Category Archives: geodata

The Pragmatic Mapper (part deaux)

My pragmatic mapper post brought some interesting responses so I thought I’d outline it a bit more.

First, the title is a riff on The Pragmatic Programmer, a neat book.

It’s also an explicit riff on Linus Torvalds explicit pragmatism and differences of opinion with the more political FSF. The FSF is freedom for freedoms sake, Linus is freedom because it’s just better. I like the latter.

On the points of does the OS matter, generally, for some value of matter. Of course they do. They’re much like Microsoft. They’re a monopoly, they have lots of money and giggly lawyers and everyone hates them. They also have maps of the whole country. But do they matter to me, personally, or many OSMers? No. The political side and bringing down The Man doesn’t really motivate me. We make OSM because it’s better and cooler.

I do find it interesting that the responders think that a national mapping monopoly is somehow a good thing. I fail to see how this is different from a national monopoly on tea bags or cars. It seems that the argument is that the OS is relevant because it has lots of data and we could use this (pulls rabbit out of hat) in a flood emergency. Bit of a poor use of a lot of money just for that. Anyway, to respond to points in turn;

…I want someone to organise data cross the entire country. I want to operate a business dependent upon that consistency.

Thats fine, but you don’t need a NMA to do that. You could have regional agencies with individual contracts where an overseer body puts it together. You could, god forbid, use OSM when we’re there. On an economic note, even if you want all that, and you want an NMA that’s fine, just please don’t force me and the 65 million other citizens to pay for it as we might not want it.

…I want to know that environmental policies in the north and south are based on similar data, its analysis and methodology and applied fairly.

Cool, but still don’t need an NMA for it. We have policies in the north and south are based on similar data, its analysis and methodology and applied fairly in schools, hospitals, roads, universities, water, gas, electricity…. most of these just need an overseer not an NMA.

…In the case of an emergency, I want to know someone can put together a river network and all its tributaries and work on solving a hydrological problem effectively, if there is a need.

Still don’t need an NMA for that. Being a bit hardcore, the emotive issues of ‘think of the children!’ or ‘what if we get flooded’… well the insurance market is very clear about that. Don’t live in a flood plain. If you want amazing disaster recovery maps of your area, then pay for it but please don’t force us all to. We might want to, of course, just don’t force us. And Katrina is not a good example, the federal government distorts the market by forcing flood insurance through FEMA. It’s the same argument as keeping rural post offices. If people in the countryside want them, then pay for them. There’s no god-given right to maps and post offices.

…I want someone to survey and record the entire country in case I want to visit other parts, know what is there and understand where I can go.

What do you do when you go to the united states then? The country isn’t falling apart because they have different mapping providers in different parts of the country. I can find my way around Orlando and San Francisco just fine, despite them being thousands of miles apart and one with a map from Hertz and the other from the county sheriff.

…I want an agency who supports governmental operations in a neutral manner with spatial information.

So do I. It need not be a country-wide monopoly. And the OS are far from neutral. By definition they stifle competition and progress, without even waving around OFT reports.

…I want someone responsible for ensuring the education system produces infrastruture and knowledge to people so the geography of the land is know, recorded and stimulated.

University Geography departments would not be impotent without the OS. 11 year olds can still learn about geography without a free map.

I wonder if Openstreetmap honestly feel that they are ready to provide disaster response mapping, or have the resources in place to ensure that their coverage of the entire country is current to within one year or less.

Not yet, but we or someone like us will. And anyway, you don’t need an NMA for disaster response mapping.

It’s perfectly fair for OpenGeoData to think that Openstreetmap suits his mapping needs, but to call it superior, and to say that the Ordnance Survey is irrelevant is a little short-sighted.

Navteq are letting you submit errors and so are teleatlas with map insight. Our way of making maps is most definitely superior and it’s the future.

If you can, listen to this podcast which excellently summarises The Wealth of Networks.

To come back to the original post for a second, really the pragmatic point was to say should we spend our time campaigning against the OS, or just building our own systems and maps? Campaigning for open data from the OS, or change to government policy is just sticky tar. Would we have got anywhere in the past two and a half years by just campaigning? It’s very doubtful. We’d have publicity no doubt and a few more high-placed friends and enemies… but this way we have that and a mapping system, and a community of 5,700 people, and maps of Baghdad, and vast sections of the UK mapped.

I have some idea of what I’m talking about here, as I’ve been involved to varying degree with fipr, no2id and stuff.

But one thing I think would be cool to do is make a map of map charges. The idea is that the OS basically don’t respond to awkward questions through the Freedom of Information Act as they’re commercially sensitive… but if we all write to our councils and ask them then they have to give us at least some idea. My council just sent me a letter with the new council tax bill breaking it down by police, schools and so on.. but not maps. So, we can figure out who’s paying the OS too much or little. It’ll be interesting. What you need to do is find your council website and information freedom officer and write them a letter asking for this stuff. There’s a wiki page with a sample letter to help you get started.

As far as I know this data doesn’t exist anywhere.

Another Week on the OSM Mailing lists…

Happy first birthday to Osmarender. As Blackadder/Andy Robinson pointed out this week Osmarender is now a year old. Before Osmarender there was nothing that generated maps from OSM data, and right from the start it performed well. Over the year Osmarender has added support for layers, areas, symbols and bridge/tunnels to give the Osmarender we have now. Compare the two maps below of Weybridge from the first to the latest incarnation of Osmarender. Also, Osmarender is the rendering engine in the times@home project which turns round updates to the map very quickly. It will be interesting to see what the next year will bring. Great inspiration Etienne & great work to all the collaborators ….

Weybridge by Osmarender

Mapping and rendering tunnels came to the fore this week with Steve Chiltern starting a discussion around just how do you really map out a tunnel (the astute of you will realise that GPS coverage is not the best in a tunnel). Various solutions were proposed with video-based surveying & inertialmeasurement units sounding promising. Not sure about the wire-guided missile idea though! Mind you, with a tunnel engineer amongst the OSMers some progress should be made. The issues of layering and tunnels was also covered here.

The question of how to get tiles updated on the map seems to come up again and again. Well, for the Osmarender view go to http://www.informationfreeway.org/ zoom in to level 12 on the area you want, and click ‘request update’. There’s talk of getting this implemented as part of the slippy map too. On the dev list there is discussion of a JOSM patch to allow the same kind of thing.

A great discussion kicked off by Andy Allen on tagging MBR trails (Ed: he really meant MTB trials). Should they be covered by a route tag? Should they be given grades? Should there be an ascii-art description of the route surface ( nnn=cobbles, __~~~___=flooding etc.). Looks like there will be crossover from the MTB community to get some mapping done. Great!

SteveC’s thread on Finding new stuff while mapping points to the joy of getting out on the road. Being ‘forced’ to go down every lane/track/path brings out lots of things you’d probably never noticed about your local area. So what are you waiting for? Get out there & see your place in a a whole new light!

This gets my thread-of-the-week award: The set of problems thrown up by the question of maximum segment length by Scott Walde. It started off by problems with things not rendering properly and segments being missed because the start/end of the segment was out of the area being rendered (though it did cross the area). Adding a point in the area ‘fixed’ this. Being a global discussion it brought up a whole range of issues on road networks – a 50K straight road is not uncommon in some places, as are grids of roads running north/south & east/west. And what happens with different projections – well worth a look. So how long is the longest stretch of straight road in the world?

There were a few things of general note on the dev list this week. Thomas Lunde brought up the US TIGER import problem. TIGER is a database of all the streets in the USA. Lets say that again: TIGER is a database of ALL the streets in the USA. And there are no copyright restrictions on using the data. Handy eh? There was an import of the data running for quite a while, but it had to be purged because of data cleanness problems. A new version of the TIGER data has come out (as of the 6th March) and there is discussion of re-running the import, and how best to do this given the existing mapping that has happened in the USA. Separately on dev Andreas Voltz asked about Writing an OSM map application and received a bunch of useful information about OSM data & appropriate databases.

The number of mapping parties is growing like topsy. Last year there were seven. There are three this March, and ten so far either happened or planned this year to June. These parties have given step changes in the mapping around Manchester, Bath, Reading, Brighton, The New Forest, VāsterÃ¥s, Rutland, Munich, Val Thorens, London, Toronto, Bristol, South London, Amsterdam (last weekend), and planned for Sheffield, Karlshrue, Madrid, Southampton and North Wales. This doesn’t include the recent funding where 12,000 euros have been earmarked to subsidise another ten mapping parties in the Netherlands this year. Way to go!

Don’t forget: If you want to submit something to OSM’s State Of The Map Conference, you had better get your skates on. The closing date for submissions is the 30th of March.

Legal-talk has been strangely quiet this week…..

by Barry Crabtree.

Law in Action programme and debate

Tomorrow at 4pm there is a programme on Law in Action on Radio 4 in which they examine crown copyright, openstreetmap and the OS. You can listen online.

Then on Thursday there is a debate with Andrew Gowers at the Francis Bancroft Lecture Theatre, Mile End Campus, Queen May, University of London at 5pm. You need to RSVP to this address to get in.

This week on the OSM mailing lists

Its been another high-traffic week on the OSM mailing lists, so lets get ready for a whistle-stop tour to the world of OSM.

The legal debate continues, but the focus has shifted to an attempt to aggregate information about Case Law, Statute Law and general legal questions onto the Legal FAQ wiki page. The principal behind this is firstly to gather together information that we can present to a friendly (and perhaps charitable) copyright lawyer and secondly to get information about similar cases. There are a couple of papers that are definitely worth looking at, this one, written by Dr Charlotte Waelde from the AHRC Research Centre for Studies in Intellectual Property and Technology Law at the University of Edinburgh and this page from the Science Commons website. If reading a journal paper is just too much like hard work, Richard has summarised the main points of the paper here. If direct action is more your thing, some OSM users have been declaring their contributions Public Domain on this wiki page.

If you are new to OSM, Geo and OpenSource projects, there’s probably a lot of things that confuse you. If you are wondering what the difference is between GPX and GPS, head over to the glossary page of the wiki. If there’s a term you want defining, just add it without a definition and someone will update it before too long. There’s also a new mailing list for newbies – sign up here and start asking all those questions that have been bugging you.

There’s been a load of development going on in the OSM commuinty this week. Tom Higgy has written a script that will calibrate OSM image tiles letting you use them with the opensource Treck Buddy navigation software. Take a look at Tom’s work on his site. Next up, Mathew has been working on a new rendering engine that takes OSM data and produces a postscript file, an interesting new take on the rendering issue. The pragmitic developer says; “Maybe there is a use for it, maybe not. I just like hacking PostScript, so this seemed a sensible merge of two enjoyable things…”.

The “Applet” is the editing tool that you use when you click on the “Edit” tab on the OSM website. Its been around since the very beginning of OSM, the first version was written by Steve, the next by Tom Carden, the way support was put in by Imi and inbetween that a few other people have had a hack at it. The Yahoo imagery has prompted a fair bit of development work on the applet. Nick Burch put in some fixes a few weeks ago, and this week Dan Moore has made quite a few bug fixes, including enhanced panning and zooming functions – all of which are now on the live site.

Keith Sharp’s been bussy too, adding a script to extract polygons from the weekly planet dump – useful if you want to extract a particular area of data, rather than load the entire dataset. You can grab a copy here.

My personal “app of the week” is Franciso Santos’ fantastic new Yahoo! imagery plugin for JOSM. Download the plugin from here, then fire up JOSM and get ready for some productive mapping. The plugin works by caching image tiles via Firefox and then loading the image into JOSM – get more info here.


Putney Bridge as seen using the JOSM Yahoo! plugin

The tagging discussions have continued this week too. One problem that’s been identified is that there is currently no way to distinguish a tube station from another type of railway station. Currently we tag railway stations with:

railway=station

Robert “Jamie” Munro’s not happy though, “that’s just wrong”, he says, and he has a point. After all a railway station is not part of the railway, and so shouldn’t be tagged with “key: railway”. One suggestion is to allow multiple tags with the same key, something like:

railway=rail;subway;tramway

Whilst this is supported by XML, it is not currently supported by the OSM API. You can follow discussions about tagging and the OSM ontology here. There are a load of new proposed features that you can vote on this week too- take a look at this wiki page and have your say.

As the amount of data in the OSM database grows, and as the uses of OSM data become more sophisticated and diverse, the underlying data structure will become increasingly important. The idea of “superways” is not a new one, but David Earl has brought the matter to light this week, largely in response to mailing list discussions about tagging motorway junctions. David’s proposal is firstly to introduce the idea of a “superway” as a higher level structure on top of the existing ways, as well as encouraging the API and editors to enforce the existing “rules” of ways (they are contiguous, ordered, unidirectional, non-branching sets of segments). David sees several advantages of the proposed scheme; a road in a housing estate that has numerous branches for example, can be grouped as one object, likewise a complex road junction which is not a single node or a roundabout could also be represented as one object.

Basically all we’re trying to do is assign tags to groups of ways without having to do it to each way, or have an easy method of selecting multiple ways – John McKerrell

There are several strands to this debate, but primarily there is the need to represent real-world objects in the OSM database and however this is done it will involve some level of abstraction. Aled Morris makes a good point that the ability of a renderer to render a particular data-model should not affect the development of that data model. The problem is, that the output of renderers will inevitably affect the way people tag objects and the way people perceive the OSM data model. Nigel Magnay makes the point that whilst you could feasibly tag your way around many situations, this is an inefficient way of dealing with the underlying problem, and what we should be doing modeling the fundamentals properly. Of course different people have different ideas of what the “fundamentals” are and what “properly” is. The herding of cats springs to mind, but then this is all part of OSM’s open, community based way of doing things.

Thats all for this week. With all the new development thats been going on there’s plenty of new toys for every kind of mapper to play with – so get out there (or fire up JOSM) and have some fun.

Another week rolls by on the OSM mailing lists…

Oh, my head hurts. I wasn’t around last weekend to do the roundup so Nick agreed to do it. I mailed him suggesting it was a fairly quiet week. “have you been following Legal-talk?” he asked. “Better subscribe. ” I said. Big mistake! It’s not the quantity of mailings that’s the problem, its the passion, quality & general level of understanding. This is where Baz the Blogger’s head really starts to hurt, I’ll keep it brief:

There are three main debates going on at the moment “Is the share-alike road navigable“, “Moving up the stack“, and “The big license debate“. I don’t pretend to do them justice in this one-sentence-per-subject summary so if the subjects interest you and you want to join in, please take a look at the debate. Thanks to David Groom, SteveC, and Tom Chance for starting off these particular threads.

The share-alike road in one sentence is this. Can the license to OSM data be made so that the people that use and enhance OSM data are obliged to make their enhancements available to OSM. To make this concrete, if someone took the OSM data and, for a particular map they placed street names at ‘good’ locations, they would have to return those locations to OSM. There are also other debates on licensing for derivative works. “The big license debate” discusses the pros and cons of CC vs. PD licenses. Should the data be made public domain – a complete free-for-all, or will CC licensing work? One issue with making the data public domain is that everyone that contributed must state that their work is in the public domain. “Moving up the stack” covers whether OSM should just stick to the data layer or move up in some way – afraid this one made my head really hurt!

It is good to see that the FAQ keeps getting updated. For example, people are naturally keen to see the edits and additions they make appear on the slippy map.  Now you keen peeps can now go to the FAQ to find how to get your freshly edited area updated quickly… then clear your cache, have a nice relaxing meal & come back to see the results of your hard labours (on a good day :-) ).

Something weird happened with a question by Franciso Santos posted on “Tagging urban streets” – whats the tagging scheme for roads when you get in the centre of a city. Plenty of discussion to agree the scope of the problem, then Mike Collinson proposed something that seemed to be agreed. Wow! Take a look – it doesn’t happen that often.

It seems to have been a busy week for applications. Or should I say nacient application-lets. OJW announced an atlas generation programme. Richard Fairhust announced a test version of Potlatch, a flash-based editor for OSM data and again OJW let us know about an easy way to generate map jpegs on the fly here, surely these should have a CC license logo on them? Or are they derived works – oh, my head is starting to hurt!

Like most (all?) distributed collaborative projects, OSM is discussed at many levels: on the mailing lists (some that we summarise here, legal, talk, dev, party ) the wiki and IRC. There’s only one thing to say about the IRC channel #osm……  “There be sxpert’s” ;-)

Fancy a trip to Spain? Iván (who I’m sure was in rehab) is organising a mapping party in Madrid – why not join in? Oh, and be careful mapping in China.

And finally, you know you’ve got Obsessive Compulsive Open Street Mapping Disorder when “…going to see relatives suddenly sounds interesting…” – too close Dutch! And it just goes on…

Enjoy!
.
By Barry Crabtree

This week’s round-up

The debate surrounding the use and licensing of OSM’s data that was reported in last week’s round-up has continued in earnest this week, resulting in an explosion of traffic on the OSM legal mailing list:

There are several components to the problem. First, there is the issue of copyright. Currently each individual contributor retains the copyright to their contributions. This situation is unlikely to change, as in many countries, such as Germany, it is simply not possible to transfer copyright held by an individual to another body. Next is the issue of the OSM license. All the data held by OSM is licensed under the Creative Commons Attribution-Sharealike 2.0 license (CC-by-SA). CC-by-SA means that anyone can use OSM data for anything they like, provided they give attribution to the creator(s) of the data (the guys with the GPS units and bicycles) and providing that they apply the same license to any derivative works that they make and distribute. I think I speak for the whole OSM community when I say that we all want the best free map of the world we can get; how we define freedom and how we get there is another matter. Richard Fairhurst gives his opinion on some of the problems with the current OSM license:

No webmapping company or cartographer will contribute any data back to OSM except through the goodness of their own heart. That’s because CC-SA doesn’t require you to publish the source code, just the final product.

So if a cartographer produces a beautifully illustrated map using OSM data and then redistributes the map, they are required to redistribute it under the CC-by-SA license. This would make selling the derived work very difficult and Richard and others have suggested that this factor will stop cartographers from using OSM’s data in the future. Instead, they will go to organisations from whom they can buy data under licenses that allow them to create works that they can use commercially, effectively bolstering the position of organisations who sell proprietary data. Frederick Ramm and Robert “Jamie” Munro also point out that this scenario contradicts OSM’s stated aim, to produce maps that are free of the “legal and technical restricions” that most geo-data is subject to.

Take another scenario, one in which a big provider of geo-data gets hold of OSM’s data-set, cleans it, augments it and then uses it to produce map tiles. The current CC-by-SA license does not require this organisation to release the augmented data back into theOSM data set , they would only release their finished product. Those who see this as a problem have suggested making a distinction between the source data (a planet dump for example) and a “rendered” work. It is suggested that the source data could be licensed so that any improvements have to be given back to OSM, whilst derived works would not need to be licensed under such a license, allowing more widespread use of OSM data.

This is a very brief round-up of a pretty complex problem that many people have strong opinions about. Whatever the final outcome of the debate is, it is going to play an important part in shaping the future of OSM. We’ll be meeting with some Creative Commons people in the next couple of weeks, so if you have any ideas about licensing, send a mail to the legal list, or take a look at this wiki page, and make sure your opinion is counted.

Almien published a list of interesting tags, based on the most recent Planet dump. One of my personal favourites is “wrong=oh yes!”, with “description=Official home of the President of The Philippines”, coming a close second. After seeing “blackadder:service=cobler” I know where I will be going to get my shoes fixed in the future, after which I can visit “blackadder:cuisine=fish” for a nice bit of seafood.

OSM@home is a project that uses distributed computing power to render SVG images of places from OSM data. You can browse through the different places here. Through the work of Barry Crabtree, J.D. “Dutch” Schmidt and others, the latest OSM@home client post-processes Osmarender data to produce images with beizer curves, like these:

There’s more information about the curve smoothing algorithm here and you can download the OSM@home client from here.

Free-Map – the countryside oriented open mapping project – now have a similar client that allows people to render Free-Map tiles (with really nice looking SRTM contours) at home. If you want to help out, you can get it from here.

The OSM Cartography meet-up on Saturday saw Artem, SteveC, Steve Chilton and Richard Fairhurst spend a few hours going through the Mapnik config file with the aim of improving the cartography produced by Mapnik. About 10 people met up in Oxford afterwards and there was some interesting discussion about cartography in OSM. Steve Chilton and Richard Fairhurst made the point that a lot of the discussions about rendering OSM data focus on getting as much data as possible onto the maps. They pointed out that the secret of cartography is to represent the information that is needed for a particular use case, and had some convincing examples of traditional cartographic products, like maps oriented towards cyclists, that don’t include many features at all, but succeed in conveying the information needed by the cyclist by choosing the most useful features. This is definitely something for OSM’s renderers to think about. As the database gets larger and more diverse, and the demand for cartographic outputs of OSM data gets greater its going to be increasingly important to carefully select the features that are represented on our maps.

Saturday’s “You know you got OCOSMD (Obsessive Compulsive Open Street Mapping Disorder) when….” thread, started by J.D. “Dutch” Schmidt has thrown up some funny suggestions, along the lines of “…when you read other peoples gpx files as blogs” or “… You forget to pick the girlfriend up from work, but remember the GPS”. It would be so much funnier if it wasnt so painfully true. There’s definitely a fair amount of obsessiveness in the OSM community. I’ve always found that you appreciate things a lot more when you start trying to do or make them for yourself – thats certainly the way its been for me with mapping. Thomas Walraet reported on a French project that’s taking the do-it-yourself ethos to the extreme, making their own GPS circuit boards:

The boards retail for around €100, with the board’s manurfacturers making no profit. More information is available in French, here.

Thats all for this week, keep tuned to OpenGeoData for all latest news from the world of Open Mapping.

Submissions for State of the Map 2007 are now open

Now’s the time to submit your papers for The State of the Map 2007. If you’ve been planning your talk, workshop or poster session and are ready to submit, click here. If you need some more inspiration, take a look at the submisions page of the State of the Map site. This is shaping up to be a fantastic weekend, we’re looking forward to receiving you submissions.

Turk meets GIS

Theres an absolutely fascinating use of Amazons Mechanical Turk (?) right now. There’s a HIT (a small task you get paid almost nothing to complete) that involves GIS:

Geospatial Vision are paying people to do image recognition on sequential video stills from a car that they are apparently then recombining in to videos. These are on their (flash only, sigh) site.

You are paid 5 cents to tag 50 images with yellow lines, manholes, drains, bollards and pedestrian crossings. They are also, from looking at the videos, using these locations to then magically classify the sign type (one way, no entry, speed limit etc). Most images have only one feature if at all, there were about 2,000 HITs last night and at 25 frames a second that puts it at about an hour of footage for $100. That is insane.

If you wanted to get data out of it, the video stills themselves could be captured from your screen like the above screen shot and put back in to a movie. People and number plates can be seen in the images… and street signs so you could figure out where they are. You could add bad data – bollards in the sky or whatever. Amazon have various methods to combat these attacks. But it’s all academic as they’re putting at least some of the work on their site anyway.

It strikes me that this is just scratching the surface of the potential of this class of problem, Mechanical Turk is still only known to a small subset of tech people really. People with big data sets would want entire teams of lawyers to look at this and have Snow Crash-esque schemes to keep people from ‘stealing’ their precious data. Could you imagine the OS ever touching this with a barge pole?

The barrier to entry is a little high in that you have to create a flash app or similar if you want to do more interesting HITs but simpler ones are done automagically with forms by amazon. The other barrier is that like many other companies they think the entire world ends at the edge of CONUS so you can’t really make use of it unless you have a US bank account.

There’s another HIT which just asks for an idea from you – what small program would you like to see that doesn’t exist? I imagine the person who did that one just sitting back and browsing ideas for things to work on. How meta can you get?

So we now live in a world where you can effectively treat data storage (Amazon S3), processing (Amazon ECC) and mass non-linear human intelligence (Mechanical Turk) as infinitely cheap and available. You can get programming, design, legal advice and more from rentacoder, elance and more.

Given this, I can’t think of much that you don’t have covered in Phase 1 of your average business plan. Or to look at it another way, the google kids have been living in this world for maybe 4-5 years.

So readers, if you had a big dataset what would it be and how would you get it processed using the above? Best idea gets $10 worth of HITs.