Well it's an interesting prospect, but in London myself I find that I prefer the opposite of what TFL's cycle routing thing thinks I would. I like main roads. Pedestrians tend to look before stepping out in main roads. They also have bus lanes. Bus drivers tend to be nicer to cyclists, they aren't in a hurry.
Some things are fairly simple, avoiding right turns for instance, as often its much faster and safer to go on a bit further until some lights or roundabouts.
Avoiding any road that is single car width, the amount of problems I find with drivers (despite me often doing in excess of the speed limits) rushing to overtake me, because i try to be courteous and let them at all times, only to then get stuck behind them due to oncoming traffic on single lane.
So whilst I applaud their efforts, I'll be a bit cynical and suggest they are going about it the wrong way, it would be better to crowd source some scores for route sections aimed at specific kind of cyclists (many don't like doing 10mph+ because they will get sweaty) and somehow weave them together.
They offer 3 routes (fastest, quitest, balanced) for this reason already. I asssume the improvements they talk about based on more in-depth data would also take into account speed/confidence of the cyclist.
The "public sector provider or limited company" is a tad limiting. This looks like the kind of thing you want to throw a talented CS grad student who's familiar with graph algorithms at for a couple of weeks, but in general they're not going to have their own limited companies. I suppose you could find an umbrella at short notice if you had to, but the hassle factor would be high for a single contract.
Yes, this is the requirement of the funder, not our own requirement. Collaboration formally with the containing University could potentially be a possibility.
Trouble with university funding depts is that they're usually slow, although my University isn't too bad: at least they let their Researchers take on outside contracts like this & I believe will handle all the paperwork / legals for a small cut. I'll forward your link round the department. Someone might bite...
We're almost there with the open-sourcing, I'm pleased to say.
Sadly this process has taken far longer than we hoped because the codebase, which was started about 6 years ago and which was not started with open-sourcing in mind, ended up with a lot of things like embedded passwords, deploy scripts, people's names, API keys, etc. There's been a mammoth effort to deal with that. The entire deployment system has had to be rewritten, for instance, and that done while trying to keep a very busy production system running that also hosts a pile of third-party sites without downtime.
Some things are fairly simple, avoiding right turns for instance, as often its much faster and safer to go on a bit further until some lights or roundabouts.
Avoiding any road that is single car width, the amount of problems I find with drivers (despite me often doing in excess of the speed limits) rushing to overtake me, because i try to be courteous and let them at all times, only to then get stuck behind them due to oncoming traffic on single lane.
So whilst I applaud their efforts, I'll be a bit cynical and suggest they are going about it the wrong way, it would be better to crowd source some scores for route sections aimed at specific kind of cyclists (many don't like doing 10mph+ because they will get sweaty) and somehow weave them together.