[GraphHopper] Bike way type conversion

[Peter]

Hey Stuart,

thanks - would be awesome and very appreciated!

Regards,
Peter

On 05.06.2015 12:39, Stuart Adam wrote:
> Hello Peter
>
> I am just awaiting a corporate decision on the contributor license and
> then I will see what I can do.
>
> Sincerely
> Stuart Adam
>
> On 3 Jun 2015, at 08:12, Peter <graphhopper at gmx.de
> <mailto:graphhopper at gmx.de>> wrote:
>
>> Hi Stuart,
>>
>> for me this looks also okay. So what changed is e.g. isPushingSection
>> && bicycle=designated => before CYCLEWAY, now OTHER_SMALL_WAY?
>>
>> Maybe just create a pull request and we'll see e.g. what the tests
>> (and ratrun) says ;)
>>
>> Regards,
>> Peter
>>
>> On 02.06.2015 17:29, Stuart Adam wrote:
>>> Hello ratrun
>>>
>>> I can understand the usage of a small amount of flags however I
>>> would have thought logic more akin to the following.  This would
>>> only highlight something as a cycleway if it is intentionally marked
>>> as such rather than merely having a right of way.
>>>
>>> Edited in the email so apologies for any formatting issues.
>>>
>>> if (way.hasTag("bicycle", intendedValues))
>>> {
>>> if(isPushingSection)
>>> wayType = WayType.OTHER_SMALL_WAY
>>> else if (“cycleway”.equals(highway)) 
>>> wayType = WayType.CYCLEWAY;
>>> else if (way.hasTag(“bicycle”,”designated) 
>>>     wayType = WayType.CYCLEWAY;
>>>     else if (roadValues.contains(highway))
>>> wayType = WayType.ROAD;
>>>
>>> Any thoughts?
>>>
>>> Sincerely
>>> Stuart Adam
>>>
>>>
>>> On 2 Jun 2015, at 16:10, ratrun <ratrun at gmx.at
>>> <mailto:ratrun at gmx.at>> wrote:
>>>
>>>> Hello Stuart,
>>>>
>>>> the waytype information is just used for the routing instructions
>>>> such that a cycle rider gets a better idea what kind of way to look
>>>> at. As the bits are limited and a more exact infomration is
>>>> superflous for that purpose, I intentionally mangled all kind of
>>>> ways somehow marked for bicycle usage together into "CYCLEWAY".
>>>>
>>>> I also had some code which produced a statistic of the tour. It
>>>> calcualted the distances per each paved and unpaved waytype. This
>>>> code didn't make it into the master, although I think that it is an
>>>> important feature for bicycle routing. The problem was that the
>>>> code was too bicycle specific and required changes in all other
>>>> flag encoders as well - see issue #209.
>>>>
>>>> ratrun
>>>>
>>>> Am 02.06.2015 um 16:17 schrieb Stuart Adam:
>>>>> Hello all
>>>>>
>>>>> I am starting to look at bike routing and I note in the
>>>>> handleBikeRelated method in BikeCommonFlagEncoder the following
>>>>> logic applies.
>>>>>
>>>>>  if (way.hasTag("bicycle", intendedValues))
>>>>>         {
>>>>>             if (isPusingSection && !way.hasTag("bicycle",
>>>>> "designated"))
>>>>>                 wayType = WayType.OTHER_SMALL_WAY;
>>>>>             else
>>>>>                 wayType = WayType.CYCLEWAY;
>>>>>         } else if ("cycleway".equals(highway))
>>>>>             wayType = WayType.CYCLEWAY;
>>>>>         else if (roadValues.contains(highway))
>>>>>             wayType = WayType.ROAD;
>>>>>
>>>>>
>>>>> This does not seem correct to me as from my understanding this is
>>>>> taking the fact that a way has been marked as having a right of
>>>>> way for bicycles (but not a pushing section) then it is a full
>>>>> blown cyclepath.  In my mind at least cycleway implies dedicated
>>>>> and marked (normally with differently coloured tarmac) sections
>>>>> which is a much stronger indication for cycle use than just a
>>>>> bicycle right of way.
>>>>>
>>>>>
>>>>> Am I correct and if so should this be changed or was there a
>>>>> reason for this decision in Graphhopper.
>>>>>
>>>>> Sincerely
>>>>> Stuart Adam
>>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/graphhopper/attachments/20150605/b75c40a3/attachment.html>