Govanus's diary

Recent diary entries

Tonights work on rail ref's

Posted by Govanus on 15 November 2013 in English (English)

this has crashed with a relation placeholder error when saveing and substatially regressed so wiill need to be redone - GW260 from NESA. It might look a bit complex but it is designed to still work when lines are cut up for features like bridges, level crossings and future midspan abandonments. the ref GW260 is a rail equivalent to nameing a road A34 so all the sub sections [of individual OSM ways] have a close relation to the ref (known as Routecode and LOR). I also put in a grouping level for named lines (traditionaly termed roads in the trade) along with there relations. this should still function when ways are split. I was going to add the speed limits but gave up today as they can sometimes very along a named line/road as well as chnage See NR info for latest speeds, till I think how best to mark them. OSM has what look like informal levelcrossings for footpaths along this stretch but NESA dosn't tell the train driver of them and you might be likely to by fined for tresspass!


Posted by Govanus on 14 November 2013 in English (English)

Intreasting site to fold into my osm db useage.(

Maybe its reinventing the wheel but a mashup with mapcss, GeoVRML, OSMdata, kensi3D engine (in josm plugin) and some runtime intergration possibly with some html5 friendly version of ECMA/Java might give a {text-editor level(/maybe also by an html form element)} easily user maluable 3d render solution that could include multiuser realtime 3D interactive vectored MIcons and sound responces...

video inset signposts incoperated into kensi based landscape with mapcss colour and texture ques intergated: While users and deamon/bots feed info into the landscape (sea and rail traffic openseamap style [but fed with VRML ship models to annotate](and weather links...

For those wondering what happend to VRML97 in standerdised officialy at iso went though some impoving versions and then had the core effort pulled into making it like the xhtml effort at the same time to make it XML capatible and turned into X3D. Early bandwidth restrictions and some user awkwardness with dimlly lit models in cosmos made it lose widespread permanace as browser wars began to focus on pandering to the consistant visual layout Dreams of corperate marketing teams used then to still working in paper white than user definable world of html's 1-2. with big browser support week people gave up sticking to a standard and mainly evolved there own for use in things like games and more widely supported java/Flash filled the rest of the gap.

I think if people are putting the effort to evolve new things like mapcss that don't depend on a wide 3rd party browser support then loking at these older (fadeing from the mainstream memory) initiatives might prove useful to some. Software and algorithums are often already made and teasted (though admitadly some systems could benifit with some updateing of the old s/w to make it run on newer enviroments).

recent work

Posted by Govanus on 11 November 2013 in English (English)

Stopped mid way though seperating highway boundry from the main roadsurface as previous attempts to get the tagging right for wide compound road sites like this was makeing a confused mess. the road is getting an area likeat the north and the boundry and footway are yet to be seperated in the south end the result will render all parts in true areas (at bing sat accuracy) this I hope to then extend beyound the estate when finished.

British Rail routecodes

Posted by Govanus on 7 November 2013 in English (English)

I was looking some more into to the live map idea of mainline train movements similar to openseamap's: and I found this useful list of all the route numbers:-

RouteCode From Open Rail Data Wiki

Network Rail divide the network up in to a set of routes, identified with a code, e.g. MD101. These codes are used in the Sectional Appendix.

A list of route codes from the Sectional Appendix - correct as of 12th June 2012, is available.

As I get time I am still looking to be annoting these to OSM with the help of NESA pdf ed diagrams. The list of things hoping to do is still larger than the time available but the finite nature means that I'll get there eventually.

I found out how to make relationships work with some of the editors (I don't think id will work on [my] today's computer). So I will try and do them in a grand relationship. NESA also holds the abrviated official name of each "road"/track that is on some public faceing signs at places like bridges so I want to do these at the same time like I did nr padington.

going into the station the editor became a confusing bunch of parrell lines that began not fit easily with the NESA Diagram and using bing beneath just made it worse to comprehend! so staed near a flyover instead. earlier work is nr didcot jcn.

Garage tagging at Southfield Park

Posted by Govanus on 9 October 2013 in English (English)

When choosing the tags that relate to garages at Southfield park I took reference initially to the fields in potlatch and id: in addition I chose the:

name to be the figures and symbols written on the garage itself {what people would recognise it to be called} Note these have been changed overtime and sometimes relate to a house name or number of an user rather than a number that logically relates to the garage (ie in a sequence)

address & postcode to be discoverable postal information ascribed to the Garage (or block). Note not all are known to have been given such addresses yet.

Designation to be the identifying code used by the operator this is more likely to be in a more logical order than the name in instances where the number there is follower a more randomly allotted users home address.

Southfield Park is not uniform in ownership or operator so different conditions exist even on this small estate. Also time has seen renovation and renumbering on some while not on all.

To make things easy I picked the most ordered and logical garages first as well as the easiest to find full information for. For this reason it looks like I've overkilled the Numbering tags with repetition; though when I finish the OCiC ones and move on to the much harder private ones gaps of missing tags and illogical sequences are likely to abound.

To compound the look of repetition of the first drawn garages I also decided to split the Full designation over a series of grouping relationships to better describe the levels of the designation and make it easier to add tags appropriate to describing unique features of different levels. This matches management of the large numbers of garages controlled by the same operator.

If you have major objections to this approach it would save me a lot of time to discuss them here first, before unilaterally deciding to make major edits that mess up the numbering schemes described here.

If you have or know of sources of reliable designations and address listings for the non-attached private garages in Oxford and elsewhere it would be useful in finishing this name and numbering exercise on OSM.

Thank you for takeing the time to read this message.

Location: Lye Valley, Oxford, Oxfordshire, South East England, England, United Kingdom

Designatory grouping relationships

Posted by Govanus on 9 October 2013 in English (English)

I've been trying to find and then complete a way to describe a set of designations made of nested groups of sub members (that look like private address schemes) that apply to geographically identifiable places (for instance rooms, buildings, blocks and out-houses. I aimed to find away that I could avoid repeating the designation in each way or node, so a relation was the natural solution: in addition I wanted to be able to tag a whole group at different levels to give relevant thematic information relating to only that subgroup.

I also was keen to allow software to comprehend automatically the structure and all relations between the groups their levels and descriptive tags, in such a way that it automatically matched easily with owner supplied data and possible streaming (possibly private) that they could use to make automatic overlays(or slippery map layers) on there website in future. An attractive benefit when promoting its adoption over the current weaker Google based offering. Matching this with simple totally custom automated rendering when needed also adds to the potential appeal.

Nested Relationships solve this neatly fit all requirements and can be made to fit the current schema neatly. I therefore will look to make a proposal for wider adoption for places with similar needs. It is noted that renderers and other software is unlikely to become confused with this unknown type of relationship as most relationships are currently only selectively processed by programs made to look for special types they use: so I expect that uninterested software will ignore the new relations while the new (local) software written to support them will use them instead.

Location: 51.747, -1.224

recent upload failures.

Posted by Govanus on 1 October 2013 in English (English)

I made 101 changes on friday but they didn't appear to upload properly: so I'm looking though if they are permently los or stageing somewhere...

Offical OxCAB Tagging

Posted by Govanus on 26 September 2013 in English (English)

I'm adding tags on behalf of Oxford CAB of the sites they run in and a few other currently unshown advice services they refer people to. I hope to use the resultant osm data to render maps for the bureau for public guidance. The information is a bit in-depth and some sites of outreaches are shared with other organisations. At first I hope to leave a node with tags at the shared sites that others marking out the buildings can incorporate in their own shared use method. Internally I hope to use a mapcss to stick a logo with some edge coded to show if is a restricted outreach (one serving people using another service at the site that have requested that we help only them at their place). I hope to find a way to upload some logo's for general rendering of CAB advice sites and give a josm mapcss for their intended usage.

The Models of Southfield park is coming on steadily I'm working on improving the accuracy. British National grid is missing from Josm at the moment that slows stuff down a lot for me, I hope the result will surpass the Mastermap by Ordnance Survey when done.


Posted by Govanus on 12 August 2013 in English (English)

I'm looking at the praticalities of doing what open seamap does for ships movements with rail data feads.

Exploring the system link at the moment (NR uses stomp for open access live data)...


Posted by Govanus on 12 August 2013 in Korean (한국어)


Posted by Govanus on 24 July 2013 in English (English)

New Parkway shopping centre

Posted by Govanus on 23 July 2013 in English (English)

I had a visit to Newbury last weekend to this new site. Previously someone asked that I do the mapping for this and so I put something quick in at the time but now I am having some problems deriving how to OSM encode the new buildings along with other sites I'm looking to match they have multiple uses configured in a 3 dimensional separation. I.e. all stacked on top of each other. These new buildings comprise of large open parking halls underground with shops on top and then on the top is a roof-top housing estate complete with footpaths small gardenscape features and several purpose-built houses of multiple occupation! Now this is a good use of limited land resource, how ready is OSM's editors and renderers to handled features like ways that stack with others like two ways that share all the same (2d) nodes but relate to very different spaces in reality and have there own possibly mismatching sets of attribute tags to render. OSM at a XML data level can easily handle this and give level type tags to help define the hight separations but the editors that novices and those without there own home programs can have a hard time with these, so could editors have a mechanism clear to users that will allow them to chose the right features when stacked. I already have trouble with id and potlatch when trying to disengage joint node use with out needing to regenerate a replacement node in the right place for the stable features to be rejoined to at an XML level it isn't necessary as a one ref leaves the edited way leaving everything else unchanged.

I Oxford I'd also like to stack the dwellings and other categories in housing blocks that can carry there own address information, for navigation programs to use. this is step between HMO's while buildings and not the same as wanting to define interior walls and and other layout features.

Renderer authors and maintainers need to have some thought to how they want stacked data to work in there algorithms to give they results they'd like.

The website at gives high quality layout information to work from. Though I don't think I can easily achieve a high accuracy transfers at this time but I can start to put some building details on the map from it to help in the meantime.

Location: Wash Common, Enborne Row, West Berkshire, South East England, England, United Kingdom


Posted by Govanus on 14 May 2013 in English (English)

Activity on rail

Posted by Govanus on 1 May 2013 in English (English)

writing quickly;- I began to add some internal code names for the rails starting around Didcot.

I also was looking at adding some basic uprn code to some buildings

I'll explain lattter more on this.

Activity 2

Posted by Govanus on 24 April 2013 in English (English)

I looked at some local areas to were I live in Newbury and Oxford and try to fill in some of the Garage [block] driveway roads and interlinking footpaths that are oftern ignored in large maps targeted more at drivers.

I did some croschecking with site vists to some and a few other means for others. It is fun to see that OSM is now showing roads that Google maps only can hint at; with the ocasional opertune streetveiw! Newbury contains like other places in the UK roads that don't work like a normal road you can drive down. these example I was working on oftern end their name in Walk and have only a footpath were the road goes and the front gardens comeing upto it from the houses on both sides. So you can only walk to these houses. Restictions are light and just phsycal so takeing a horse though {that didn't eat the flowerbeds} would probably work.

The density is quite large in the area I was working in

Activity 1

Posted by Govanus on 24 April 2013 in English (English)

The first activity begun was to Number the international roads in the UK as designated by the local area offices of the United Nations.

I started with my local one that goes from Gibralter to Greenook Primary route E05.

The only part I have done so far is on the British south coast from Gosport to the M3. Unfortunately I found I need to edit a lot of very small segements so only reached the southern end of the M3.

They aren't regulaly marked on the roads in the UK though a sign was chosen for them and the British Goverment did do a lot of leaseing at least in the past.

what you tend to find is that Motorway projects and major trunk roads (central funded and maintained) to take complementry paths to the scheme; so HGV use is still good for these roads.

Older Entries | Newer Entries