Road Atlas?

peterwalkley · 22 Jul 2016 at 18:44

Yeah - I realised that, but what is "minor" ? That's a lot of potholes, line repainting, fixing of street lightbulbs.

Clarkey · 22 Jul 2016 at 18:49

That must be a mistake, no way in hell nearly 11 million light bulbs got changed on the roads in 6 weeks never mind alterations to the road layout.

muon · 22 Jul 2016 at 18:58

Burnsy2023 said:
From 12/05/16 to 23/06/16 there have been 10,820,550 changes to the road network recorded by Ordnance Survey in GB. That's just in a six week period.

BS.

They are counting "changes" as something you or I wouldn't.

Clarkey said:
That must be a mistake, no way in hell nearly 11 million light bulbs got changed on the roads in 6 weeks never mind alterations to the road layout.

Lol. Yep. Really need to sense check numbers before posting them.

[TW]Fox;29816175 said:
They'll be minor changes that would not have any bearing on an actual map in the way we'd use one.

I'm surprised he posted such a blatantly ridiculous number in support of the claim that a paper map is 'out of date as soon as you buy it'.

I'd be amazed if it was even that.

[TW]Fox · 22 Jul 2016 at 19:14

lolBurnsy

Burnsy2023 · 22 Jul 2016 at 19:20

[TW]Fox;29816175 said:
They'll be minor changes that would not have any bearing on an actual map in the way we'd use one.

I'm surprised he posted such a blatantly ridiculous number in support of the claim that a paper map is 'out of date as soon as you buy it'.

Around 820,000 of those are changes to traffic flows and directly influence routing.That may be new one-way streets or turn restrctions.

Around 4,000,000 of them are changes on junctions to roads, either existing or new roads.

For reference, if you had a new roundabout with three exits when previously there was a T-junction that would be 3 new changes. New estates will generally be a few hundred changes with roads, links and signage etc.

[TW]Fox · 22 Jul 2016 at 19:25

Burnsy2023 said:
Around 820,000 of those are changes to traffic flows and directly influence routing.That may be new one-way streets or turn restrctions.

Around 4,000,000 of them are changes on junctions to roads, either existing or new roads.

For reference, if you had a new roundabout with three exits when previously there was a T-junction that would be 3 new changes. New estates will generally be a few hundred changes with roads, links and signage etc.

Cool so basically nothing of much relevance to a GB road atlas dated 2016 then. Glad we clarified.

'Oh no, my atlas said this was a roundabout and it's now a T-junction, I'd better just park up and wish I'd used Google Maps'

muon · 22 Jul 2016 at 19:29

Burnsy2023 said:
Around 820,000 of those are changes to traffic flows and directly influence routing.That may be new one-way streets or turn restrctions.

Around 4,000,000 of them are changes on junctions to roads, either existing or new roads.

For reference, if you had a new roundabout with three exits when previously there was a T-junction that would be 3 new changes. New estates will generally be a few hundred changes with roads, links and signage etc.

Over 6 weeks?

Do you have a link to where you got these figures?

Burnsy2023 · 22 Jul 2016 at 19:30

[TW]Fox;29816400 said:
Cool so basically nothing of much relevance to a GB road atlas dated 2016 then. Glad we clarified.

'Oh no, my atlas said this was a roundabout and it's now a T-junction, I'd better just park up and wish I'd used Google Maps'

*Sigh*

Even if 10% of those changes affect routing (a very conservative estimate) then it can make really quite a big different. Also remember this is only within a 6 week period.

If you're only driving trunk roads, then you probably don't even need a road atlas, signage would be enough. If you are going to any more detail then these changes can make a material difference.

I'm sure you'll argue that your gut feeling is indeed correct despite having little to go on, so I won't bother trying to further convince you otherwise.

muon · 22 Jul 2016 at 19:33

I am fairly sure what you'll find is that 1 new road is counted as a ridiculous number of changes. Simply because of the detailed amount of information that has to be added for the simplest of roads.

Rename a new road? If it's a long one, may add up to a lot of changes depending on the resolution of the data.

All things along these lines.

It's simply infeasible that OS can physically make that number of changes over a 6 week period.

Clarkey · 22 Jul 2016 at 19:40

Face it burnsy, it's a complete load of utter nonsense that there have been millions of changes to the roads within weeks.

Even if your argument had any validity at all, digital maps are not really that great a solution. Just as an example, Burnt tree island in Dudley a few years ago was changed to a light controlled junction. A complete change to a notorious bottleneck, yet google maps still showed the roundabout for a considerable period of time after the change was completed.

muon · 22 Jul 2016 at 19:41

I suspect he is reading marketing spiel from somewhere, where they have an incentive to stretch the definition of "change".

muon · 22 Jul 2016 at 19:49

I've figured out what Burnsy must be referring to.

https://www.ordnancesurvey.co.uk/business-and-government/products/itn-layer.html

Now to work out what a single change is define as.

DXP55 · 22 Jul 2016 at 19:57

Here ya go best post so far and just what you asked for

Everyone else must have miss read your post.

http://www.az.co.uk/?nid=60&iid=128...of4lnM3WURhYtZW45eT6shoC7vnw_wcB#.V5JrxS0qOpo

Not everyone has unlimited data

Burnsy2023 · 22 Jul 2016 at 20:01

muon said:
I am fairly sure what you'll find is that 1 new road is counted as a ridiculous number of changes. Simply because of the detailed amount of information that has to be added for the simplest of roads.

Rename a new road? If it's a long one, may add up to a lot of changes depending on the resolution of the data.

All things along these lines.

It's simply infeasible that OS can physically make that number of changes over a 6 week period.

Yes, you are correct that describing a road can be a complex problem and a single road will have many features that describe it, but I haven't suggested that they've added 10m new roads. I'm not arguing otherwise, indeed your description is pretty accurate.

Even if you strip out all the road features and only take a look at road routing information features which directly affect where you can drive, you've still got 820,000 changes every 6 weeks, which is still sizable.

Clarkey said:
Face it burnsy, it's a complete load of utter nonsense that there have been millions of changes to the roads within weeks.

Even if your argument had any validity at all, digital maps are not really that great a solution. Just as an example, Burnt tree island in Dudley a few years ago was changed to a light controlled junction. A complete change to a notorious bottleneck, yet google maps still showed the roundabout for a considerable period of time after the change was completed.

And this is one of the issues around Google mapping data, it's currency is pretty bad as you've highlighted.

muon said:
I've figured out what Burnsy must be referring to.

https://www.ordnancesurvey.co.uk/business-and-government/products/itn-layer.html

Now to work out what a single change is define as.

No, that's not my reference.

[TW]Fox · 22 Jul 2016 at 20:02

Bottom line is that your number was worthless and there is nothing wrong with a decent road map. Its a nice thing to have kicking around in the car.

The_Dark_Side · 22 Jul 2016 at 20:07

[TW]Fox;29816580 said:
Bottom line is that your number was worthless and there is nothing wrong with a decent road map. Its a nice thing to have kicking around in the car.

Very true.
Having a backup is always a good plan.

Burnsy2023 · 22 Jul 2016 at 20:09

[TW]Fox;29816580 said:
Bottom line is that your number was worthless and there is nothing wrong with a decent road map. Its a nice thing to have kicking around in the car.

You don't like cars without certain features and I don't like out of date maps.

My number is accurate though within the context for which it was given.

muon · 22 Jul 2016 at 20:10

Burnsy2023 said:
No, that's not my reference.

Well that is the most detailed level data OS have for the road network, so if anything it should contain more changes.

Your number is simply nonsensical.

What is your source comparison data? I am beginning to think you might have access to these files and have done the change comparison yourself. In which case it depends on how you have defined a change.

How many changes will renaming a road trigger? Every connecting node would have to be redefined, every segment of the road (and anything else?).

What about a new road? How many different characteristics have to be defined with each connecting node and segment? Is traffic flow defined from one connecting node to another, in which case every junction matters? Or even worse is it defined for a specific segment at a time rather than one node to another? How are speed limits defined for roads? Do details like number of lanes matter?

Simple things like this can turn 10,000 changes into millions.

Burnsy2023 · 22 Jul 2016 at 20:16

muon said:
Well that is the most detailed level data OS have for the road network, so if anything it should contain more changes.

Your number is simply nonsensical.

What is your source comparison data? I am beginning to think you might have access to these files and have done the change comparison yourself. In which case it depends on how you have define a change.

How many changes will renaming a road trigger? Every connecting node would have to be redefined, every segment of the road (and anything else?).

What about a new road? How many different characteristics have to be defined with each connecting node and segment? Is traffic flow defined from one connecting node to another, in which case every junction matters? Or even worse is it defined for a specific segment at a time rather than one node to another? How are speed limits defined for roads? Do details like number of lanes matter?

Simple things like this can turn 10,000 changes into millions.

This is easy to look up in the ITN specification. Whilst there is a scaling of logical changes vs actual feature changes, it's not as low as you're suggested. Anyway, I'm clearly wasting my breath, so I'll stop dragging this thread off topic.

muon · 22 Jul 2016 at 20:20

Burnsy2023 said:
This is easy to look up in the ITN specification. Whilst there is a scaling of logical changes vs actual feature changes, it's not as low as you're suggested. Anyway, I'm clearly wasting my breath, so I'll stop dragging this thread off topic.

I would be more than happy to accept what you are saying if you could show it. I have no particular knowledge with geospatial data.

I also have no idea by how much a "single" change compounds through the data (other than logically it would in this case), but I know when people quote such high numbers (in any field) they are usually using the largest defendable number.

What is the scaling factor?

edit:

If you go through the specifications

https://www.ordnancesurvey.co.uk/do...stermap-itn-layer-technical-specification.pdf

an incredible amount of information is captured for single features and points in a road. A single change or addition could trigger so many things to change, making the 10m number more believable but less relevant to this thread.