Australian place names - rationale ?

Firstly, congratulations on a great informative tool

I am aware from the FAQ that errors in the location of place names in Australia is a known problme - names are some kms away from their true location

However what puzzles me is the selection of place names - predominance appears to be given to a seemingly random selection of local government areas rather than capital cities eg Adelaide is shown as the City of West Torrens (the West Torrens district is but one of a dozen or so local governmen areas within Adelaide). Likewise in Victoria the DiamondV alley Shire (another local government areas - in some states they are known as Shires, in others as 'Cities') apeears but not Melbourne. Same is true in Sydney where a couple of local govt districts to the north of Sydney take predominance.

The inaccuracies in both place names and location are so severe that it leaves me with little confidence regarding locations in other areas of the world with which I am less familiar.

[Beansprout]

Interesting. From what I've seen it's an algorithm which decides what to show, rather than a huge list. Seems the algorithm needs some tweaking

As for the inaccuracy in locations - yes, as you've read, this is known and will be fixed.

It is almost certainly a quatity control issue and not a selection algorithm that is the underlying issue - almost without exception the names shown in bold yellow (before the more detailed set of towns/suburbs in smaller print) are for practical purposes meaningless.

The names shown in bold yellow are of little relevance and are simply misleading. It is difficult to believe that there was a deliberate decision to provide a random selection of county/shire/council districts and not display capital city names.

I suspect the problem is that the list that used to generate these names has been drawn from some source without any quality control or local knowledge.

[TomServo]

Actually, and you can search back about 3 months to be sure, the list you see in WW now was compiled from several sources. Which is why there are multiple tags for location, locations named wrong and miss alignments.

It has all been noted, and in the next full release will be fixed.

[Beansprout]

The best bit is that Wales is left out of the UK if you look

Same reason as it was left off the front cover of the European COmmission's report - they had a map of Europe but with no Wales - the coastline is apparently too complex for the algorithm they used to manage, so it skipped it!

Thanks = having played around with WW a little more I am even more enthusiastic. It has been great to be able to 'fly' across areas I have travelled over the years. It adds a new geopgrpahic perpsective. I will in future be interested in adding GPS tracks for historical interest as well as putting trips into perspective,

Having looked at various locations in Australia, Europe and the US there is a high degree of inconsistency in the location naming. It looks as though locations were just tipped into the bucket from any available source with few if any rules fo consistency. Some areas (eg Australia) the names are of little use in others (eg France) they are so dense as to be unreadable. At present this is the least useful aspect of WW but I recognise it is a work in progress. If you need any assistance with the Australian bit please let me know.

In terms of the registration errors in Australia. Not sure if this is known but the Landsat 7 and topo data do not align at least in central Australia. This can be clearly seen with the town of Alice Springs. To the south of the township there is a distinct mountain (?) range (part of the Macdonnell Ranges. This range is a sharp delineating feature and is sufficiently steep that there is no building on it. The Topo data puts this range a km or two to the north of the Landsat 7 image - hence if you fly across it you see two obvious effects - the township appears to drape across the gap in the ranges (this is in fact a quite steep and relatively narrow gap). On closer examination you see an EW ridge in the town a km or two north of the actual range as depicted in the Landsat 7 image. No sure which one (or both) are misregistrered but this may give another pointer if it is not already known.

Cheers
Bruce

Originally posted by TomServo@Jan 8 2005, 05:05 PM
Actually, and you can search back about 3 months to be sure, the list you see in WW now was compiled from several sources. Which is why there are multiple tags for location, locations named wrong and miss alignments.

It has all been noted, and in the next full release will be fixed.

Quoted post

I've seen the same problem in germany. The names seem to be placed at the crossings of an virtual grid with about 2km height and less width. It looks to me like a rounding problem.

[Beansprout]

Originally posted by Guest@Jan 10 2005, 08:40 PM
Thanks = having played around with WW a little more I am even more enthusiastic. It has been great to be able to 'fly' across areas I have travelled over the years. It adds a new geopgrpahic perpsective. I will in future be interested in adding GPS tracks for historical interest as well as putting trips into perspective,

Having looked at various locations in Australia, Europe and the US there is a high degree of inconsistency in the location naming. It looks as though locations were just tipped into the bucket from any available source with few if any rules fo consistency. Some areas (eg Australia) the names are of little use in others (eg France) they are so dense as to be unreadable. At present this is the least useful aspect of WW but I recognise it is a work in progress. If you need any assistance with the Australian bit please let me know.

In terms of the registration errors in Australia. Not sure if this is known but the Landsat 7 and topo data do not align at least in central Australia. This can be clearly seen with the town of Alice Springs. To the south of the township there is a distinct mountain (?) range (part of the Macdonnell Ranges. This range is a sharp delineating feature and is sufficiently steep that there is no building on it. The Topo data puts this range a km or two to the north of the Landsat 7 image - hence if you fly across it you see two obvious effects - the township appears to drape across the gap in the ranges (this is in fact a quite steep and relatively narrow gap). On closer examination you see an EW ridge in the town a km or two north of the actual range as depicted in the Landsat 7 image. No sure which one (or both) are misregistrered but this may give another pointer if it is not already known.

Cheers
Bruce

Quoted post

Indeed it is a work in progress, you are absolutely right

I ask you to remain patient and enjoy what's there at the moment - the trails and MODIS functions, as well as the USGS .25m data, are all excellent to play with. I've spent ages simply browsing the globe looking at the landmark catalogue addon

The mis-aignment of topography data is another issue which will be corrected in the forthcoming release of WorldWind, unless anything major happens.

I assure you the devs are working harder than Michael Moore at voting time (am I allowed to say that, or will Jessi bring in that space laser again? :P) to get the next version out, and the NASA team are being deafened by all the extra servers they've brought in to convert and produce the data