Re: [OSM-talk] recent changes in rendering the map make it worse

[Andy Townsend]

On 23/09/2015 11:53, joost schouppe wrote:
... I don't know how we could expect OSM-carto to reflect all our 
needs. Apart from being a tool for mappers (i.e. showing as much as 
possible), it also wants to be pretty and useful for non-mappers. 
Trying to make everyone happy might make everyone unhappy in the end.


I completely agree with that. 
https://github.com/gravitystorm/openstreetmap-carto/blob/master/CARTOGRAPHY.md 
sums up what it's trying to do, and to be fair that does say that the 
three main aims "pull in different directions".  For comparison I 
recently had a look at the OSM map style as it was in April 2014, and 
personally I'm not convinced that progress has been made in two of the 
three areas mentioned as goals (specfically detail's less visible, the 
design _is_ clearer (as in more coherent) and it's not any easier to 
customise).




If you want to make people who need paths happy, just send them to 
waymarkedtrails.org  or to opentopomap.org 
 . And if that doesn't make you happy, use 
overpass-turbo (e.g. http://overpass-turbo.eu/s/bCm) to highlight the 
features you need. And show them in a pretty map with umap (see 
http://www.mappa-mercia.org/2014/09/creating-an-always-up-to-date-map.html 
for instructions).
(Yes, the performance of umap/overpass combo isn't great, that could 
really use some improvement. And yes, it would be nice if osm.org 
 was more of a portal for all of us special needs people)


To add to that, creating "a personal map style" of a small area using 
https://switch2osm.org/serving-tiles/manually-building-a-tile-server-14-04/ 
whilst not for everyone, isn't that difficult these days, and it doesn't 
require huge server resources.  There are also canned options such as 
Mapbox of course if you want someone to host it for you.  It'd be nice 
if there was a way to see the tiles on osm.org without resorting to 
cludges such as 
http://wiki.openstreetmap.org/wiki/User:SomeoneElse/Your_tiles_from_osm.org 
though.


Cheers

Andy (SomeoneElse)

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

Re: [OSM-talk] recent changes in rendering the map make it worse

[joost schouppe]

While I am guilty of some of the same reflexes (see
https://github.com/gravitystorm/openstreetmap-carto/issues/110 ), I don't
know how we could expect OSM-carto to reflect all our needs. Apart from
being a tool for mappers (i.e. showing as much as possible), it also wants
to be pretty and useful for non-mappers. Trying to make everyone happy
might make everyone unhappy in the end.

If you want to make people who need paths happy, just send them to
waymarkedtrails.org or to opentopomap.org . And if that doesn't make you
happy, use overpass-turbo (e.g. http://overpass-turbo.eu/s/bCm) to
highlight the features you need. And show them in a pretty map with umap
(see
http://www.mappa-mercia.org/2014/09/creating-an-always-up-to-date-map.html
for instructions).
(Yes, the performance of umap/overpass combo isn't great, that could really
use some improvement. And yes, it would be nice if osm.org was more of a
portal for all of us special needs people)

2015-09-23 11:42 GMT+02:00 moltonel 3x Combo :

> On 23/09/2015, henk van der laan  wrote:
> > Now I'm really getting confused. osm-carto is the current version and
> > gsoc the new one, right?
>
> There's been a tremendous amount of discussion, feedback, a tweaks
> about this change:
> * https://github.com/gravitystorm/openstreetmap-carto/pull/1736
> * http://www.openstreetmap.org/user/Mateusz Konieczny/diary
> * various calls for participation on this mailing list.
>
> OSM, and osm-carto in particular are developed in the open. If
> anything, they are often flooded by feedback. Please use open source
> to your advantage and check the process, the challenges and the
> compromises before passing judgement, and contribute using those
> channels if you can.
>
> Osm-carto is full of compromises trying to reach a balance between all
> the requirements, it is no easy task. For example, the tertiary roads
> styling you bemoan has been heavily debated before reaching the
> current conclusion; a tough requirements was keeping a high enough
> contrast for non-ideal viewing conditions.
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>



-- 
Joost @
Openstreetmap  |
Twitter  | LinkedIn
 | Meetup
 | Reddit
 | Wordpress

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

Re: [OSM-talk] recent changes in rendering the map make it worse

[Lester Caine]

On 23/09/15 13:59, Andy Townsend wrote:
> To add to that, creating "a personal map style" of a small area using
> https://switch2osm.org/serving-tiles/manually-building-a-tile-server-14-04/
> whilst not for everyone, isn't that difficult these days, and it doesn't
> require huge server resources.  There are also canned options such as
> Mapbox of course if you want someone to host it for you.  It'd be nice
> if there was a way to see the tiles on osm.org without resorting to
> cludges such as
> http://wiki.openstreetmap.org/wiki/User:SomeoneElse/Your_tiles_from_osm.org
> though.

I'm STILL struggling to get a private tile server working just covering
the UK. I DO have my own tweaked version of the style shhet running via
Tilemil, but while the database SHOULD be updating, I'm not seeing
changes I made rendered on the local copy, and I still need to get a
proper tile cache working.

But what I do have working shows in house rendering is not at all
difficult. Just need to nail all the cludges in a simple to roll package
of tools?

-- 
Lester Caine - G8HFL
-
Contact - http://lsces.co.uk/wiki/?page=contact
L.S.Caine Electronic Services - http://lsces.co.uk
EnquirySolve - http://enquirysolve.com/
Model Engineers Digital Workshop - http://medw.co.uk
Rainbow Digital Media - http://rainbowdigitalmedia.co.uk

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

Re: [OSM-talk] recent changes in rendering the map make it worse

[Mateusz Konieczny]

On Wed, 23 Sep 2015 13:59:43 +0100
Andy Townsend  wrote:

> For comparison I 
> recently had a look at the OSM map style as it was in April 2014 (...)
> and it's not any easier to customise.

This is less subjective than two others - and in that case I am pretty
sure that situation improved. For example replacing many hardcoded
colour values by variables and png to svg icon conversion makes
customization easier.

There were also improvements and I am pretty sure that that changes
that made customization harder were really rare and limited.

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

Re: [OSM-talk] recent changes in rendering the map make it worse

[henk van der laan]

> W dniu 22.09.2015 15:49, Mateusz Konieczny napisa?(a):
> 
> > Note that it is not including the latest changes that are not deployed
> > on the OSM website (currently it includes the new road style).
> 
> One can preview those changes here (for some testing places):
> 
> http://bl.ocks.org/pnorman/raw/c61d6b11193081910866/#6.00/40.000/-75.000

Now I'm really getting confused. osm-carto is the current version and
gsoc the new one, right?
It seems road importance has dropped a level, and tertiary is no longer distinct
by colour. That would have severe impications for the map in Europe,
where tertiary roads are the main roads in rural areas. They link
villages, or in case of city's neighberhoods. As so they are the
preferred roads for travelling and should be emphasized over
unclassified.
This may work at a low zoomlevel but at higher zoomlevels it adds
confusion because tertiary and unclassified have the same colour.


regards

Henk


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

Re: [OSM-talk] recent changes in rendering the map make it worse

[Mateusz Konieczny]

On Wed, 23 Sep 2015 09:10:44 +0200
henk van der laan  wrote:

> 
> 
> > W dniu 22.09.2015 15:49, Mateusz Konieczny napisa?(a):
> > 
> > > Note that it is not including the latest changes that are not
> > > deployed on the OSM website (currently it includes the new road
> > > style).
> > 
> > One can preview those changes here (for some testing places):
> > 
> > http://bl.ocks.org/pnorman/raw/c61d6b11193081910866/#6.00/40.000/-75.000
> 
> Now I'm really getting confused. osm-carto is the current version and
> gsoc the new one, right?

osm-carto is currently deployed version and gsoc is version with the
new road style and without some relatively minor recent changes.

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

Re: [OSM-talk] recent changes in rendering the map make it worse

[moltonel 3x Combo]

On 23/09/2015, henk van der laan  wrote:
> Now I'm really getting confused. osm-carto is the current version and
> gsoc the new one, right?

There's been a tremendous amount of discussion, feedback, a tweaks
about this change:
* https://github.com/gravitystorm/openstreetmap-carto/pull/1736
* http://www.openstreetmap.org/user/Mateusz Konieczny/diary
* various calls for participation on this mailing list.

OSM, and osm-carto in particular are developed in the open. If
anything, they are often flooded by feedback. Please use open source
to your advantage and check the process, the challenges and the
compromises before passing judgement, and contribute using those
channels if you can.

Osm-carto is full of compromises trying to reach a balance between all
the requirements, it is no easy task. For example, the tertiary roads
styling you bemoan has been heavily debated before reaching the
current conclusion; a tough requirements was keeping a high enough
contrast for non-ideal viewing conditions.

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

Re: [OSM-talk] recent changes in rendering the map make it worse

[henk van der laan]

Hi Mathijs,

In response I've combined some answers for readablilliity.

> Hi Henk,
> 
> On 22 September 2015 at 12:43, henk van der laan  
> wrote:
>
> To aid the discussion, here you can compare the old and new rendering
> (Netherlands/Belgium/Luxembourg only):
> http://bl.ocks.org/tyrasd/raw/67a00dbb869456094373/#9.00/52.3535/5.0572

Another tool would be this:  
http://tools.geofabrik.de/mc/#16/52.3581/4.8658=2=mapnik=geofabrik
Side by side view of the Vondelpark. 
At what zoomlevel somthing 

> What mobile device are you using? We received this complaint more
> often, but I don't seem to have the same problems on my own device.

All. Mobile devices are by nature difficult to read in bright light and
allmost impossible to read in direct sunlight. But PC with calibrated
monitor suffers as well.

> > - the map lost it's contrast and colour-opposition. The total map is
> > dimmed, and looks flat. This is not an improvement: with the current
> > generation of mobile screens it is nearly impossible to read the map in
> > bright conditions, let alone in direct sunlight.
> 
> Could you be a bit more specific about which objects and zoomlevels
> you are referring to?
> Are there any specific landuse areas where the problem is most severe?

In fact this is at all zoomlevels and against all landuse. The point
here is that red and green are closely together in gray-value. Small red
dots on a green background are more difficult to read than brown lines
on a green background. For that reason red on green is a not a good
choice. As you may know this colorscheme was discussed massively with
the introduction of the new Dutch ANWB 'mushroom'. Many complaints from
visually impaired and elderly people followed. As well as 'normal'
cyclist who found that in twilight this turned out to be unreadable.
Real bad is a path in red-dot rendering against landuse=farm or heath.

Take a look here: 
http://tools.geofabrik.de/mc/#15/52.9294/6.2958=2=mapnik=geofabrik

> > - dimmed colours seem to make the difference between highways less
> > visible: secundary, tertiary and unclassified do not stand out as
> > different as they used to.
> 
> I agree. We are rolling out a chance soon that will make the highways
> much brighter.

In that case increasing the contrast for all ways would be better. 
So including cycleways and path.

> > Apart from this the underlying thoughts for the changes are a point of
> > concern: reading through the different threads I detect an alarming
> > trend to favor roads for motorized traffic over pedestrians/cyclist etc.
> 
> This is certainly not a deliberate choice by the team that maintains
> the style. Interesting enough, I don't think any of the people who
> have been working on the changes you mention even own a car...

That surprizes me. Seeing the rendering does not give that impression.

> > I even read one comment that suggests to get rid of bridleways in
> > rendering because there are no bridleways in his aerea ...
> > Others doubt the use of different colours for cycleways, because they
> > do not form a special roadtype ..  only a legal distinction.
> > Following the last thought there would not be any reason to have a
> > distinction between primary/secundary/trunk etc. either.
> 
> Everyone is free to comment whatever they like, of course... I don't
> think any of these arguments were considered as serious reasons to
> change the style by the style maintainers.

No, it's not in the style .. yet. But to me it's a warning sign. 

> > - Colourscheme does not adhere to even the basic cartographic rules.
> 
> Could you expand on that?

I'll expand on that in a general sense. Colour schemes should not be
judged on colour alone. Every colour has a corresponding grayscale.
When testing a map ( or any graphic for that matter ) you should also
judge your work in grayscale. Consider the old mapping-problem of how
many colours to use to have no adjecent colours the same: the colours
used also give a good grayscale.
The artist's way of judging this is by squinting your eyes. This mimics 
a grayscale view. Using this on the last linked examples clearly reveals
the flaw in red dots.


> > - Readabillity is rapidly becoming worse. In the past I directed many
> > to OSM, praising it's very complete and readable map over Google/Bing.
> > That is no longer true.
> 
> You're always free to help us improving the style sheet at
> https://github.com/gravitystorm/openstreetmap-carto
 
I've never felt the need to do so because I was totally satisfied with
the exisitng rendering.
As for my dedicated maps I'm using different rules alltogether. But they 
are cycle/hiking maps with an ancient look. ( since you're Dutch:
Flakplan fietskaart oude stijl, met die dikke rode fietspaden )

> > - The new rendering omits the hard work of many contributers who
> > walked/cycled miles to give OSM more than just highways.
> 
> The stylesheet maintainers are amongst this group, so that's not
> really fair criticism.
> 

In that 

Re: [OSM-talk] recent changes in rendering the map make it worse

[Mateusz Konieczny]

On Tue, 22 Sep 2015 13:32:42 +0200
Matthijs Melissen  wrote:

> To aid the discussion, here you can compare the old and new rendering
> (Netherlands/Belgium/Luxembourg only):
> http://bl.ocks.org/tyrasd/raw/67a00dbb869456094373/#9.00/52.3535/5.0572

Note that it is not including the latest changes that are not deployed
on the OSM website (currently it includes the new road style).

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

Re: [OSM-talk] recent changes in rendering the map make it worse

[Matthijs Melissen]

Hi Henk,

On 22 September 2015 at 12:43, henk van der laan  wrote:
> Over the past couple of month some big changes in the rendering of the
> map have been made.
> These changes do not improve the readabillity of the map, and despite
> it's good intentions do not seem to take into account the basics of map
> rendering, but rather look at tagging and decide on that.

To aid the discussion, here you can compare the old and new rendering
(Netherlands/Belgium/Luxembourg only):
http://bl.ocks.org/tyrasd/raw/67a00dbb869456094373/#9.00/52.3535/5.0572

What mobile device are you using? We received this complaint more
often, but I don't seem to have the same problems on my own device.

> - the map lost it's contrast and colour-opposition. The total map is
> dimmed, and looks flat. This is not an improvement: with the current
> generation of mobile screens it is nearly impossible to read the map in
> bright conditions, let alone in direct sunlight.

Could you be a bit more specific about which objects and zoomlevels
you are referring to?

> - several highway-types have gotten a smaller line-width, up to a point
> where the lines are impossible to see even on a high resolution screen.
> In particular footway/path/cycleway are lost.

This was done (only on low zoomlevels) to prevent clutter of footways
in dense areas, which also caused serious readability problems. See
for example Vondelpark on zoom level 13. We realise it makes it harder
to read footways in less dense areas on zoom levels 13 and 14 (and
only on these), but that's a trade-off we had to make.

> - The combining of footway/path ( apart from beeing wrong to start with )
> led to little red dots. These do not stand out , especially against some
> landuse colours. And again: especially on mobile device in bright
> light they are invisible.

Are there any specific landuse areas where the problem is most severe?

> - dimmed colours seem to make the difference between highways less
> visible: secundary, tertiary and unclassified do not stand out as
> different as they used to.

I agree. We are rolling out a chance soon that will make the highways
much brighter.

> Apart from this the underlying thoughts for the changes are a point of
> concern: reading through the different threads I detect an alarming
> trend to favor roads for motorized traffic over pedestrians/cyclist etc.

This is certainly not a deliberate choice by the team that maintains
the style. Interesting enough, I don't think any of the people who
have been working on the changes you mention even own a car...

> I even read one comment that suggests to get rid of bridleways in
> rendering because there are no bridleways in his aerea ...
> Others doubt the use of different colours for cycleways, because they
> do not form a special roadtype ..  only a legal distinction.
> Following the last thought there would not be any reason to have a
> distinction between primary/secundary/trunk etc. either.

Everyone is free to comment whatever they like, of course... I don't
think any of these arguments were considered as serious reasons to
change the style by the style maintainers.

> Imo OSM is a general map, that is: no particular feature should be left
> out or favored over another. I'm not really interested in powerlines,
> but do not oppose rendering them on the map.

The problem is that we're at a point where we have so many features,
that we need to make choices. For ex
Also, we simply have so many different features now, that we have no
longer enough colours to render them all...

> So the conclusions:
> - Currently the new rendering favors motorized traffic over others.
> If the goal of OSM is to become a road-map like Google that would mean
> the loss of thousands of mappers who use OSM as the only complete map.

No, that's not the goal of OSM or the default rendering.

> - Colourscheme does not adhere to even the basic cartographic rules.

Could you expand on that?

> - Readabillity is rapidly becoming worse. In the past I directed many
> to OSM, praising it's very complete and readable map over Google/Bing.
> That is no longer true.

You're always free to help us improving the style sheet at
https://github.com/gravitystorm/openstreetmap-carto

> - The new rendering omits the hard work of many contributers who
> walked/cycled miles to give OSM more than just highways.

The stylesheet maintainers are amongst this group, so that's not
really fair criticism.

> I vote to re-enstate the old rendering, honouring the work of many and
> getting a readable map that is supports all users and contributers.

We're always happy to consider improvements to the map, but simply
reverting changes is not the way to go in this case, as that would
also reintroduce the problems solved by these changes in the first
place.

> I wonder how many users would back me up on this.

OSM is not a democracy. :) In particular, concrete proposals on how to
do something are more likely to be considered than 

[OSM-talk] recent changes in rendering the map make it worse

[henk van der laan]

Over the past couple of month some big changes in the rendering of the
map have been made.
These changes do not improve the readabillity of the map, and despite
it's good intentions do not seem to take into account the basics of map
rendering, but rather look at tagging and decide on that.

- the map lost it's contrast and colour-opposition. The total map is
dimmed, and looks flat. This is not an improvement: with the current
generation of mobile screens it is nearly impossible to read the map in
bright conditions, let alone in direct sunlight. 
- several highway-types have gotten a smaller line-width, up to a point
where the lines are impossible to see even on a high resolution screen.
In particular footway/path/cycleway are lost.
- The combining of footway/path ( apart from beeing wrong to start with )
led to little red dots. These do not stand out , especially against some
landuse colours. And again: especially on mobile device in bright
light they are invisible.
- dimmed colours seem to make the difference between highways less
visible: secundary, tertiary and unclassified do not stand out as
different as they used to.

Apart from this the underlying thoughts for the changes are a point of
concern: reading through the different threads I detect an alarming
trend to favor roads for motorized traffic over pedestrians/cyclist etc.
I even read one comment that suggests to get rid of bridleways in
rendering because there are no bridleways in his aerea ...
Others doubt the use of different colours for cycleways, because they 
do not form a special roadtype ..  only a legal distinction.
Following the last thought there would not be any reason to have a
distinction between primary/secundary/trunk etc. either.
Imo OSM is a general map, that is: no particular feature should be left
out or favored over another. I'm not really interested in powerlines,
but do not oppose rendering them on the map. 

So the conclusions:
- Currently the new rendering favors motorized traffic over others.
If the goal of OSM is to become a road-map like Google that would mean 
the loss of thousands of mappers who use OSM as the only complete map.
- Colourscheme does not adhere to even the basic cartographic rules. 
- Readabillity is rapidly becoming worse. In the past I directed many
to OSM, praising it's very complete and readable map over Google/Bing.
That is no longer true.
- The new rendering omits the hard work of many contributers who
walked/cycled miles to give OSM more than just highways.

I vote to re-enstate the old rendering, honouring the work of many and 
getting a readable map that is supports all users and contributers.

I wonder how many users would back me up on this. 

Henk ( Noordfiets on OSM )


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

Re: [OSM-talk] recent changes in rendering the map make it worse

[Christoph Hormann]

On Tuesday 22 September 2015, henk van der laan wrote:
> Over the past couple of month some big changes in the rendering of
> the map have been made.
> These changes do not improve the readabillity of the map, and despite
> it's good intentions do not seem to take into account the basics of
> map rendering, but rather look at tagging and decide on that.

Not sure if you are aware of this but with that introductory statement 
you essentially say that everyone who contributed to the standard style 
in the last months has done a poor job and essentially is incompetent.

That certainly is an opinion that can be respected but you need to 
expect that quite a few people will disagree here which is not a very 
good basis regarding the rest of your critique.

Long story short: development of the standard style is open to 
everyone - if there are things you would like to improve and feel you 
can do better you are welcome to contribute.  There are several open 
issues in the bug tracker already dealing with problems introduced by 
recent changes and you are welcome to make specific suggestions how to 
improve things there.  But just stating a sweeping 'every change is 
bad' without showing that you have considered the reasons why this 
change has been made in the first place does not make you very 
believable.

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

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

Re: [OSM-talk] recent changes in rendering the map make it worse

[Daniel Koć]

W dniu 22.09.2015 15:49, Mateusz Konieczny napisał(a):


Note that it is not including the latest changes that are not deployed
on the OSM website (currently it includes the new road style).


One can preview those changes here (for some testing places):

http://bl.ocks.org/pnorman/raw/c61d6b11193081910866/#6.00/40.000/-75.000

--
"The train is always on time / The trick is to be ready to put your bags 
down" [A. Cohen]


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

Re: [OSM-talk] recent changes in rendering the map make it worse

[henk van der laan]

> On Tuesday 22 September 2015, henk van der laan wrote:
> > Over the past couple of month some big changes in the rendering of
> > the map have been made.
> > These changes do not improve the readabillity of the map, and despite
> > it's good intentions do not seem to take into account the basics of
> > map rendering, but rather look at tagging and decide on that.
> 
> Not sure if you are aware of this but with that introductory statement 
> you essentially say that everyone who contributed to the standard style 
> in the last months has done a poor job and essentially is incompetent.
> 
> That certainly is an opinion that can be respected but you need to 
> expect that quite a few people will disagree here which is not a very 
> good basis regarding the rest of your critique.
> 
> Long story short: development of the standard style is open to 
> everyone - if there are things you would like to improve and feel you 
> can do better you are welcome to contribute.  There are several open 
> issues in the bug tracker already dealing with problems introduced by 
> recent changes and you are welcome to make specific suggestions how to 
> improve things there.  But just stating a sweeping 'every change is 
> bad' without showing that you have considered the reasons why this 
> change has been made in the first place does not make you very 
> believable.

Stating that something is a bad job by someone does not mean that
person is incompetent. That is merely an interpretation of the person(s)
involved fighting of critique.
You may find it a harsh thing to say, but yes: I think it's a bad job.

Not every change is bad, but than again a change is not per definition
good. Only change for the sake of change is bad per definition.
And I've considered many of the reasons and found few really convincing
for the changes made.


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

Re: [OSM-talk] recent changes in rendering the map make it worse

[Christoph Hormann]

On Tuesday 22 September 2015, henk van der laan wrote:
>
> Stating that something is a bad job by someone does not mean that
> person is incompetent.

No, but if someone does a bad job he/she is either incompetent or has 
bad intentions and since you specifically say you assume good 
intentions...

It is always a good idea to consider that just because you do not 
recognize the advantages of a certain change does not mean there aren't 
any.  Hence it is usually more productive to concentrate on the 
negative effects of a change than to insist that the change is 
intrinsically negative.

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

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