Hi all, thanks for the input. I think we have at least two different issues here. 1) Some styles add attributes like "Path", "unpaved" etc. to the road name. Those attributes are obsolete in the index, you don't want to search for them and they should not appear in result lists. If that is true we should try to find a method to remove all those attributes from the name before it is written to the index. I can think of various ways to do that: 1a) While name is not empty: check if last word is in exclude list, if yes, remove it, if not, exit loop. 1b) For each word in name: check if it appears in the exclude list, if so, remove it including the rest. 1c) Use a special character sequence (like the highway shield codes) to separate the real name from the rest If the name is empty after such processing it would not be added to the index. 1a) and 1b) may need special code for unicode maps when name is written from right to left (hebrew for example). No idea if mkgmap works at all for those names.
2) The other problem is that road names in some languages have lots of rather meaningless words, either as prefix (Rua, Rue, Chemin) or as suffix (Street, Road, Weg). My understanding is that we don't want to remove those words from the name. The index should at least contain the full name, but we don't all the variations produced by the current implementation of the --x-split-name-index option. 3) A road has up to four different labels, the processing happens for each label. In some cases the --housenumber option adds or overwrites a label to make sure that an address is found. The default style creates quite often three labels like this: label1: "Ahlhorner Straße [B 213]" label2: "B 213" label3: "Ahlhorner Straße" I am not sure if we really need all the index entries which are now produced for this. 4) @Felix: Your ideas to put names of POI in different languages into the index are a bit off-topic as they don't appear in the mdr7 index which is only for roads. Not sure if other indexes allow those partial names. I guess no because Steve did not implement the --x-split-name-index for them. Gerd ________________________________________ Von: mkgmap-dev <[email protected]> im Auftrag von Carlos Dávila <[email protected]> Gesendet: Mittwoch, 5. April 2017 08:34:30 An: Development list for mkgmap Betreff: Re: [mkgmap-dev] Meaning of option --x-mdr7-excl El 04/04/17 a las 20:34, Andrzej Popowski escribió: > Hi Gerd, > > I think the most useful version would be to drop last word from name. > > If you would like to process prefixes too, then maybe limit it to > first word only. For name like "Chemin de Pierre Froide", you would > remove "Chemin" but not "de". Alternatively you could implement > prefixes with multiple words, like for example: > > --x-mdr7-excl="chemin,chemin de" > I don't find much use in removing only the first word. The probability that someone searches "Chemin de Pierre Froide" typing "de Pierre Froide" is very low. Using multiple words would be much more useful, but the list may be really large. _______________________________________________ mkgmap-dev mailing list [email protected] http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev _______________________________________________ mkgmap-dev mailing list [email protected] http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
