When I was a working man, for twenty three years I did the same 25 mile journey into work every day. Over the years, the duration of the journey became less and less predictable, sometimes 50 minutes, sometimes nearly two hours. It didn't, take much to cause a delay, a burst water main in Teddington, a burst lorry tyre on the M3, road works in Kingston, whatever. In the end I resorted to leaving home before 7 am to avoid the main part of the rush hour, but still the journey home after work was totally unpredictable.
Why am I telling you all this? Because it's getting a bit like that on some canals and I've been thinking about it while playing around with programming new CanalOmeters. Every day, I think of improvements to make it easier to use, but in the end anything that estimates a journey time has to be taken with a big pinch of salt, even on a canal.
My first cardboard Canal0meter, way back when, was based on journeys into London from the Slough area. At that time, you could keep up a steady speed and with there being no locks until Camden, the journey times were pretty consistent. Even heading north up the GU, the traffic was so light that queueing at locks was rare and there weren't that many moored boats to slow you down. These days, with all the moored boats, it's a different story.
Then I've been thinking about other unpredictable places. Take Watford locks for example. On a good day, you might arrive and go straight in, passing through in forty minutes. On a bad day you might have to wait a couple of hours to be let in.
Or how about the Oxford canal which gets very busy in high season. A lock taking less than fifteen minutes in the spring, might take an hour in the school holidays.
Or how about a river which might slow you down to one mph against a strong flow or sweep you along at 7 mph the other way?
Then there's your crew. When we went up the Hatton flight alongside Nb Chertsey, good old Jim went ahead on his bike and every lock was open and waiting when we got there. We probably shaved a minimum of an hour off the ascent that day. Probably more.
So any calculating system for predicting journey times is fraught with difficulty. Nick Atty's wonderful Canalplan gets closest if you can be bothered to set the defaults. He lets you choose different speeds for wide and narrow canals, wide and narrow locks, and locks in flights. Maybe he should also set it for areas with moored boats, although that would be a non trivial and ever changing exercise to keep the data current.
In the back of my mind I can hear dear old Maffi saying "Why the hell do you want to do all that anyway? Why not just set off and see how far you get?" He has a point, but then again, he's a continuous cruiser and usually not on a schedule. Lots of other folk have a finite time for a trip and they want to plan how far they can get in the time available, or perhaps to plan stopping places overnight, or to check whether to stop at this pub, or risk reaching the next one in time. I think I'm right in saying that Canalplan is the number one visited canal related website for those very reasons, plus of course sometimes to compare different routes to the same place. It's also worth being aware before you decide to set off from,say, Banbury that the journey to Stratford-upon-Avon, about 36 minutes in a car, involves 101 locks each way!
So despite all the 'if's and 'but's I'm still fooling around with making myself CanalOmeters because , being a bit of a Nerd, (or should that be Geek?) I enjoy doing it, and sometimes they come in really handy. The latest Python version I knocked up only yesterday is dead whizzy, but I have some ideas on how to make it even better.
I think they call it displacement activity.
Why am I telling you all this? Because it's getting a bit like that on some canals and I've been thinking about it while playing around with programming new CanalOmeters. Every day, I think of improvements to make it easier to use, but in the end anything that estimates a journey time has to be taken with a big pinch of salt, even on a canal.
My first cardboard Canal0meter, way back when, was based on journeys into London from the Slough area. At that time, you could keep up a steady speed and with there being no locks until Camden, the journey times were pretty consistent. Even heading north up the GU, the traffic was so light that queueing at locks was rare and there weren't that many moored boats to slow you down. These days, with all the moored boats, it's a different story.
Then I've been thinking about other unpredictable places. Take Watford locks for example. On a good day, you might arrive and go straight in, passing through in forty minutes. On a bad day you might have to wait a couple of hours to be let in.
Or how about the Oxford canal which gets very busy in high season. A lock taking less than fifteen minutes in the spring, might take an hour in the school holidays.
Or how about a river which might slow you down to one mph against a strong flow or sweep you along at 7 mph the other way?
Then there's your crew. When we went up the Hatton flight alongside Nb Chertsey, good old Jim went ahead on his bike and every lock was open and waiting when we got there. We probably shaved a minimum of an hour off the ascent that day. Probably more.
So any calculating system for predicting journey times is fraught with difficulty. Nick Atty's wonderful Canalplan gets closest if you can be bothered to set the defaults. He lets you choose different speeds for wide and narrow canals, wide and narrow locks, and locks in flights. Maybe he should also set it for areas with moored boats, although that would be a non trivial and ever changing exercise to keep the data current.
In the back of my mind I can hear dear old Maffi saying "Why the hell do you want to do all that anyway? Why not just set off and see how far you get?" He has a point, but then again, he's a continuous cruiser and usually not on a schedule. Lots of other folk have a finite time for a trip and they want to plan how far they can get in the time available, or perhaps to plan stopping places overnight, or to check whether to stop at this pub, or risk reaching the next one in time. I think I'm right in saying that Canalplan is the number one visited canal related website for those very reasons, plus of course sometimes to compare different routes to the same place. It's also worth being aware before you decide to set off from,say, Banbury that the journey to Stratford-upon-Avon, about 36 minutes in a car, involves 101 locks each way!
So despite all the 'if's and 'but's I'm still fooling around with making myself CanalOmeters because , being a bit of a Nerd, (or should that be Geek?) I enjoy doing it, and sometimes they come in really handy. The latest Python version I knocked up only yesterday is dead whizzy, but I have some ideas on how to make it even better.
I think they call it displacement activity.
2 comments:
You probably won't be surprised to know that I've had many of the same thoughts, and taken part in discussions around suggestions, for CanalPlanAC.
I've also reached the same conclusions: I don't think it's doable, the best thing is to suggest to people that they use it as an indication as to how they are doing overall - are they sliding behind or building up a reasonable "buffer" of time. Anything more than that ("you'll be at Little Venice at 14:20 three days from now" is wishful thinking, and then some!
But, as you say, it's endless fun playing with the calculations.
Nick, you're right. I'm having curb my enthusiasm by not worrying too much about detail like exact lock times. Nevertheless, next time we're out on Herbie I may have a go at noting how long the locks take etc. A sort of CanalOmeter calibration cruise:-)
Post a Comment