Hi Gerd I've just built a map with --split-name-index and --road-name-config as below, using latest trunk.
On my old eTrex Legend, I still get the behaviour as described earlier. I didn't use --split-name-index before, but it seems to behave as expected - all words except the suffix words are in the street list and can be found quickly and the result list then shows the full name. On my new eTrex 30x the behaviour is slightly different and worse. In the example from before, if I enter 'Xx' into the Street part of the address search, the selection list shows just "Xx Avenue". If I select this, the result list shows all addresses in Xx Avenue/Close/Way... I don't think this is relevant, but none of these streets have house number information and I've used option --no-housenumbers. The result list has 1 entry per street. Regards Ticker On Tue, 2019-02-19 at 18:21 +0000, Gerd Petermann wrote: > What you describe would be an error in my understanding. Can you > still reproduce it? > > Gerd > > ________________________________________ > Von: mkgmap-dev <mkgmap-dev-boun...@lists.mkgmap.org.uk> im Auftrag > von Ticker Berkin <rwb-mkg...@jagit.co.uk> > Gesendet: Dienstag, 19. Februar 2019 19:06 > An: Development list for mkgmap > Betreff: Re: [mkgmap-dev] Sample basic mkgmap config file > > Hi Gerd > > I experimented with --road-name-config=../roadNameConfig.txt just > after > it was committed to trunk a few months ago, using my old eTrex Legend > HCx. Nothing strange in my style. > > roadNameConfig.txt included: > > # english > #no prefix1:en = "East ", "North ", "South ", "West " > suffix:en = " Avenue", " Close", " Court", " Crescent", " Drive", > "Gardens", " Gate", " Grove", " Lane", " Mews", " Parade", " Park", > "Passage", " Place", " Rise", " Road", " Square", " Street", > "Terrace", > " View", " Walk", " Way", " Yard" > ... > lang:GBR = en > > Around where I live there are "Xx Road", "Xx Avenue", "Xx Close", > "Xx > Mews" and "Xx Way" > > Doing Find > Address, I select the "Region" (=country), enter the > "City" and, into the "Street" field, enter "X" and it shows the list > of > matching streets, just "Xx". Then entering a "Number" (or clearing > the > Number field to show all addresses in the street) it shows, in a > single > list, all the matching addresses in Xx Road, Avenue, Close, Mews and > Way. > > It is much quicker and I consider it more natural, to select the > particular Street in the "Street" stage (from the small, presented > list > of 5 in this case) and then be shown the matching addresses in just > this street. > > Does this make sense? > > Ticker > > > On Tue, 2019-02-19 at 17:26 +0000, Gerd Petermann wrote: > > Hi Ticker, > > > > I don't understand what you mean with > > "Removing all the common suffixes (road/street/etc) to the second > > stage > > of address searching just makes finding an address awkward and > > unnatural. It is normal to select the full street in 1 stage then > > the > > locations within the street in the next stage." > > Is that something that you do in your style? > > > > Gerd > > > > ________________________________________ > > Von: mkgmap-dev <mkgmap-dev-boun...@lists.mkgmap.org.uk> im Auftrag > > von Ticker Berkin <rwb-mkg...@jagit.co.uk> > > Gesendet: Dienstag, 19. Februar 2019 12:32 > > An: Development list for mkgmap > > Betreff: Re: [mkgmap-dev] Sample basic mkgmap config file > > > > Hi Gerd > > > > --split-name-index and --road-name-config behave as expected, but: > > > > UK roads are always indexed by their first word, even when the word > > is > > 'The'. > > > > Removing all the common suffixes (road/street/etc) to the second > > stage > > of address searching just makes finding an address awkward and > > unnatural. It is normal to select the full street in 1 stage then > > the > > locations within the street in the next stage. > > > > My opinion is that using --order-by-decreasing-area and then, if > > you > > have a typ-file, setting all [_drawOrder] to the same value, gives > > a > > good map, whereas attempting to pre-define which polygons overwrite > > others using [_drawOrder] will be wrong in some cases. > > > > Ticker > > > > On Thu, 2019-02-14 at 17:35 +0000, Gerd Petermann wrote: > > > Hi Ticker, > > > > > > thanks for the comments. > > > - I forgot route, that was not intended. > > > - no-tdbfile and nsis are a bad combination, and I see no problem > > > when we generate all possible formats > > > unless one starts to create really large maps. > > > - I am not aware of problems with --split-name-index or --road > > > -name > > > -config, please give examples in a new thread. Besides that > > > I see no problems to remove those. > > > - I agree that --add-pois-to-lines is a rather problematic > > > option, > > > but I see no need to add no-add-pois-to-lines > > > - My understanding is that we don't need order-by-decreasing-area > > > once we have a default typ? > > > > > > Gerd > > > > > > > > > ________________________________________ > > > Von: mkgmap-dev <mkgmap-dev-boun...@lists.mkgmap.org.uk> im > > > Auftrag > > > von Ticker Berkin <rwb-mkg...@jagit.co.uk> > > > Gesendet: Donnerstag, 14. Februar 2019 18:19 > > > An: Development list for mkgmap > > > Betreff: Re: [mkgmap-dev] Sample basic mkgmap config file > > > > > > Hi > > > > > > I have a few suggestions / comments: > > > > > > I don't agree with --recommended because this file will often > > > required > > > some tweaking and eventually become the basis of the new users > > > building > > > environment. > > > > > > I think forward slash (/) works as a directory seperator on > > > DOS/Windows > > > so can always use this. > > > > > > It should be written to assume various items are in the current > > > directory, including the .cfg file itself, the outputs from > > > splitter, > > > bounds.zip, sea.zip and mkgmap release subdirectory > > > > > > It should be aimed at generating a GMAPSUPP.IMG file to put on a > > > Garmin > > > device. All the PC bits for basecamp/mapsource are a distraction > > > and > > > a > > > great deal of time can be wasted trying to get these installed on > > > a > > > PC > > > and then trying to use them install the map image on a device; so > > > shouldn't have 'gmapi' and should have 'no-tdbfile' > > > > > > Should specify: code-page=1252 > > > > > > I'm not convinced of the general benefits of 'split-name-index' > > > and > > > 'road-name-config'. They don't give good behaviour for the UK > > > > > > The 'bounds' option to location-autofill doesn't do anything > > > > > > No need to specify the style - it will use 'default' > > > > > > name-tag-list=name:XX,int_name,name,place_name,loc_name is a > > > useful > > > option where XX is the map user's language code > > > > > > Should specify: route > > > > > > I don't think 'check-roundabouts' is of any benefit for a new > > > user > > > > > > Option 'add-pois-to-lines' generates a lot of 'not very useful' > > > POIs > > > and I think it is better turned off. This is a topic that could > > > be > > > re > > > -visited in 'default style improvements' > > > > > > 'make-opposite-cycleways' is wrong for a lot of countries > > > > > > I'd specify: 'order-by-decreasing-area' as the way to get a map > > > that > > > shows polygon features overlayed in the best order, without > > > having > > > to > > > make fixed arbitrary decisions about all the [_drawOrder] levels > > > in > > > a > > > typ-file. > > > > > > Following, inline, is a version with some changes - feel free to > > > take/ignore whatever you think is best. It should probably be > > > more > > > heavily commented. > > > > > > Regards > > > Ticker > > > > > > # > > > # Set options to create a routable map for a Garmin GPS > > > # Assumes that the files bounds.zip and sea.zip exist > > > # > > > gmapsupp > > > code-page=1252 > > > no-tdbfile > > > nsis > > > index > > > #split-name-index > > > #road-name-config=mkgmap/examples/roadNameConfig.txt > > > bounds=bounds.zip > > > location-autofill=is_in,nearest > > > housenumbers > > > #name-tag-list=name:en,int_name,name,place_name,loc_name > > > max-jobs > > > route > > > drive-on=detect > > > no-add-pois-to-lines > > > add-pois-to-areas > > > precomp-sea=sea.zip > > > #make-opposite-cycleways > > > link-pois-to-ways > > > process-destination > > > process-exits > > > order-by-decreasing-area > > > > > > #that's it > > > > > > On Thu, 2019-02-14 at 14:29 +0100, Andrzej Popowski wrote: > > > > Hi Gerd, > > > > > > > > maybe treat it like default style, which is kind of build in? > > > > For > > > > example, when invoking mkgmap with an option --recommended, > > > > mkgmap > > > > could > > > > include all options from default config. > > > > > > > _______________________________________________ > > > mkgmap-dev mailing list > > > mkgmap-dev@lists.mkgmap.org.uk > > > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > > _______________________________________________ > > mkgmap-dev mailing list > > mkgmap-dev@lists.mkgmap.org.uk > > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev > _______________________________________________ > mkgmap-dev mailing list > mkgmap-dev@lists.mkgmap.org.uk > http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
