[mkgmap-dev] [Fwd: Re: Suggestions for roadNameConfig.txt]

Thu Apr 16 09:35:30 BST 2020

Hi Gerd

Could you consider this change to roadNameConfig.txt for GBR

I've attached an updated version based current trunk.o1

Thank you
Ticker

-------- Forwarded Message --------
From: Ticker Berkin <rwb-mkgmap at jagit.co.uk>
Reply-to: Development list for mkgmap <mkgmap-dev at lists.mkgmap.org.uk>
To: Development list for mkgmap <mkgmap-dev at lists.mkgmap.org.uk>
Subject: Re: [mkgmap-dev] Suggestions for roadNameConfig.txt
Date: Thu, 23 Jan 2020 16:56:17 +0000

Hi Gerd

I've just been going through some old mails/patches.

I think the updated contents of roadNameConfig.txt is correct and
better than the previous version.

The fact that it doesn't work consistently across devices and, on some,
is misleading shouldn't stop the file being available as an example in
the distribution.

The user has the choice about how to use it and it contains some
comments and a link relating to these problems.

I've attached an updated patch with UTF-8 encoding markers, a different
section for UK road names and improved some comments.

This file has always been read by mkgmap as UTF-8. The function of the
markers is to make editors aware of this; it contains characters
outside the 7-bit ascii range. 

Ticker

On Mon, 2019-12-16 at 11:30 +0000, Gerd Petermann wrote:
> Hi Ticker,
> 
> reg. search: You are probably right, address search works very
> different on the Garmin devices. On the Oregon, I have several
> dialogs, first to chose country, next to chose city or "all", next to
> type house number, last one asking for road name. In this dialog
> there is no autocompletion, I just have to type characters. Sometimes
> the device switches automatically to a result list while I am typing,
> sometimes it doesn't. When I press OK I get a list of matching roads.
> So, for the Oregon it works fine. Maybe it doesn't for those devices
> with auto-completion. Cannot test this right now. I tried it once
> with a Nüvi and it seemed to work well, but I don't own one.
> 
> Reg. combination of   --road-name-config and --split-name-index: Yes,
> this gives confusing results. Not sure if it worked better in the
> past.
> I tried with Essex extract:
> java -jar d:\mkgmap\dist\mkgmap.jar --bounds=f:\osm\bounds.zip -
> -gmapi --net --index  --road-name
> -config=d:\mkgmap\resources\roadNameConfig.txt --gmapsupp --split
> -name-index f:\dwnload\temp\essex-latest.osm.pbf
> 
> 
> According to Mapsource a search for "Mace" without giving a city name
> should list "John Mace Road", "Mace Avenue", "Mace Court", and "Mace
> Walk", maybe in a different order.
> Without your patch I  get a list showing "Mace Road", "Mace Avenue",
> and "Mace Walk", so "Mace Court" is missing and when I select "Mace
> Road" the device shows the "John Mace Road".
> With your patch this gets worse: The list shows only "Mace Road",
> when I select that entry I get a new list containing
> ",Thurrock","Braintree,Essex","Chelmsford,Essex", and "Colchester,
> Essex". Each of them points to just one of the expected roads.
> 
> Results look much better when I specify a city name.
> Result also looks very different without the --split-name-option.
> I have to do more tests, I think it worked better in the past. Maybe
> sorting is broken.
> 
> Gerd

-------------- next part --------------
A non-text attachment was scrubbed...
Name: roadNameEN_v3.patch
Type: text/x-patch
Size: 2989 bytes
Desc: not available
URL: <http://www.mkgmap.org.uk/pipermail/mkgmap-dev/attachments/20200416/671780c0/attachment.bin>