Re: [OSM-talk] [HOT] Nametagging: Local script versus
On 03/03/2011 02:19, andrzej zaborowski wrote: On 28 February 2011 21:40, Steve Doerrsteve.do...@blueyonder.co.uk wrote: On 28/02/2011 16:55, Ed Avis wrote: If the name is still in Arabic, but Arabic written with the Latin alphabet, then name:ar@Latin would be correct. Did you just make that up, or is this use of the @ symbol a pre-existing standard? The part after @ is the modifier in posix locales and is often used for script type, see for example http://wiki.services.openoffice.org/wiki/LocaleMapping Looks like a good scheme. Conflicts slightly with the suggestion at http://wiki.openstreetmap.org/wiki/Key:name, under 'Localization', of using name:ko_rm for romanized Korean. Presumably name@Latin would also be desirable? -- Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus
andrzej zaborowski wrote: On 28 February 2011 21:40, Steve Doerr steve.do...@blueyonder.co.uk wrote: On 28/02/2011 16:55, Ed Avis wrote: Jean-Marc Liotierjmat liotier.org writes: By the way, for latin script names, should we use int_name, name:en or both ? If the name is still in Arabic, but Arabic written with the Latin alphabet, then name:ar@Latin would be correct. Did you just make that up, or is this use of the @ symbol a pre-existing standard? The part after @ is the modifier in posix locales and is often used for script type, see for example http://wiki.services.openoffice.org/wiki/LocaleMapping Citation from above link : @modifiers that indicate a non-default script, e.g. uz_UZ@cyrillic, be_BY@latin. Reusing this standard sounds like a great idea to me - much more flexible than my earlier proposal of recycling int_name in this role. Anyone else in favor of name:ar@latin for the romanized version of the local Arabic name ? ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus
2011/3/3 Jean-Marc Liotier j...@liotier.org: Anyone else in favor of name:ar@latin for the romanized version of the local Arabic name ? +1 cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus
On 03/03/2011 12:38, M∡rtin Koppenhoefer wrote: 2011/3/3 Jean-Marc Liotierj...@liotier.org: Anyone else in favor of name:ar@latin for the romanized version of the local Arabic name ? +1 with the caveat that name@latin is probably just as important if not more so (renderers can more easily know where to find the romanized form of the local name that way). -- Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus
On 28 February 2011 21:40, Steve Doerr steve.do...@blueyonder.co.uk wrote: On 28/02/2011 16:55, Ed Avis wrote: Jean-Marc Liotierjmat liotier.org writes: By the way, for latin script names, should we use int_name, name:en or both ? If the name is still in Arabic, but Arabic written with the Latin alphabet, then name:ar@Latin would be correct. Did you just make that up, or is this use of the @ symbol a pre-existing standard? The part after @ is the modifier in posix locales and is often used for script type, see for example http://wiki.services.openoffice.org/wiki/LocaleMapping Cheers ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
Considering the difficulties I had getting florists to display on a map I would technically agree with you but add having the data in OSM means you just need a custom rule set to display rather than setting up a preprocessor for the renderer. Cheerio John On 1 March 2011 02:39, Stephan Knauss o...@stephans-server.de wrote: On 01.03.2011 00:42, john whelan wrote: I have a couple of basic VB programs that can take a local .OSM file and add a name:fr field to street names, it also does a little look up to translate the street name type. You just take the output and feed it into JOSM and upload the changes. If a script can do it, then there is no need to upload it to the DB. It can simply be a preprocessing step for the renderer. Stephan ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
Le 28/02/2011 01:00, Stephan Knauss a écrit : On 27.02.2011 21:14, Jean-Guilhem Cailton wrote: Maybe I can convince german community member Stephan Knauss to join the HOT team as he set up great bilingual maps like http://thaimap.osm-tools.org or http://iran.osm-tools.org Would it help to have bilingual rendering? I think porting a map similar the Iran one should be done quite easy. It already has bilingual names and font adjustments to improve readability of the map. Currently the minutely diff update is tailored to the thaimap. With some rework I guess it should be possible to enable the minutely diff also for Libya. Until then a manual daily update is the thing I can offer. Would this be of any help? Please let me know. Stephan Of course, this would help. Please go, ahead. The progressive deployment and generalization of such tools will facilitate proper handling of several languages. It would not solve the problem for all tools, of course. For example, MapOSMatic of Tripoli apparently uses the name field. So my recommendation, given the feedback received so far, would still be to use Latin + Arabic in the name field. Regarding readability of the map, I'd like to raise the issue of place definition and rendering (city/town/village...) The global convention is based on uniform population numbers. ( 10 / 1 / ...). While these may be well suited for densely populated countries, like United Kingdom or Germany, coupled with the current renderings, it produces maps that look empty (compared with other providers) in other countries. Typically in Libya, with its large desert areas (but also in France, for example). For roads, for example, the convention can be adapted to the local context. An unpaved surface is not considered a road in GB, for example, but in other countries this might be the standard surface. Similarly, it might be appropriate to define local conventions (on a country by country basis, for example) for the urban fabric. Maybe your tool, specific to a country, can make it possible to experiment on the rendering of places. In the meantime, maybe the global definition could be applied to Libya with some flexibility. For example, a city of slightly less than 100 000 inhabitants, or a town of slightly less than 10 000 inhabitants, but isolated in the desert, might be bulged to the upper OSM standard category, (which would then actually better match the semantics). It would thus be given a chance to appear on the radar, which might not be the case otherwise. At least, this is what I confess of doing sometimes. The population field could of course be a useful information for this issue. Best wishes, Jean-Guilhem pgp 0x5939EAE2 ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
Jean-Guilhem Cailton wrote: It would not solve the problem for all tools, of course. For example, MapOSMatic of Tripoli apparently uses the name field. So my recommendation, given the feedback received so far, would still be to use Latin + Arabic in the name field. Do that if it is the most expedient way to satisfy the requirements of users... But isn't it recommended to also use the tags that will let people generate maps in a specific language ? By the way, for latin script names, should we use int_name, name:en or both ? Regarding readability of the map, I'd like to raise the issue of place definition and rendering (city/town/village...) The global convention is based on uniform population numbers. ( 10 / 1 / ...). While these may be well suited for densely populated countries, like United Kingdom or Germany, coupled with the current renderings, it produces maps that look empty (compared with other providers) in other countries. Typically in Libya, with its large desert areas (but also in France, for example). Tagging should not be done with rendering in mind, but the relative importance of the place in the urban hierarchy must be taken into account. Documents showing the administrative hierarchy may provide useful hints, though administrative status is not always correlated to actual importance to the local social fabric. For roads, for example, the convention can be adapted to the local context. An unpaved surface is not considered a road in GB, for example, but in other countries this might be the standard surface. Very important indeed in Africa where the topological hierarchy in the road network is the most important factor to consider. In Libya in particular, I have seen a lot of residential, tertiary and even secondary roads with unpaved surfaces. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
On 28/02/2011 10:33, Jean-Marc Liotier wrote: Jean-Guilhem Cailton wrote: It would not solve the problem for all tools, of course. For example, MapOSMatic of Tripoli apparently uses the name field. So my recommendation, given the feedback received so far, would still be to use Latin + Arabic in the name field. Do that if it is the most expedient way to satisfy the requirements of users... But isn't it recommended to also use the tags that will let people generate maps in a specific language ? By the way, for latin script names, should we use int_name, name:en or both ? In my opinion, the int_name can usefully be used for the direct romanization of the Arabic name, as it now (as of yesterday) is for Tripoli, i.e. Ṭarābulus (previously, it said 'Tripoli'). http://www.openstreetmap.org/browse/node/27564957/history. -- Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
Le 28/02/2011 11:33, Jean-Marc Liotier a écrit : Jean-Guilhem Cailton wrote: It would not solve the problem for all tools, of course. For example, MapOSMatic of Tripoli apparently uses the name field. So my recommendation, given the feedback received so far, would still be to use Latin + Arabic in the name field. Do that if it is the most expedient way to satisfy the requirements of users... But isn't it recommended to also use the tags that will let people generate maps in a specific language ? Of course, that goes without saying. (But thanks for reminding it). By the way, for latin script names, should we use int_name, name:en or both ? name:en has the advantage of specifying the language, and can thus also be used by language specific renderings (or not used, depending on their policy). I would also use int_name only if it differs from name:en (can't think of an example right now), or if what is meant is really Internationally known as. Of course, if they are different and if your question is about the Latin part in name tag, I guess int_name should have precedence, in the framework of international humanitarian intervention. What do you think? Steve, (about Ṭarābulus) that's an idea, though it does not fit the wiki definition. At least, that's a way of storing it. A name:xyz where xyz would mean romanized Arabic would be better, in principle. (But there does not seem to be a recognized code for it). Is int_name used by any renderer? Jean-Guilhem pgp 0x5939EAE2 ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
Jean-Guilhem Cailton wrote: I would also use int_name only if it differs from name:en (can't think of an example right now), or if what is meant is really Internationally known as. Of course, if they are different and if your question is about the Latin part in name tag, I guess int_name should have precedence, in the framework of international humanitarian intervention. What do you think? Is int_name used by any renderer? Reading http://taginfo.openstreetmap.de/keys/int_name#values suggests that what international name actually means is the romanized version of the local name - which may differ from the English name. Romanized version of the local name makes more sense to me than just international name. Anyone else here who thinks that is would make a better definition of int_name ? Steve, (about Ṭarābulus) that's an idea, though it does not fit the wiki definition. At least, that's a way of storing it. A name:xyz where xyz would mean romanized Arabic would be better, in principle. (But there does not seem to be a recognized code for it). Which romanization standard should be used ? http://en.wikipedia.org/wiki/Romanization_of_Arabic#Comparison_table It looks like no standard dominates. Let's hope we'll find a way to avoid OSM edit wars about romanized Arabic... ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
Steve Doerr wrote: In my opinion, the int_name can usefully be used for the direct romanization of the Arabic name, as it now (as of yesterday) is for Tripoli, i.e. Tarābulus (previously, it said 'Tripoli'). Looks like we posted the same thing at the same time... And you beat me to it... Anyway, this looks like the beginning of a consensus. If that holds, I'll edit the wiki to mention romanized version of the local name as the definition of the int_name tag. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
Le 28/02/2011 13:27, Jean-Marc Liotier a écrit : Reading http://taginfo.openstreetmap.de/keys/int_name#values suggests that what international name actually means is the romanized version of the local name - which may differ from the English name. Romanized version of the local name makes more sense to me than just international name. Anyone else here who thinks that is would make a better definition of int_name ? I do. And also it seems it would be more used, and useful, in practice. Steve, (about Ṭarābulus) that's an idea, though it does not fit the wiki definition. At least, that's a way of storing it. A name:xyz where xyz would mean romanized Arabic would be better, in principle. (But there does not seem to be a recognized code for it). Which romanization standard should be used ? http://en.wikipedia.org/wiki/Romanization_of_Arabic#Comparison_table It looks like no standard dominates. Let's hope we'll find a way to avoid OSM edit wars about romanized Arabic... Let's just discuss it here if the need appears. :) Jean-Guilhem pgp 0x5939EAE2 ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
2011/2/28 Steve Doerr steve.do...@blueyonder.co.uk: In my opinion, the int_name can usefully be used for the direct romanization of the Arabic name, as it now (as of yesterday) is for Tripoli, i.e. Ṭarābulus (previously, it said 'Tripoli'). http://www.openstreetmap.org/browse/node/27564957/history. until now, the int_name was defined as Internationally known as e.g. # name=Channel Tunnel - Default name # int_name=Eurotunnel - Internationally known as from: http://wiki.openstreetmap.org/wiki/Key:int_name therefor I feel that using Ṭarābulus instead of Tripoli is not according to the general usage of int_name, because Tripoli is IMHO not internationally known as 'Ṭarābulus' but I might be wrong. cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
2011/2/28 Jean-Marc Liotier j...@liotier.org: Steve Doerr wrote: In my opinion, the int_name can usefully be used for the direct romanization of the Arabic name, as it now (as of yesterday) is for Tripoli, i.e. Tarābulus (previously, it said 'Tripoli'). Anyway, this looks like the beginning of a consensus. If that holds, I'll edit the wiki to mention romanized version of the local name as the definition of the int_name tag. Generally breaking all applications in use by redefining the meaning of a tag is not a sustainable approach. If you want to introduce a romanized version of the local name why don't you simply introduce a new tag for it? Have a look at current int_name-values to see that romanized version of the local name is not the predominant current use. cheers, Martin ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
M∡rtin Koppenhoefer wrote: 2011/2/28 Steve Doerr steve.do...@blueyonder.co.uk: In my opinion, the int_name can usefully be used for the direct romanization of the Arabic name, as it now (as of yesterday) is for Tripoli, i.e. Ṭarābulus (previously, it said 'Tripoli'). http://www.openstreetmap.org/browse/node/27564957/history. until now, the int_name was defined as Internationally known as e.g. # name=Channel Tunnel - Default name # int_name=Eurotunnel - Internationally known as from: http://wiki.openstreetmap.org/wiki/Key:int_name therefor I feel that using Ṭarābulus instead of Tripoli is not according to the general usage of int_name, because Tripoli is IMHO not internationally known as 'Ṭarābulus' but I might be wrong. Yes, but http://en.wikipedia.org/wiki/Tarabulus_District mentions Tarabulus District (Arabic: شعبية طرابلس, Shab'iyat Ṭarābulus, English: Tripoli District). So Wikipedia does distinguish between Tripoli as an English name and Tarabulus as its canonical name. But at the same time, the page for the capital is http://en.wikipedia.org/wiki/Tripoli - so the issue is not clear cut. This one is an interesting example because Tarabulus is the internationaly used transliteration of the local name while Tripoli is fundamentally international as it comes from the Greek Τρίπολις (Trípolis) which means Three Cities. This does not help us to settle the issue easily... We might end-up having to keep int_name for the internationally used name, even if it is a somewhat fuzzy definition - and then add a name:romanized or romanized_name tag for the romanized transliteration of the local name. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
M∡rtin Koppenhoefer wrote: Generally breaking all applications in use by redefining the meaning of a tag is not a sustainable approach. Yes, I have been too hasty in suggesting that. But it remains one of the possible options nevertheless... Subject to prior debate and consensus, of course. We need to take our time investigating the actual use of the int_name tag to assess the impact of altering the interpretation of that tag and decide if it is better left alone or not. If you want to introduce a romanized version of the local name why don't you simply introduce a new tag for it ? That is indeed the folksonomic democracy option: create a new tag and see if other adopt it too. It is a well known method in OpenStreetMap, but it carries the cost of introducing an extra tag and extra confusion in an already crowded space. That is why I leaned toward other options... But I don't deny it may end up being the right one. Have a look at current int_name-values to see that romanized version of the local name is not the predominant current use. I disagree about that assessment. I may be wrong since I don't know any of the non-latin languages that seem to be the source of many of the names who appear in values of int_name. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus
Jean-Marc Liotier jm at liotier.org writes: By the way, for latin script names, should we use int_name, name:en or both ? If the name is still in Arabic, but Arabic written with the Latin alphabet, then name:ar@Latin would be correct. (There may be some more precise specification of exactly which romanization scheme you use - see the 120 different Latin spellings of 'Gadaffi' - but just ar@Latin would probably be enough.) If the name is like 'Tripoli', a non-Arabic name, then name:en (plus in this case name:it) or int_name would be appropriate. -- Ed Avis e...@waniasset.com ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
Hi, On 28.02.2011 17:00, Jean-Marc Liotier wrote: We might end-up having to keep int_name for the internationally used name, even if it is a somewhat fuzzy definition - and then add a name:romanized or romanized_name tag for the romanized transliteration of the local name. I can't understand Arabian. I'm no expert. Is there a way to do the transliteration automatically in a consistent way? Acconding to wikipedia there exist multiple ways to do so: http://en.wikipedia.org/wiki/Arabic_transliteration If it is possible in a consistent way then there is no need to store this in the database. A map wanting to display it can easily calculate the romanized form. Stephan ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus
On 28/02/2011 16:55, Ed Avis wrote: Jean-Marc Liotierjmat liotier.org writes: By the way, for latin script names, should we use int_name, name:en or both ? If the name is still in Arabic, but Arabic written with the Latin alphabet, then name:ar@Latin would be correct. Did you just make that up, or is this use of the @ symbol a pre-existing standard? -- Steve ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
Stephan Knauss wrote: Is there a way to do the transliteration automatically in a consistent way ? According to wikipedia there exist multiple ways to do so: http://en.wikipedia.org/wiki/Arabic_transliteration For each of the transliteration standards, an automated method is probably possible... But Wikipedia cites sixteen different standards, some of them with different focuses, such as phonetic expressiveness, spelling, simplicity, lack of diacritics or teaching. I have not read about any of those methods being more authoritative, so I guess that we have to rely on user input for each named object. On top of all that, Arabic has local variations, sometimes strong and always co-existing with literary Arabic... So I guess that there won't be a substitute to user wisdom. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
hi We might end-up having to keep int_name for the internationally used name, even if it is a somewhat fuzzy definition - and then add a name:romanized or romanized_name tag for the romanized transliteration of the local name. I can't understand Arabian. I'm no expert. Is there a way to do the transliteration automatically in a consistent way? Acconding to wikipedia there exist multiple ways to do so: http://en.wikipedia.org/wiki/Arabic_transliteration I'm also not able to read Arabic, but I'm able to read some Asian writing systems. Lets take Bangkok for example. The City is called: ? (/Krung Thep Mahanakhon)/ Most of the people there call it Krung Thep There is also a long (offical) name: ? ??? ??? ?? ??? ? ?? Lets cross the border to Laos In Laos Bangkok is also called Krung Thep. If people don't speak English, they don't know the name Bangkok. Laos has it's own writing system an the Lao transliteration of Bangkok is ??. Next country Cambodia - Khmer writing system: = krung thep ? = bangkok Sound file: http://en.wikipedia.org/wiki/File:Th-Krung_Thep.ogg --- 3 countries 3 writing systems. I don't see an international name. Maybe the name as written at the airport could be the int_name. name:? name_offical:? ??? ??? ?? ??? ? ?? name_en:Bangkok name@lo:?? name_lo:don't know (There is no R in Lao) name@km: name_km:? name@de:Krung Theep name@en:Krung Thep name_int:Bangkok name@ipa: kru-? t^(h)ê?p máha(? nák^(h)???n http://en.wikipedia.org/wiki/Wikipedia:IPA_for_Thai_and_Lao --- name:? (Bangkok) is definitly wrong. If I want to render a bilingual map it is not possible - except I make an English/Thai map. It is not possible to render a Thai/Lao map or a Thai/Chinese map. --- There isn't an offical transliteration system for every writing system. Laos and Cambodia have been French colonies. In Cambodia they have an official translitaration system based on english pronunciation. For me as a German speaking person the English based transliteration is nearly useless. In Laos they use more often the French sounds of Latin letters - but there is no official transliteration system. In Cambodia I have never seen a map using Khmer letters, but it is easy to buy a Cambodian map with English letters. The local people have no maps if they are not able to read English. The Khmer alphabet has 70 letters. Transliteration is not possible without losses. Mapnik is still not able to render Khmer fonts without mistakes. Bernhard ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
Hi Bernhard, On 28.02.2011 23:27, Bernhard Zwischenbrugger wrote: Lets take Bangkok for example. [...] lots of examples name:กรุงเทพมหานคร (Bangkok) is definitly wrong. If I want to render a bilingual map it is not possible - except I make an English/Thai map. It is not possible to render a Thai/Lao map or a Thai/Chinese map. --- There isn't an offical transliteration system for every writing system. great explanation why we need more than just a name tag. Some of the transliterations might be calculated. For the Thai map I thought of letting the map processor do an automatic RTGS. But even with RTGS being some form of standard, it's not that widely used. Most transliteration is based on the personal taste of the person designing the sign. At least it often looks like this. So the only sure thing is the local name. We put this into name. In Thailand I often repeat it into name:th. While not really needed it won't do any harm and might be useful for bilingual maps crossing country borders. Mapnik is still not able to render Khmer fonts without mistakes. Is this an issue with Mapnik or due to missing fonts? Stephan ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
I have a couple of basic VB programs that can take a local .OSM file and add a name:fr field to street names, it also does a little look up to translate the street name type. You just take the output and feed it into JOSM and upload the changes. As with all bots and such tools its best to have some one else verify the output before going mad with it. Maperitive can be set to render using one of the alternative name fields such as name:fr and I have an example if need be. Cheerio John On 28 February 2011 15:25, Stephan Knauss o...@stephans-server.de wrote: Hi, On 28.02.2011 17:00, Jean-Marc Liotier wrote: We might end-up having to keep int_name for the internationally used name, even if it is a somewhat fuzzy definition - and then add a name:romanized or romanized_name tag for the romanized transliteration of the local name. I can't understand Arabian. I'm no expert. Is there a way to do the transliteration automatically in a consistent way? Acconding to wikipedia there exist multiple ways to do so: http://en.wikipedia.org/wiki/Arabic_transliteration If it is possible in a consistent way then there is no need to store this in the database. A map wanting to display it can easily calculate the romanized form. Stephan ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
Mapnik is still not able to render Khmer fonts without mistakes. Is this an issue with Mapnik or due to missing fonts? Khmer Font exists but letter spacing is wrong and there are wrong rectangles around some letters. Lao rendering is ok, Thai fonts are too small, Myanmar fonts also have problems with letter spacing, the letter ြ for example is rendered too big - but I can't read Burmese. Bernhard ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
Hi all Thanks for having opened the thread. I feel that on the HOT side there will be a need to have name:en and name:ar which can be by default name as possible tags. Ciao Nicolas On Tue, Mar 1, 2011 at 12:42 AM, john whelan jwhelan0...@gmail.com wrote: I have a couple of basic VB programs that can take a local .OSM file and add a name:fr field to street names, it also does a little look up to translate the street name type. You just take the output and feed it into JOSM and upload the changes. As with all bots and such tools its best to have some one else verify the output before going mad with it. Maperitive can be set to render using one of the alternative name fields such as name:fr and I have an example if need be. Cheerio John On 28 February 2011 15:25, Stephan Knauss o...@stephans-server.de wrote: Hi, On 28.02.2011 17:00, Jean-Marc Liotier wrote: We might end-up having to keep int_name for the internationally used name, even if it is a somewhat fuzzy definition - and then add a name:romanized or romanized_name tag for the romanized transliteration of the local name. I can't understand Arabian. I'm no expert. Is there a way to do the transliteration automatically in a consistent way? Acconding to wikipedia there exist multiple ways to do so: http://en.wikipedia.org/wiki/Arabic_transliteration If it is possible in a consistent way then there is no need to store this in the database. A map wanting to display it can easily calculate the romanized form. Stephan ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk -- Nicolas Chavent Humanitarian OpenStreetMap Team http://wiki.openstreetmap.org/wiki/WikiProject_Haiti Mobile (Haiti): +509 389 583 05 Mobile (France): +33 6 89 45 54 58 Landline (FRA): +33 2 97 26 23 08 Email: nicolas.chav...@gmail.com Skype: c_nicolas Twitter: nicolas_chavent ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
On 01.03.2011 00:42, john whelan wrote: I have a couple of basic VB programs that can take a local .OSM file and add a name:fr field to street names, it also does a little look up to translate the street name type. You just take the output and feed it into JOSM and upload the changes. If a script can do it, then there is no need to upload it to the DB. It can simply be a preprocessing step for the renderer. Stephan ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
Hi, Thanks Claudius for bringing this up here. I am the Arabic-illiterate you got in touch with. I had also gotten in touch previously with Esperanza36 on this same subject. Also addressing this to the talk list, since some of the mappers involved in the current Libya effort may very well be only there, as well as some of the relevant technical competencies. In fact, my main point was that paper maps are essential in field deployments. GPS maps might also be useful. And of course, the standard OSM online renderings should not be forgotten. I think we all agree that, normally, name= data should be in the local language. The question of the tools has been raised recently. As far as I know, it does not seem to be perfectly solved for renderings on all devices yet, nor everywhere. For example, wikimedia tools server ( http://toolserver.org/~osm/locale/ ) provide online maps that use a chosen language when it is available. Unfortunately, from what I saw using it for Egypt, it seems that it is not updated very frequently, which makes it ill-suited for the current emergency mapping effort. In this crisis context, my personal, maybe naive, take at this has been to use both Arabic and Latin alphabets in the name tag when I added new features, including even new Libyan cities that were missing. They get rendered quickly in the standard OSM renderings, Mapnik and Osmarender, and my guess is that they also get printed properly in most likely settings. Esperanza36 mentioned problems regarding GPS maps with 2 alphabets. What are your experiences about this, regarding technical solutions and best compromise for usability in crisis context? Keeping in mind that a crisis is generally not the best time to deploy new tools. Maybe we should keep this focused on Libya for now, otherwise the situations might be so diverse that the discussion might become difficult to follow. If a consensus rule emerges, it would then be possible to normalize the current practices to it, which would improve the rendered maps of Libya (currently a bit messy). (I.e. all names in Latin+Arabic, or in Arabic only). Best wishes, Jean-Guilhem Le 27/02/2011 13:39, Claudius Henrichs a écrit : Hi there HOT team, I recently got in touch with a fellow mapper who had added Tripoli to the name-tag in addition to the local used arabic script during his participation in his H.O.T. remote mapping efforts. His argument was that he like many others who want to help can't read arabic and many other tools like emergency frameworks rely on the mapnik rendering. The first point (illiteration in certain scripts of the world) probably applies to 40% (just a guess) of populated surface of the earth so I guess this questio is quite relevant. I would like to get some insight on the second point he made. Maybe you can shed some light on this: 1) Do emergency relief organistions rely on the default Mapnik rendering? 2) Are there frameworks in place to help them quickly generate English-labelled maps of a certain area or even better bilingual ones? 3) What's the general recommendation you give for HOT mapping efforts regardings the main name-tag? I don't want to see the naming efforts of the Iranian mapping community being countered by some well-intended addition of English names once we have the next earthquake there. I think no one wants to rename a chinese province affected by the next flooding for the same reason. Maybe I can convince german community member Stephan Knauss to join the HOT team as he set up great bilingual maps like http://thaimap.osm-tools.org or http://iran.osm-tools.org Regards, Claudius ___ HOT mailing list h...@openstreetmap.org http://lists.openstreetmap.org/listinfo/hot -- Jean-Guilhem pgp 0x5939EAE2 ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [HOT] Nametagging: Local script versus international latin script (for Libya)
On 27.02.2011 21:14, Jean-Guilhem Cailton wrote: Maybe I can convince german community member Stephan Knauss to join the HOT team as he set up great bilingual maps like http://thaimap.osm-tools.org or http://iran.osm-tools.org Would it help to have bilingual rendering? I think porting a map similar the Iran one should be done quite easy. It already has bilingual names and font adjustments to improve readability of the map. Currently the minutely diff update is tailored to the thaimap. With some rework I guess it should be possible to enable the minutely diff also for Libya. Until then a manual daily update is the thing I can offer. Would this be of any help? Please let me know. Stephan ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk