/CartoCSS in general.

-- 
Christoph Hormann
http://www.imagico.de/

___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Christoph Hormann]

On Sunday 24 June 2018, Sven Geggus wrote:
>
> May I mention the fact, that Mapserver is still actively maintained
> and that there even is a carto preprocessor called magnacarto which
> is able to produce mapserver style files.

Yes, Mapserver has some very powerful features - including even the 
ability to code styling logic in javascript:

http://mapserver.org/de/mapfile/styleitem.html

The problem i have with Mapserver from a designer point of view is that 
it offers a complex high level feature set with many functions clearly 
intended for direct use.  But at the same time the language to 
formulate this in (the mapfiles) is not really suited to write and 
manage more complex stuff directly.  And Magnacarto only supports the 
basic features offered by both Mapserver and Mapnik which does not 
really work for Mapserver in the end since it does not make use of what 
Mapserver can but Mapnik cannot.

-- 
Christoph Hormann
http://www.imagico.de/

___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Sven Geggus]

Daniel Koć  wrote:

> Moreover this particular approach to surface rendering has been
> developed for about a year, so I'm surprised you missed the whole
> discussion.

The problem is not with surface rendering but this:

"* code re-ordering - no rendering change"

I just try to manually re-add all my changes to a vanilla v4.12.0 roads.mss 
which
seem to work so far.

Regards

Sven

-- 
"Thinking of using NT for your critical apps?
  Isn't there enough suffering in the world?"
   (Advertisement of Sun Microsystems in Wall Street Journal)
/me is giggls@ircnet, http://sven.gegg.us/ on the Web

___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Sven Geggus]

Christoph Hormann  wrote:

> That's a really tough call.

May I mention the fact, that Mapserver is still actively maintained and that
there even is a carto preprocessor called magnacarto which is able to
produce mapserver style files.

Sven

-- 
Freiheit ist immer die Freiheit des Andersdenkenden
(Rosa Luxemburg)

/me is giggls@ircnet, http://sven.gegg.us/ on the Web

___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Christoph Hormann]

On Sunday 24 June 2018, Mateusz Konieczny wrote:
>
> > It is all open source of course so anyone can pick up development
> > but you need to ask yourself if this would be worthwhile of course.
> >  Both the technological and cartographic environment have changed a
> > lot since these programs were originally designed.
>
> What would you recommend for someone starting a new project?

That's a really tough call.

The only real recommendation i can give is a quite abstract one: Go 
small, concentrate on doing one thing and one thing only and avoid the 
impetus to go broader and then get bogged down by conflicting goals and 
the complexity of the whole thing.

It is interesting you mentioned 3d rendering since i think (and also 
have frequently written in the past) map rendering can learn a lot from 
the historic development of general 3d rendering (which has a much 
longer and more faceted history than digital map production)

For a map design project with similar scope and goals as OSM-Carto the 
situation today is quite impossible indeed.  But you need to keep in 
mind the OSM standard style (combining the old XML style and OSM-Carto 
here) and Mapnik essentially grew up together so to speak.  You can't 
expect today to be able to find a framework that can directly take over 
this role Mapnik essentially grew into over many years.  So the goal 
would be to look for a toolchain that has the potential to grow into 
such a role over time which is both a technical requirement and one 
regarding the developer community.

But it is also possible that what the OSM standard style represents, the 
one size fits all map for OpenStreetMap, will end in the near future - 
as a dinosaur incapable to adjust to a changing world.  And that my 
recommendation to go small and concentrate on doing one thing and one 
thing only not only is to be applied to software development but also 
map design projects.

But i don't see this happening at the moment yet because i don't see the 
open, non-commercial limited scope map design projects flourishing.  
How likely this is to happen in the future depends largely on the 
burden of suffering felt by the OSM community because of OSM-Carto not 
being able to deliver certain things.  And to get back to the original 
subject of this thread - the issue regarding rendering of paved/unpaved 
was open for nearly five years, the desire to have this rendered had 
existed probably already significantly longer (surface tag reached one 
million uses in 2009).  And despite this pressure and the incapability 
of OSM-Carto to deliver on this it has none the less not led to the 
development of a specialized roads map style to fill this gap.  So the 
inertia in the social environment surrounding the OSM standard style is 
immense.  It is quite possible that from this point of view OSM-Carto 
could continue more or less indefinitely and the only contraint would 
be to mitigate the scaling problems (as it is already done with 
aggressive way_area filtering for example).

> Recently Mapzen ecosystem looked interesting for me (though
> making maps good from cartographic design viewpoint was not one of
> its strength), but I picked poor moment to start with Mapzen :)

Well - as i wrote some time ago 
(http://blog.imagico.de/on-permanence-in-it-and-cartography/) Mapzen 
did not have any truly revolutionary and innovative ideas in their 
portfolio.  They had the advantage of not being as dominated by 
immediate business needs as some of the other commercial players which 
allowed them to think things through a bit better in some cases and 
make tools a bit more generic which also makes them look more 
attractive to third parties obviously.  Still there is a lot of 'just 
another version of the same thing essentially' in what they did.

-- 
Christoph Hormann
http://www.imagico.de/

___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Mateusz Konieczny]

23. Jun 2018 12:19 by o...@imagico.de :


> Fixes for major bugs like here:
>
> https://github.com/mapnik/mapnik/issues/3558 
> 
>
> can take a year.




Well, at least it was fixed. It is much better experience than what I had after 
reporting

some bugs to Firefox/Libre Office bug trackers or other major projects.




 There are issues with Mapnik but at least project is not dead.



> It is all open source of course so anyone can pick up development but 
> you need to ask yourself if this would be worthwhile of course.  Both 
> the technological and cartographic environment have changed a lot since 
> these programs were originally designed.
>




What would you recommend for someone starting a new project?

Sorry if I missed recommendations on your page but I found only 
http://www.imagico.de/pov/tutorials_en.php 
 and from what I understand it is 
for 


making images like http://earth.imagico.de/view.php?site=newzealand_whole 


rather than for making maps.







Recently Mapzen ecosystem looked interesting for me (though 


making maps good from cartographic design viewpoint was not one of

its strength), but I picked poor moment to start with Mapzen :)



___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Christoph Hormann]

On Saturday 23 June 2018, Yves wrote:
>
> Which is not going to change if everybody scripts its own way :)
> I did not followed the discussion around that change, but it would be
> a good idea to make changes upstream to come up with an equivalent
> style without special code. Yves

The question is how long would you be committed to wait for an upstream 
change to be implemented?  

Fixes for major bugs like here:

https://github.com/mapnik/mapnik/issues/3558

can take a year.  Not sure how requests for feature additions fare on 
average.  But i think without an existing implementation being 
submitted as a pull request the chances are relatively small.  It is 
commendable that Mapbox still does some basic maintainance and 
bugfixing there but this is hardly more than keeping the lights on.  
And any feature addition would also need support in Carto which - as 
Daniel already mentioned - is definitely and officially not maintained 
any more.

It is all open source of course so anyone can pick up development but 
you need to ask yourself if this would be worthwhile of course.  Both 
the technological and cartographic environment have changed a lot since 
these programs were originally designed.

-- 
Christoph Hormann
http://www.imagico.de/

___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Daniel Koć]

W dniu 23.06.2018 o 11:18, Yves pisze:
> | Mapnik and CartoCSS being essentially
> | unmaintained,
>
> Which is not going to change if everybody scripts its own way :)
> I did not followed the discussion around that change, but it would be
> a good idea to make changes upstream to come up with an equivalent
> style without special code.

CartoCSS is dead end now - first it has been abandoned by Mapbox, and
lately by nebulon42, who made 1.0.0 release shortly before his departure.

Mapnik is a different beast. It's still being developed by artemp
(founder) and talaj at least and has some big deployments, like Mapy.cz.
I don't remember right now if we asked about adding support for our
solution, but solving things upstream is what I try to promote.

However it always takes some coding to be done and it might be not what
maintainers are focusing on. So if there's somebody interested in the
issue, please let us know.

-- 
"My method is uncertain/ It's a mess but it's working" [F. Apple]



___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Daniel Koć]

W dniu 23.06.2018 o 09:21, Sven Geggus pisze:

> Are you unaware of the fact that there are forks of OpenStreetMap Carto like
> German style which are keept in sync by merging them with every upstream
> release?

Hi, Sven!

I was aware that your fork is quite closely following osm-carto, unlike
many others, so I did not expect such problem.

Moreover this particular approach to surface rendering has been
developed for about a year, so I'm surprised you missed the whole
discussion.

-- 
"My method is uncertain/ It's a mess but it's working" [F. Apple]


___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Yves]

| Mapnik and CartoCSS being essentially
| unmaintained,

Which is not going to change if everybody scripts its own way :)
I did not followed the discussion around that change, but it would be a good 
idea to make changes upstream to come up with an equivalent style without 
special code.
Yves ___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Christoph Hormann]

On Saturday 23 June 2018, Sven Geggus wrote:
>
> I'm doing this for German style for more than 2 years now without
> that much of a hassle up till this release.
>
> This will now kill my ability to keep on doing this at least as
> roads.mss is concerned!

This is to be expected.  You are kind of spoiled because during the last 
year this is about the only larger technologically more sophisticated 
change in the style.

Discussion of this change can be found in:

https://github.com/gravitystorm/openstreetmap-carto/pull/2640

and discussion leading up to this in:

https://github.com/gravitystorm/openstreetmap-carto/issues/110

The reason for the technological complexity of this and the resulting 
mangling of the code is the lack of support to render lines with a 2d 
pattern in Mapnik.

I still hope Lucas will write up a few details about the technical 
approach that would enable you to more easily re-implement this in a 
different style.

Note there have been voices to choose a more conservative approach to 
rendering unpaved roads in OSM-Carto - like yet another form of dashing 
in addition to the various dashing types already used in road 
rendering.  There are very good cartographic reasons not to do that 
though.

All in all this is a good example for OSM-Carto being at a crossroads 
(and having been for quite some time) between staying avant-garde and 
pushing the boundaries of cartographic design and technology or being 
satisfied with shuffling the options offered by the cartographic 
mainstream within the technological framework used - and which, due to 
Mapnik and CartoCSS being essentially unmaintained, becomes more narrow 
and limiting every year.

-- 
Christoph Hormann
http://www.imagico.de/

___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] OpenStreetMap Carto release v4.12.0

[Sven Geggus]

Daniel Koć  wrote:

> Changes
> - Added rendering “surface” tag on roads with a pattern

This is not supposed to be a huge change, right?

Are you unaware of the fact that there are forks of OpenStreetMap Carto like
German style which are keept in sync by merging them with every upstream
release?

I'm doing this for German style for more than 2 years now without that much
of a hassle up till this release.

This will now kill my ability to keep on doing this at least as roads.mss is
concerned!

The commit which causes the hassle is most likely this one:

Author: Lukas Sommer 
Date:   Fri May 11 10:57:04 2018 +

Render “surface” tag on roads with a pattern (#2640)

* code re-ordering - no rendering change

* Render unpaved road surface with a pattern

WTF. Did you have a good reason which was worth killing your forks ability to
stay in sync with upstream?

Regards

Sven

-- 
"Whenever there is a conflict between human rights and property
rights, human rights must prevail." (Abraham Lincoln)

/me is giggls@ircnet, http://sven.gegg.us/ on the Web

___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

[OSM-dev] OpenStreetMap Carto release v4.12.0

[Daniel Koć]

Dear all,

Today, v4.12.0 of the OpenStreetMap Carto stylesheet (the default
stylesheet on the OSM website) has been released. Once changes are
deployed on the openstreetmap.org it will take couple of days before
all tiles show the new rendering.

Changes include:

Major changes
- Dropped subpixel accuracy for areas, which includes tuning some
database indexes. When deployed, it might speed up reading data.

Changes
- Added rendering “surface” tag on roads with a pattern
- More vertical objects rendering and tuning (man_made=tower types,
man_made=chimney, man_made=communications_tower)
- tourism=information types rendering and tuning
(information=audioguide, board, guidepost, map, office, tactile_map,
tactile_model and terminal)
- Added rendering for place=quarter
- Added rendering of historic=city_gate
- Added rendering of lock_name
- Ditch and drain name labels are rendered with some offset
- Pixel aligned ford icon
- Made amenity=shelter icon brown
- Finer man_made=pier width rendering
- Rendering living street tunnels different from residential
- Added rendering of overground power=cable like power=line
- Small documentation and code fixes

Thanks to all the contributors for this release including Adamant36 and
M1dgard, new contributors. I also like to thank nebulon42 who left our
team due to a change of his priorities for all the work on this style
and the tools we're using!

For a full list of commits, see
https://github.com/gravitystorm/openstreetmap-carto/compare/v4.11.0...v4.12.0

As always, we welcome any bug reports at
https://github.com/gravitystorm/openstreetmap-carto/issues

-- 
"My method is uncertain/ It's a mess but it's working" [F. Apple]

___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev