logo separator

[mkgmap-dev] FW: --make-opposite-cycleways option

From Gerd Petermann gpetermann_muenchen at hotmail.com on Tue Mar 10 06:21:53 GMT 2015

Hi Mike,

I still don't understand the effect on wrong routing through oneways.

Anyway, I don't see that anyone else is interested in changing this option,
so maybe I'll just commit the patch that changes mkgmap to delete
the tags handled in inc/access.

Gerd

From: mike at tvage.co.uk
To: mkgmap-dev at lists.mkgmap.org.uk
Date: Sun, 8 Mar 2015 17:04:27 +0000
Subject: Re: [mkgmap-dev] FW:  --make-opposite-cycleways option

HI Gerd, thanks - I didn't spot the entry in the documentation because
mkgmap:road-speed-class is displayed over two lines in the PDF document, so
searching for it appears not to work.
 
I have now determined that if I also delete maxspeed from the cycleway, so
that mkgmap:road-speed-class does not get set in inc/roadspeed, then the
routing works correctly.
 
Updated patch attached.
 
Regards,
Mike
 
-----Original Message-----
From: GerdP [mailto:gpetermann_muenchen at hotmail.com] 
Sent: 08 March 2015 11:08
To: mkgmap-dev at lists.mkgmap.org.uk
Subject: Re: [mkgmap-dev] FW: --make-opposite-cycleways option
 
Hi Mike,
 
the tag mkgmap:road-speed-class is documented in 
d:\mkgmap\doc\styles\internal-tags.txt 
I think it works as documented. Not sure why it breaks
routing when it is not used.
 
Gerd
 
 
Mike Baggaley wrote
> HI Gerd, I have now investigated further and have discovered that the
> difference between my lines file that works and the default that doesn't
> is
> that I don't include inc/roadspeed. The setting of mkgmap:road-speed-class
> in this file seems to break the routing when an opposite cycleway is added
> from the style file. This setting is not mentioned in the style manual, so
> I
> don't know what it is supposed to do. Can you take a look at this?
>  
> Thanks,
> Mike
>  
> From: Gerd Petermann [mailto:
 
> gpetermann_muenchen@
 
> ] 
> Sent: 07 March 2015 18:29
> To: 
 
> mkgmap-dev at .org
 
> Subject: Re: [mkgmap-dev] FW: --make-opposite-cycleways option
>  
> Hi Mike,
> 
> yes, also sounds like a good idea to me,
> but doesn't solve the initial problem regarding 
> routing in the cycleway.
> 
> Gerd
>   _____  
> 
> From: 
 
> mike at .co
 
>  <mailto:
 
> mike at .co
 
> > 
> To: 
 
> mkgmap-dev at .org
 
>  <mailto:
 
> mkgmap-dev at .org
 
> > 
> Date: Sat, 7 Mar 2015 13:07:47 +0000
> Subject: Re: [mkgmap-dev] FW: --make-opposite-cycleways option
> Hi Gerd, I suggest it might be more useful to make it a more general
> purpose
> option for affecting the order of precedence used with the continue
> statement, something like mkgmap:precedence=
> <integer>
>  where negative number
> are added before positive ones and the default precedence is 0. This would
> allow you to specify a road, bridge and cycleway on the same highway
> segment
> and specify what order they should be created in. Alternatively, the
> precedence could be specified as part of the continue statement e.g [0x10
> road_class=0 road_speed=0 resolution 24 continue precedence=1] would mean
> create after all the matches have been completed
>  
> Regards,
> Mike
> From: Gerd Petermann [mailto:
 
> gpetermann_muenchen@
 
> ]  
> Sent: 07 March 2015 09:12
> To: 
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt; 
> Subject: [mkgmap-dev] FW: --make-opposite-cycleways option
>  
> Hi all,
> 
> sorry, I hit the send button by mistake.
> ...
> Did anybody think about this proposal?
> 2) Maybe we can replace the  --make-opposite-cycleways option by 
> a new special tag like mkgmap:add_cycleway=[before|after] ?
> 
> The idea is to change the code in mkgmap so that it doesn't create a copy
> of the way BEFORE style processing, instead it checks if this tag occurs
> and
> adds a copy with bicycle-only and "oneway=no" after style processing.
> The tag value before / after tells mkgmap if the cycleway should be added
> before
> or after the "normal" way.
> I see only one theoretical problem: if the style adds the same OSM way two
> or more times
> with this tag, should we also add multiple cycleways? And where exactly?
> I think we would have to ignore all sub sequent ways.
> 
> Gerd
>   _____  
> 
>  
> Hi Mike,
> 
> I think a possible problem with your patch is that you always add the
> cycle
> way
> with type 0x10, without further checking the attributes of the highway.
> 
> The -make-opposite-cycleways option would add the cycle way with the same
> type
> as that for the car.
> 
> 
> Gerd
>   _____  
> 
> From: 
 
> mike at .co
 
>  &lt;mailto:
 
> mike at .co
 
> &gt; 
> To: 
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt; 
> Date: Fri, 6 Mar 2015 00:11:00 +0000
> Subject: Re: [mkgmap-dev] --make-opposite-cycleways option
> Hi Gerd, I added the following to the lines file in my style and it works
> fine there if I remove the --make-opposite-cycleways option, allowing just
> cycling and walking against the flow. However, it doesn't seem to work
> correctly if I add it to the default style (it allows cars to go the wrong
> way along the one-way street).
>  
> highway=* & (oneway=yes | oneway=-1 | oneway=true | oneway=1 |
> oneway=reverse) & (oneway:bicycle=no | cycleway=opposite |
> cycleway=opposite_lane | cycleway=opposite_track) {delete oneway; delete
> cycleway; set access=no; delete foot; delete vehicle; delete
> motor_vehicle;
> delete motorcar; delete goods; delete hgv; delete psv; delete emergency;
> delete taxi; delete bus; add bicycle=yes; set highway=cycleway} [0x10
> road_class=0 road_speed=1 resolution 24 continue]
>  
> I can't see why this might be happening. Has anyone any ideas (the
> attached
> patch is what I changed)?
>  
> Regards,
> Mike
>  
> From: Gerd Petermann [mailto:
 
> gpetermann_muenchen@
 
> ] 
> Sent: 04 March 2015 16:22
> To: 
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt; 
> Subject: Re: [mkgmap-dev] --make-opposite-cycleways option
>  
> Hi Mike,
> 
> the style cannot create a way, but it can add multiple routable ways for
> one
> OSM way.
> Use "continue" or "continue with actions" for that.
> 
> The comment is a bit misleading and I think the code for "old-style" is
> now
> obsolete as well.
> 
> Gerd
>   _____  
> 
> From: 
 
> mike at .co
 
>  &lt;mailto:
 
> mike at .co
 
> &gt; 
> To: 
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt; 
> Date: Wed, 4 Mar 2015 16:01:56 +0000
> Subject: Re: [mkgmap-dev] --make-opposite-cycleways option
> Hi Gerd, I was not clear what you meant by change the style to produce the
> same result, so I provided a change to the style that with the code
> changes
> produces the same result as now. Unless a style can create an extra way, I
> can't see how a style can be set up to allow cycling in both directions,
> but
> other traffic only in one direction without using the extra way created by
> the --make-opposite-cycleways code. Can a style create a way? I got the
> list
> of tags to delete from the following code which suggested that the
> following
> lines were all access tags (plus a few tags mentioned elsewhere that I
> also
> included):
>  
>                 public boolean init(ElementSaver saver, EnhancedProperties
> props) {
>                                 
>                                 if (props.getProperty("old-style", false))
> {
>                                                 // the access tags need to
> be loaded if the old style handling
>                                                 // is active and access
> restrictions are handled by the java
>                                                 // source code and not by
> the style
>                                                 usedTags.add("access");
>                                                 usedTags.add("bicycle");
>                                                 usedTags.add("carpool");
>                                                 usedTags.add("delivery");
>                                                 usedTags.add("emergency");
>                                                 usedTags.add("foot");
>                                                 usedTags.add("goods");
>                                                 usedTags.add("hgv");
>                                                 usedTags.add("motorcar");
>                                                
> usedTags.add("motorcycle");
>                                                 usedTags.add("psv");
>                                                 usedTags.add("route");
>                                                 usedTags.add("taxi");
>                                 }
>  
> I gather that route is not an access tag, so I agree that it should not be
> included.
>  
> Regards,
> Mike
>  
> From: Gerd Petermann [mailto:
 
> gpetermann_muenchen@
 
> ] 
> Sent: 04 March 2015 05:31
> To: 
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt; 
> Subject: Re: [mkgmap-dev] --make-opposite-cycleways option
>  
> Hi Mike,
> 
> as I said, I'd be more happy to have a patch that remove's the option
> and shows how to change the default style instead to get the same result 
> as with your patch.
> 
> Your patch tries to remove all kinds of tags which might lead to wrong
> routing:
> +        cycleWay.deleteTag("vehicle");
> +        cycleWay.deleteTag("motor_vehicle");
> +        cycleWay.deleteTag("carpool");
> +        cycleWay.deleteTag("delivery");
> +        cycleWay.deleteTag("emergency");
> +        cycleWay.deleteTag("foot");
> +        cycleWay.deleteTag("goods");
> +        cycleWay.deleteTag("hgv");
> +        cycleWay.deleteTag("motorcar");
> +        cycleWay.deleteTag("motorcycle");
> +        cycleWay.deleteTag("psv");
> +        cycleWay.deleteTag("route");
> +        cycleWay.deleteTag("taxi");
> +        cycleWay.deleteTag("bus");
> +        cycleWay.deleteTag("truck"); 
> 
> Please explain why you remove route=* .
> I think that one should be kept.
> 
> Gerd
>   _____  
> 
> From: 
 
> mike at .co
 
>  &lt;mailto:
 
> mike at .co
 
> &gt; 
> To: 
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt; 
> Date: Tue, 3 Mar 2015 19:45:00 +0000
> Subject: Re: [mkgmap-dev] --make-opposite-cycleways option
> HI Gerd, the attached updated patch adds a line to the default style that
> can be uncommented if you want to see the opposite cycleway names as they
> are currently.
>  
> Regards,
> Mike
>  
> From: Gerd Petermann [mailto:
 
> gpetermann_muenchen@
 
> ] 
> Sent: 03 March 2015 08:26
> To: 
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt; 
> Subject: Re: [mkgmap-dev] --make-opposite-cycleways option
>  
> Hi Mike,
> 
> I think you are right  regarding the access tags.
> The current solution in combination with the default style
> might route a truck through the wrong direction of a oneway.
> Thanks for pointing this out.
> 
> If I got it right, most style developers don't use this option,
> they prefer to have the logic in the rules.
> Maybe this is another argument to remove the option
> instead of adding code to make it work in special cases?
> 
> Is anybody able to change the default style so that it produces
> the same result ?
> 
> Gerd
> From: 
 
> mike at .co
 
>  &lt;mailto:
 
> mike at .co
 
> &gt; 
> To: 
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt; 
> Date: Mon, 2 Mar 2015 22:46:07 +0000
> Subject: [mkgmap-dev] --make-opposite-cycleways option
> Dear all,
>  
> When using the --make-opposite-cycleways option, I notice that the address
> index adds a road name of the actual road name suffixed with "
> (cycleway)".
> For me at least, this is undesirable. Also the generated name does not
> conform to the OSM naming convention of not including any descriptive
> information in names. I therefore propose that the name of an opposite
> cycleway be set to the same as the original road name, and that users who
> want to name the opposite cycleway differently should use the style file
> to
> rename it.
>  
> In looking at the code that produces the opposite cycleway, I also believe
> there is an error in the way it handles access tags - the code currently
> sets access=no, bicycle=yes and foot=no, but ignores all other access
> tags.
> Hence if (for example) the road has psv=yes on it, the generated cycleway
> will also have psv=yes (unless I am misunderstanding something in the
> handling of access tags). 
>  
> I therefore submit the attached patch for trial.
>  
> Regards,
> Mike
> 
> _______________________________________________ mkgmap-dev mailing list
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt;
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
> 
> _______________________________________________ mkgmap-dev mailing list
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt;
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
> 
> _______________________________________________ mkgmap-dev mailing list
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt;
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
> 
> _______________________________________________ mkgmap-dev mailing list
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt;
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
> 
> _______________________________________________ mkgmap-dev mailing list
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt;
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
> 
> _______________________________________________ mkgmap-dev mailing list
 
> mkgmap-dev at .org
 
>  &lt;mailto:
 
> mkgmap-dev at .org
 
> &gt;
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
> 
> _______________________________________________
> mkgmap-dev mailing list
 
> mkgmap-dev at .org
 
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
 
 
 
 
 
--
View this message in context:
http://gis.19327.n5.nabble.com/make-opposite-cycleways-option-tp5835586p5836
256.html
Sent from the Mkgmap Development mailing list archive at Nabble.com.
 

_______________________________________________
mkgmap-dev mailing list
mkgmap-dev at lists.mkgmap.org.uk
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.mkgmap.org.uk/pipermail/mkgmap-dev/attachments/20150310/727b13e2/attachment-0001.html>


More information about the mkgmap-dev mailing list