Could you give our updated journey planner a quick test?

[Nearholmer]

Crikey!

My eyesight is so bad I hadn’t picked-up on the distinctions between the dash lengths on the brown lines.

I live and learn.

[francovendee]

I've taken a quick look and for someone who hasn't already found an app they already use then it may be of use.
I haven't spent long looking at it but so far failed to see a way of showing a profile of the route. One thing I really like about Cycle Travel is that you're able to call up a Google Street View. It gives a good idea of the sort of road you'll be riding on.
As I said I've a only taken a quick look and both features may be there and I just haven't found them

[st599_uk]

No breakdown of surfaces is a big negative.

Do any of the current maps actually include/encode surface information?

I ask, because once you move away from established roads and paved cycle paths, things get very interesting, with a huge variety of surfaces, so researching and encoding that lot would take “forever”.

(As a person who rides about on semi-knobbly 40mm tyres, my interest in this is probably the exact reverse of Wirral’s!)

Anything based on Open Street Map should have if a user has added it to the path element.

[Cycling UK website]

Thanks for so much feedback, everyone. You've given us a number of things to reinforce on our existing long wishlist, including

• drag to change route and/or click to add via points
• paved-only option, and surface breakdown
• (@francovendee - there's an 'elevation profile' button in the top right of the map).

In the meantime it sounds as if things are working fairly happily as far as it goes, except the issue on Ubuntu that emleyman had, and Nearholmer's trouble on mobile, which we're trying to replicate.

Richard is continuing to improve the speed on the maps, which is great, and the ctrl-zoom is something we can't really avoid for now, as the map is embedded in a scrollable page rather than being a single-screen implementation. (Though we will put a single-page implementation on the list of things to consider.)

So, it's a start, and we appreciate your help with it. Over the coming months we would also like to do some journey-planner research with less experienced cyclists; and possibly explore ways in which we can use our reach to help improve open-source mapping data in some small way. We'll also be rolling out more of the mapping work that Richard has done for us behind the scenes over the last year, which will improve the way we share routes and other content. We're really looking forward to that.

Sean

[SimonCelsa]

Just tried a quick test route from Inverness to Findhorn - just a random choice of places close to home.

Strangely, using cycle.travel off road it routed me across Findhorn Bay. I'm fairly sure there's not a ferry running or a newly constructed bridge in the vicinity, and I don't think the tide ever gets that low even at spring low water.

I'll have to cycle out that way shortly to have a gander.

[Nearholmer]

Nearholmer's trouble on mobile, which we're trying to replicate.

It still won’t work from “current location”, and I think that might be because I haven’t granted the gateway software permission to use my current location (I can’t see how to), whereas I have granted that permission to the two apps individually.

It does now work fine by selecting current location by dotting on it on the map …… but that isn’t ideal if you are using it live and happen to be lost at the time.

CycleTravel continues to pick needlessly long off-road routes within or through Milton Keynes, where I suspect it sort of gives up a bit (runs out of node-testing space?) because there are zillions to choose from.

[Ayseven]

I think they all have merits. But they don't read your mind, generally and sometimes you get sent somewhere unpleasant on the day. I will stick with cycle.travel, Google maps for towns, and Komoot because it integrates easily into my Garmin.

[Nearholmer]

You are right that they all have their quirks.

I gave CS, CT and GoogleMaps the same start and end points, Woburn Sands village centre to Winslow town square, trying CS in ‘quiet’, ‘balanced’, and ‘fast’, and CT in ‘ordinary’ and ‘gravel’, and got a different route from each.

CS ‘fast’, and CT ‘ordinary’ were very similar, and both included pretty decent quiet-ish road routes. GoogleMaps was an entirely mad route, using a long section of often incredibly muddy footpath (yes, footpath, not bridleway) and quite busy road, instead of the cycle paths available. CT ‘gravel’ was very strange indeed, including a great long diversion on paved cycle paths rather than the bridleways that run adjacent. CS ‘quiet’ and ‘balanced’ we’re the same, and seemed very good, using the most sensible combination of paved and cycle paths.

My guess is that if tested with other starts and ends the batting order would be different.

But, TBH, I’m in awe of any one who can create these route-finding algorithms, so I regard any suggested routes at all as miracles!

[jgurney]

Pretty good as online planners go. It did not make the errors that I have known several other make.
Unless set to 'fast' it tends to pick absurdly roundabout routes.

Some specific local issues I noticed:

- it treats Hammersmith Bridge in London as non-existent. This is open to pedestrians and makes a useful dismounted route.

- it does not alert users when it plots routes through parks, etc, which are closed at night.

- it sometimes plots routes along bridleways which are in fact mudbaths impractical as cycle routes (e.g. fropm Castle Cary to Hornblotton Green in Somerset, it recommends a very muddy bridleway instead of a slightly longer tarmac route). However it also sometimes avoids bridleways which have a good easily-cyclable surface.

- it occasionally marks 'cyclepaths' which do not exist at all or which are wholly within private property with no public access (e.g. it shows one such near Goldborough, N. Yorks, which is actually a dead-end farm track, and another inside the ICI works at Wilton, Teeside).

[sjs]

You are right that they all have their quirks.

I gave CS, CT and GoogleMaps the same start and end points, Woburn Sands village centre to Winslow town square, trying CS in ‘quiet’, ‘balanced’, and ‘fast’, and CT in ‘ordinary’ and ‘gravel’, and got a different route from each.

CS ‘fast’, and CT ‘ordinary’ were very similar, and both included pretty decent quiet-ish road routes. GoogleMaps was an entirely mad route, using a long section of often incredibly muddy footpath (yes, footpath, not bridleway) and quite busy road, instead of the cycle paths available. CT ‘gravel’ was very strange indeed, including a great long diversion on paved cycle paths rather than the bridleways that run adjacent. CS ‘quiet’ and ‘balanced’ we’re the same, and seemed very good, using the most sensible combination of paved and cycle paths.

My guess is that if tested with other starts and ends the batting order would be different.

But, TBH, I’m in awe of any one who can create these route-finding algorithms, so I regard any suggested routes at all as miracles!

I have cycled Woburn (not Woburn Sands) to Winslow a number of times, as part of longer rides. The first couple were long before online planners were a thing, and I happily chose a route using OS maps spread out on a dining table. Turns out, that route is more or less the same as CT's "paved" and "night" (how long has that option been there?) versions. More recently, I used CT's "any" route, without checking it first. That uses a longish drag on a path alongside the A5, then a tedious meander through the back streets of Bletchley, before emerging for the last few miles of pleasant countryside. But that's not CT's fault. All it shows is that if you want a route finely tuned to your own preferences, you should plan it yourself.

[Nearholmer]

That uses a longish drag on a path alongside the A5,

If it’s the path I’m thinking of, it is so unpleasant and so little maintained that it ought to be “struck off the register” of cycle paths, or be upgraded to a state where people might actually use it.

[Richard Fairhurst]

I'll reply to a couple of these over in the cycle.travel thread as they're really about c.t's routing (which is the same on both cycle.travel itself and the Cycling UK journey-planner) rather than the Cycling UK planner itself!

[pwa]

That uses a longish drag on a path alongside the A5,

If it’s the path I’m thinking of, it is so unpleasant and so little maintained that it ought to be “struck off the register” of cycle paths, or be upgraded to a state where people might actually use it.

I wonder if there is a case for allowing, as an option, short stretches of path where you will need to dismount, if it avoids long diversions or busy road junctions, etc. I have opted for a short walk section, in the past, to avoid a bit of unpleasant cycling.

[sjs]

That uses a longish drag on a path alongside the A5,

If it’s the path I’m thinking of, it is so unpleasant and so little maintained that it ought to be “struck off the register” of cycle paths, or be upgraded to a state where people might actually use it.

It's the stretch from Little Brickhill into Bletchley, largely alongside the A5, and including an unpleasant burrow through the filling station at the bottom of the hill.

[Nearholmer]

I use it maybe once a month, and find it a depressing mile, only worth bothering with to get up onto the hills. There is a nominal continuation of it south alongside the A5 for several miles, and that is even worse, because it is just as much a mess, and narrower, but one isn’t protected by the crash barrier. That bit I won’t use at all.

Next time, follow the road along the ridge to Soulbury, then bear right for Winslow if you want a scenic and reasonably quick route.

The path we are talking about is one I advocated for upgrade in the latest consultation, but it won’t get sorted out until new housing sprouts on the big fields adjacent.