Recent diary entries
I have been cycling in the Taurus Mountains in Turkey in the summer, and off course water is essentiel. Therefore I have made a waypoint everytime I met a water source, but how do I transfer it to OSM?
Thanks in advance
Question for the developers: how will Potlatch be prepared for the day when Adobe Flash finally disappears from the webscape?
Thanks in advance.
Here is my weekly update on my Outreachy project.
#Week No: Seven
##Target Milestone: Python Tags, Mockup edits, Survey results, Content
###Summary: OSM and HDM Tags were created in python, the survey results of the OSM Export Tool was analyzed, Mockup was given touch-ups and content for Help Wizard was created.
Does anyone know of an Android app that will take my location and let me add a note there? My planned use is to go biking/walking through local strip malls and take down information on each store.
Please add our company location as Envicon Engineering Solutions to MAPS.ME
Dear UK/GB OpenStreetMappers,
From time to time we talk about the potential of setting up a "UK/GB OpenStreetMap group" (name yet to be decided) but we never quite know what it should look like.
Survey time!! Please fill in the following 2 minute survey:
Your responses to page 1 will be shared when we close the survey (in a few weeks). If you respond to the optional page 2 questions (your details), your responses will be used for the purpose of administrating the group only (they will only be seen by myself and any designated administrator should a UK/GB group be set up).
So stop reading and go to the survey:
Best regards and happy mapping,
Only M1 section from Kobryn to Orsha is SIGNED AS EXPRESSWAY in White Russia (Belarus).
We are preparing a new osm2pgsql stable series release, 0.88.0. This is based off of the work done in 0.87 development series with porting to C++.
Help is needed in doing a last round of tests before release, in particular the z_order logic. 0.88.0-RC1 can be obtained from https://github.com/openstreetmap/osm2pgsql
If upgrading from a 0.86.0 or earlier created database, the schema migrations in https://github.com/openstreetmap/osm2pgsql/blob/master/docs/migrations.md are required.
Osm2pgsql is now C++ and requires the Boost libraries.
A new backend has been added, the “multi” backend. This allows multiple tables which can each contain different types of features. More documentation is available at https://github.com/openstreetmap/osm2pgsql/blob/master/docs/multi.md
In-database pending way tracking has been replaced with in-memory tracking, offering significant performance gains. This requires a schema migration for old databases.
z_order logic has been improved, taking into account recent work across multiple styles. https://github.com/openstreetmap/osm2pgsql/pull/374 has more information. This requires a schema migration for old databases.
Osm2pgsql 0.87.4 has been released. This development release is focused on improving the node cache and pending way status storage.
Lockfree Queue removal
boost::lockfree::queue implementation of the pending way queue has been removed, leaving the
std::stack based implementation which used to be available with the
--without-lockfree configuration flag. The stack implementation was found to use substantially less RAM. This should allow the
--cache value to be increased and drastically speed up import speeds, particularly with full planet imports and machines with 16-32GB of RAM and mechanical hard drives.
boost::lockfree::queue implementation used to require Boost 1.53 or later, and this difference has been removed.
Any package maintainers packaging the 0.87.x series should move to 0.87.4 and remove any usage of
--without-lockfree from their build scripts
Node cache cleanups
Cleanups have been done to the node cache, significantly expanding test coverage, allowing 0,0 as a valid coordinate with the dense cache, and fixing some bugs.
Default style fixes
Some fixes have been made to
default.style. These do not impact anyone using it as-is, but should avoid confusing behavior based on column order if the file is modified.
--oldmerc option has been removed. This used a legacy projection that no one would normally want, and if someone does want it, they can specify it by SRID
- style.lua cleanups
- Travis CI improvements. OS X support for Travis is ready, but Travis is not offering new capacity for Mac workers at this time
One large issue remains that should be addressed before 0.88.0 is release - reimplementing middle_ram_t::relations_using_way. Development help from one of the many organizations using osm2pgsql in production would be welcome.
A full list of commits is at https://github.com/openstreetmap/osm2pgsql/compare/0.87.3...0.87.4
Photo © CC BY 2.0 Seattle Municipal Archives
Over the last 8 years I have been exploring and mapping the Elizabeth Furnace area in Virginia near Seven Mountains. I am planning to update the data in this area over the next few months. I will be working on trail alignments, trail conditions, adding points like primitive campgrounds, overlooks and so on.
because of some roadwork roundabout and tunnel is closed temporarily I can put acces=no but that would make it closed forever , especially on downloaded maps
My tag motor_vehicle:acces = no @ (Jul 10, Jul 11,2015) does not work, Navigation still goes trough closed tunnel
(something like this would be great: no_acces_until=11-07-2015)
At least online navigation should avoid the road, all the navigation on openstreetmap still goes trough despite the road is closed.
We have a pretty impressive house for the president compared to your average head of state - a 340 room palace in a 320 acre estate with some amazing gardens, which I had a chance to visit on a sweaty afternoon a few months ago.
The estate was not looking too impressive on the map.
Spent close to an hour this afternoon to improve my drawing skills and its looking a lot better now:
Is there any single way to get information about administrative entities? It seems to me that there is a class of map users who are interested in admin entities. Perhaps they are interested in city boundaries. In the US, there may also be school districts, water districts and open space districts.
Once one has an id of some relation, it seems ok to get information about that entity.
But I have tried a couple of times to discover how to find these entities and I keep hitting a wall.
I can construct OverPass queries for these things, but they do not work and it is hard to see why they do not work. It is sort of like that old joke: How many prolog programmers does it take to screw in a lightbulb? false.
Here are queries I have tried:
relation ["boundary"="administrative"] ["is_in:country"="USA"] ["is_in:state"="CA"] ["admin_level"="8"] ["border_type"="city"]; out body;
<?xml version="1.0" encoding="UTF-8"?> <osm version="0.6" generator="Overpass API"> <note>The data included ....</note> <meta osm_base="2015-07-09T21:00:02Z"/> </osm>
Hm. Not too helpful.
relation ["is_in:country"="USA"] ["is_in:state"="CA"] ["admin_level"="8"] ["border_type"="city"]; out body;
relation ["is_in:country"="USA"] ["is_in:state"="CA"] ["border_type"="city"]; out body;
relation ["is_in:country"="USA"] ["is_in:state"="CA"] ["name"="Ramon Road"]; out body;
relation ["is_in:country"="USA"] ["is_in:state"="CA"] ["name"="Sunnyvale"]; (._;>;); out body;
I’ve been using mapillary for a while now for a few different projects and thought I’d share some of my experiences of using it. Before I start, I’ve no affiliation with Mapillary, I’m just a fan of it!
What is Mapillary?
“Crowdsourced street level photos... Using simple tools like smartphones or action cameras anyone can collect photos that are combined into a street level photo view.” For the average user, this means a street view that we can use to edit OSM, with added functionality like auto-detection of road signs to help things.
Mapillary is a core function of the OpenStreetMap iD editor
Open an iD editor page and click on ‘Map Data’ from the right hand menu (Shortcut: F) and select Mapillary photo overlay. For mapathon usage, Mapillary will help show people what roads, buildings and features look like in a particular area and hopefully provide some context to the aerial images they are mapping from. Sure, you probably won’t be lucky enough to have photos exactly of the area you’re looking at, but you’ve gotta start somewhere!
In all of my uses so far, I’ve used a Garmin Virb Elite to capture images and the manual uploader on mapillary’s website. There’s a blog on how to here. The easier way to do it (although my phone is too old and I can’t download it!) is to use the mapillary app which has loads of function in there to help create your street level photos. Take a photo(s) on your phone, upload it to Mapillary later on when you get some wifi.
Uses so far
The first project I put Mapillary to use for was in Zimbabwe with the Missing Maps Project. Using a Garmin Virb Elite on loan from the British Red Cross, I took around 15,000 photos during the 2 weeks that we were mapping there (part of this time was spent in Epworth, part in Dzivarasekwa). The use of Mapillary in Zimbabwe is pretty obvious; there’s no Street View coverage here and having it enables people to verify what’s on the map using the images. The American Red Cross have also used Mapillary for a project in Haiti, and there's a Missing Maps Project using it right now in Bangladesh.
Coverage of Google Street View is pretty comprehensive in London (and across most western countries, bar places like Austria and Germany who opposed it), so why do we need Mapillary? Well, there’s plenty of reasons. Mapillary serves as a good place for people to store their geotagged photos, which I know some keen mappers use to improve OSM. With the automatic recognition of roads signs, improving details like turn restrictions, no entries and speed limits is far easier. There are also plenty of places where Street View doesn’t exist in London, but would be useful to have coverage for – tow paths; cycle lanes/alleys etc. I got myself a handlebar mount for my Garmin to attach it to my bike, and off i went. There’s also a growing number of organisations opening up there imagery to Mapillary, including the City of Greater Geelong here with 1,362,869 and the state of Victoria here with 3,276,086 images, both in Australia. Hopefully this can be replicated elsewhere!
I’ve recently been on a white water kayaking holiday in the Austrian Tirol region, and used my Garmin Virb Elite to document the rivers and roads there. My thinking is that at some point, these images could help to form a more detailed kayak guidebook or be used to show different river levels and how things change. I’m still in two minds as to whether having a detailed river level view is a great thing (does it detract from the adventure?!), but given the availability of GoPro footage on most rivers now it’s not really much of a concern.
Go and use it! Either start taking photos, or turn on the photo overlay the next time you use iD. The new release of the ScoutSigns JOSM plugin from Telenav has intergrated Mapillary function, and I think the ball is rolling for further JOSM integration in the future (feel free to comment if this is/isn't the case).
The first version of the road style is prepared. It is work in progress and feedback is welcomed. I started work from designing z16 zoom level, currently I am expanding style to cover higher and lower zoom levels.
Current rendering (preview location from openstreetmap-carto readme)
Style under development - with its variants
Display of major roads on lower zoom levels is currently working quite OK and I am trying to make it better. For example, I am considering various variants how junction names and oneway arrows should be rendered. Below are some of possibilities.
Test renderings are for location from "trunk is invisible in forest and secondary on farmland" bug report. All renderings have for comparison on the left side how map is currently displayed.
gray oneway arrows and red junction names: https://cloud.githubusercontent.com/assets/899988/8582450/22903378-25c7-11e5-8e6c-b88daeb5d81a.png
gray oneway arrows and blue names: https://cloud.githubusercontent.com/assets/899988/8582452/22bd76a8-25c7-11e5-9018-44d6f6d8d0cf.png
version with modified road colors: https://cloud.githubusercontent.com/assets/899988/8582451/22b4d58e-25c7-11e5-8dba-461c8a018bc0.png)
rendering for names and arrows unmodified despite road colour change: https://cloud.githubusercontent.com/assets/899988/8582447/228aa700-25c7-11e5-8b9e-64f79a2e2d9b.png
Current rendering of highway=footway is quite noisy, also highway=pedestrian is closer to highway=tertiary in its style than to highway=footway. Dotted highway footway is not pretty - for example rendering proposed by sb12 is much prettier. But finding styling that would at the same time would be prettier than current one, do not look like a road and fulfil additional criteria turned out to be an interesting challenge. Finally, after many failed attempts, I tried as joke styling inspired by OSM landscape. It turned out to be better than expected and was the first one that was considered by testers both as prettier and with readability not worse than the current rendering.
Unfortunately, purple is already used for borders, airport infrastructure, railway landuse and industrial landuse so I am experimenting with both different colour for pedestrian infrastructure and changing landuse colours.
rural area where highway=footway are important:
highway=pedestrian and highway=footway:
pedestrian area & footways nearby landuse=railway:
Also - something that is already done.
During reworking road style I decided to try changing closely related tram style as on z13, z15, z16 tram lines were too noticeable.
During more close checking of rendering I discovered that tram tracks were rendered also on z10, z11 and z12 what was quite surprising as I never noticed it.
On reading code I discovered that railway=tram was rendered from z8, what makes no sense. Especially with current style as even during looking for tram lines at known positions I failed to notice it.
Overall this was one of many things leading to really bad rendering of cities around z10.
Even in Toronto, city with "the largest streetcar system in the Americas" and quite far on the north (relevant due to distortion of scale) it makes sense to render trams up to z12. It is possible to make a weak argument for rendering at z11. See http://overpass-turbo.eu/s/ai3 (I linked overpass turbo, as in current rendering tram lines are not possible to find - despite rendering them).
Other example, Vienna - "With 173.4 km of track, Vienna's network is one of the largest in the world." z11 is the lowest zlevel where rendering tram makes sense ( http://overpass-turbo.eu/s/ai4 - note, query returns large amount of data).
In general it seems that even with large networks rendering tram up to z12 makes sense, z11 may be justified but it is not the best idea but later it is rather pointless.
And obviously, there is no point in rendering something in way that makes it impossible to notice it.
Unfortunately it is impossible to make rendering great for all locations. In some places each tram tracks is mapped separately and in some places two tracks are together mapped as a single railway=tram line. In places where two tracks are mapped as one on low zoom levels trams are not displayed not as strong where each track is mapped. Compare for example Praha (where two tracks are mapped as a single railway=tram line) with Helsinki and Kraków (each track mapped as railway=tram line).
Tram rendering is causing exactly the same problem with current road style and my proposed version. Working on it separately makes easier to test and review so I prepared it as a separate change.
Proposed change that should improve situation was merged yesterday - thanks for received feedback!
It's been a while since my last post, so I just wanted to give a short update about the plugin development stage.
This week I've finally managed to have a working code for the upload feature ;-)
All signals and functions for the imagery tab are implemented:
- Select files from file system
- Load QGIS layers (getting the corresponding file system path)
- Add/remove selected sources
- Rearrange order of sources to upload (up/down buttons)
I still need to work on the package for QGIS which will include all the dependencies, and upload it to qgis repository with the tag experimental. In the meantime, you get get the code at our Github repository and follow the install instructions at the README file.
The interface has also evolved a bit since my last post, and an update on that will come in the short future.
As always, feedbacks are very very welcome, please don't hesitate to play with the plugin and get in touch.
Yesterday's Missing Maps party in London looked like it was going to be completely chaotic. Both Astrid (who organises the mapping parties) and I had been on holiday til the day before and the new MSF office (which we only moved into a little while ago) was only ready for us to move around about ten minutes before we started.
But, by 1820 (twenty minutes after start time) all that could be heard was the clicking of mice and the low level buzz of excited, whispered conversations...
How did it go so smoothly when the opposite was looming?
Chatting with Astrid later, we came to the following conclusions, which I want to share here...
Firstly, our Missing Maps community in London is committed and engaged. Many people come back on a regular basis and friendships and sub-communities are forming. So cool to see groups forming around certain tasks, such as supporting Carmen, who is field mapping with the OSM Bangladesh guys or helping Rupert get the field paper data he and Kieran gathered with local mappers in Epworth, Zimbabwe on to OSM.
Secondly, we made the discovery that if you split people up by experience and by what they want to achieve (we had four rooms: iD for beginners, JOSM for beginners, experienced JOSMers and validators-in-training), the conversation flows, the people in each room are much more able to support each other and peer training really works.
Lastly, our HOTties from London (and nearby) are great. Each room had a HOTty in there, responsible for leading and advising. So, as soon as people turned up, they got comfortable, knew who to ask for help and could get rolling with no fuss at all. When we have everyone in one big room, it certainly looks more impressive, but people are less confident asking questions and HOTties often end up answering the same questions countless times.
Of course there are other reasons (the wifi didn't crumble, the pizza turned up on time etc), but for those looking to put on large scale mapathons (about 70-80 people last night), I hope this helps...
So today I started mapping my most visited places, Luanda, Kitwe, Windhoek, and a bit of Lusaka. Luanda as of now looks like a huge twisted net with no direction - and the plan is making sure the place ends up being as precise as possible.
New areas of Windhoek, NAM and Kitwe, ZAM, are also being corrected.
Let's see what's going to happen. Really looking forward to seeing my updates on Maps.me.
UPDATE We are working on getting the source code out as Open Source as well. I will post with more details when I have them.
ScoutSigns is a JOSM plugin developed by my colleagues here at Telenav that displays speed limit signs (and a few other types) as a layer in JOSM. These signs are detected by users of our global Scout app if they opted in and have their phone mounted on the windshield. This is a pretty good source of sign information ready to be mapped!
JOSM showing Scout detected signs in ScoutSigns in Hamburg, Germany
Recently, Mapillary started detecting all kinds of signs in their millions of user-uploaded street view images. What would make more sense than collaborating with them to display their detected traffic signs right along with ours in Scout Signs? That's exactly what we did!
JOSM showing Scout (red) and Mapillary (blue) detected signs in ScoutSigns in Hamburg, Germany
With this new release, JOSM mappers can unleash the power of the fast growing Mapillary imagery right in your favorite OSM editor! I have updated the ScoutSigns manual with some details about the new Mapillary functionality.
I am excited to have worked with Mapillary to expose their fantastic number of recognized signs to the power JOSM mapping community! I am going to make this a brief blog post because I want to do some more speed limit mapping :)
If you have ScoutSigns installed already, it will update automatically the next time you start JOSM. If you haven't tried ScoutSigns yet, look for it in the JOSM Plugins preference pane.