A Camino Planner

[charvolant]

Hi all,

While walking the Camino last year, I spent a lot of time thinking about how to optimise the stages. We'd started with a guidebook and then switched to the Buen Camino app but I was thinking about how to globally optimise the stages, stretching and compressing things here and there to make the entire trip as "nice" as possible.

I've been doing some coding since then and the result is at https://de-calixtinus.org/

Essentially, it passes through a series of screens where you specify your level of fitness, how far you want to walk each day, what sort of place you want to stop, variants you want to take, places you absolutely have to stop at to see the sights and so on. The result is a plan that lays out each stage for you.

It's reasonably fast -- usually a couple of seconds to generate a plan, up to about ten seconds if thing get complicated -- so you can re-plan while walking.

There's only one route available at the moment, the Portuguese route, with a variety of variants and a side-trip to Fátima, if you want. I've been using the excellent Buen Camino as a major source (with their permission). However, it takes a long time to compile the data so next one is probably Fisterra, so that I can get another one up and running in a reasonable time. Translations are limited to English.

This is still relatively untested, so keep an eye on "computer says so!" problems. If you encounter problems, please add an issue at https://github.com/charvolant/de-calixtinus/issues

The code is available at https://github.com/charvolant/de-calixtinus

Let me know what you think.

 

[Antnix1]

I clicked through and I'd probably advise you that this isn't how caminos are planned or walked.

Sometimes you walk a little further because you're feeling energetic, or really want to stay somewhere special. Sometimes you split a stage because it's raining and youre wading through mud. Sometimes after a week of albergues you'll stay in a hotel in a city.

I personally use gronze for planning and buen camino while walking.

In the interest of providing good feedback, "day's end cost" shows KM instead of a cost.

 

[roving_rufus]

I had a short play - kind of get your idea and I like to plan (even if I later abandon it due to a range of other factors). The using km for input is confusing except for actual distances. The boxes with minimum, average and maximum distance/time probably need further work to make it clearer rather than just a scant sentence of text at top of page. The final "output" - the "plan" section was fairly clear but often I found I was staying in a guesthouse or hotel at an end point when generally I prefer to stay in an albergue- and will often stop short or stretch extra kms to do so. I also got a random 10km day that seemed an odd spot to do so.


But don't think I am currently converted from planning distances using the godesalco planner (even if it doesn't allow for such a number of variant routes for initially planning) then using something like gronze, wise pilgrim buen camino etc for accomodation details. The godesalco has its own issues but allows more control in my choices eg distances, primarily using albergues etc

 

[charvolant]

I clicked through and I'd probably advise you that this isn't how caminos are planned or walked.

Sometimes you walk a little further because you're feeling energetic, or really want to stay somewhere special. Sometimes you split a stage because it's raining and youre wading through mud. Sometimes after a week of albergues you'll stay in a hotel in a city.

Thanks for the reply. That's one of the reasons I've been trying to make the response as fast as I can make it, so that people who change plans can then say, "well, now what?"

I personally use gronze for planning and buen camino while walking.

In the interest of providing good feedback, "day's end cost" shows KM instead of a cost.

I've been hiding detailed explanations in the help screens attached to most of the input screens. I don't want to deluge people with too much detail but it looks like I haven't got the balance right.

 

[charvolant]

But don't think I am currently converted from planning distances using the godesalco planner (even if it doesn't allow for such a number of variant routes for initially planning) then using something like gronze, wise pilgrim buen camino etc for accomodation details. The godesalco has its own issues but allows more control in my choices eg distances, primarily using albergues etc

Thanks for the reply. It's possible to strongly prefer albergues by altering the preference levels. One of the things I should consider is asking for broad styles of travel up-front (similar to the fitness levels) and conditioning results based on that.
I've included help screens on most of the pages but it looks like I haven't got the balance of immediate/detailed information right.

 

[Antnix1]

Well good luck with it but i think you're going down a rabbit hole.

 

[krystianw]

This is pretty neat! I'd seriously consider using it on my next camino. (Provided you'd have a bunch of obscure caminos added in by then.) Most of the stuff I'd take into account when planning is in there, and I definitely like setting bounds on time and distance.

At the moment I think it's something only someone with experience would know how to use, because you need to know what your preferences are on some pretty specific things. But I'm sure you knew that already.

And the fact that it's in Haskell is... intriguing.

 

[Robo]

I had a quick look and found it very confusing. I had no idea if the distances were miles or kms, and the time? Didn't understand what that was.

I find Camino planning a very simple thing, although I think over the years I have become rather obsessed with the process! Like many others!

To me it's simple.

You can start with one of two questions.

One. How much time do I have and therefore where should I start?
Two. I'd like to start at X, how much time do I need?

Of course you first need to decide on a route, and researching here is a great start.

I then go through this process:

1. I'm going to walk route X, starting at point Y.
2. This is 700 kms.
3. My comfortable distance is 22 kms / day.
4. So I need to allow at least 31.8 (32 days)
5. I'll allow a rest/injury day ratio of 1:7
6. So I need 4.5 rest days. Call it 4.
7. So all up I'll need 36 days on the Camino.
8. I'll then factor in a buffer, which may or may not get used. Let's say 2 days.
9. So now I have 38 days on the route.
10. I can book my flights !

But..............

Before booking my flights I'll now undertake more detailed route reserach.
Any tricky areas.
Particular places to spend more time.......

Accomodation doesn't really come into it.
It is what it is. I'll use what's available.

With all the planning in the World, once I'm there, I'm free from my plan.
The plan was really only about knowing how long to allow and what date to book my return flight.

Reading a Basic guidebook and looking at the Gronze website gives me all I need.......

 

[Gerard Griffin]

Well done to create such a useful resource. What I'd like to see is an app that helps pilgrims to find Catholic mass on the Camino. I wish I had the skill to create one myself. It'd make a big difference to bible-bashers like myself

 

[David Tallan]

Hi all,

While walking the Camino last year, I spent a lot of time thinking about how to optimise the stages. We'd started with a guidebook and then switched to the Buen Camino app but I was thinking about how to globally optimise the stages, stretching and compressing things here and there to make the entire trip as "nice" as possible.

I've been doing some coding since then and the result is at https://de-calixtinus.org/

Essentially, it passes through a series of screens where you specify your level of fitness, how far you want to walk each day, what sort of place you want to stop, variants you want to take, places you absolutely have to stop at to see the sights and so on. The result is a plan that lays out each stage for you.

It's reasonably fast -- usually a couple of seconds to generate a plan, up to about ten seconds if thing get complicated -- so you can re-plan while walking.

There's only one route available at the moment, the Portuguese route, with a variety of variants and a side-trip to Fátima, if you want. I've been using the excellent Buen Camino as a major source (with their permission). However, it takes a long time to compile the data so next one is probably Fisterra, so that I can get another one up and running in a reasonable time. Translations are limited to English.

This is still relatively untested, so keep an eye on "computer says so!" problems. If you encounter problems, please add an issue at https://github.com/charvolant/de-calixtinus/issues

The code is available at https://github.com/charvolant/de-calixtinus

Let me know what you think.

I'm trying it out. I'm typing up my reactions here as I do so.

The first thing I noticed is that I was stumped on the very first question. It asks me my preferred travel style. The choices are " Walking", "Walking (Naismith's rule"), and "Cycling". I have no idea what "Naismith's rule" is. I went to the Help page and there was no help there. Consequently, I don't know whether to choose the first or second choice, nor do I have any idea of what the effect will be of choosing wrong. I'm about to go off and google "Naismith's rule" now.

Back now. I have actually seen this before, but didn't remember it. I wouldn't put that here, since I don't really think it affects one's style of travel. Rather, I would put it where you ask people's preferred distances. That seems more intuitive to me, because I think the option is really about whether you want this included in the calculation when calculating the daily distances.

I like the way you can set absolute minimum and maximum, preferred minimum and maximum, and ideal distances. I assume that the default values are based on the fitness level you selected.

After a bit of thought, I can see why you would express Day's End Cost in km but it is really counterintuitive to me. I really want to think of days end cost as how many euros I am spending at day's end, either for accommodations or all in. I suggest changing this to "Day's End Flexibility". I'm not sure how "Preferred Services during Day" interacts with Day End Cost, since I presume the difference between these and "Services at Day's End" is that you've already encountered these services before the end of the day, so why would you walk further to them? If you are walking further to them, presumably they would be in the section above (Services at Day's End).

Having done the calculations, there are a few things that are puzzling me. I thought I was expressing a strong preference for albergues over hostels (4km cost for albergues, 1km cost for hostels, guesthouses or homes, 0 cost for hotels or camping). Yet there were a number of times it put me in hostels when there were albergues well within the cost distance and I would not have been pushed over my absolute maximum distance on that or subsequent days. In some places, it put me in non-albergues where albergues are present (e.g. Ponte de Lima, which has an albergue but it put me in apartments).

I have a feeling it is weighting my preferred (rather than absolute) maximum much higher than my other preferences and is not willing to let the "cost distances" drive it to exceed that preferred maximum. My thinking was that by filling in the cost, I had said I was ready and willing to walk that much more than what I would otherwise walk.

That's my feedback. I hope it helps.

 

[Antnix1]

I'm trying it out. I'm typing up my reactions here as I do so.

The first thing I noticed is that I was stumped on the very first question. It asks me my preferred travel style. The choices are " Walking", "Walking (Naismith's rule"), and "Cycling". I have no idea what "Naismith's rule" is. I went to the Help page and there was no help there. Consequently, I don't know whether to choose the first or second choice, nor do I have any idea of what the effect will be of choosing wrong. I'm about to go off and google "Naismith's rule" now.

Back now. I have actually seen this before, but didn't remember it. I wouldn't put that here, since I don't really think it affects one's style of travel. Rather, I would put it where you ask people's preferred distances. That seems more intuitive to me, because I think the option is really about whether you want this included in the calculation when calculating the daily distances.

I like the way you can set absolute minimum and maximum, preferred minimum and maximum, and ideal distances. I assume that the default values are based on the fitness level you selected.

After a bit of thought, I can see why you would express Day's End Cost in km but it is really counterintuitive to me. I really want to think of days end cost as how many euros I am spending at day's end, either for accommodations or all in. I suggest changing this to "Day's End Flexibility". I'm not sure how "Preferred Services during Day" interacts with Day End Cost, since I presume the difference between these and "Services at Day's End" is that you've already encountered these services before the end of the day, so why would you walk further to them? If you are walking further to them, presumably they would be in the section above (Services at Day's End).

Having done the calculations, there are a few things that are puzzling me. I thought I was expressing a strong preference for albergues over hostels (4km cost for albergues, 1km cost for hostels, guesthouses or homes, 0 cost for hotels or camping). Yet there were a number of times it put me in hostels when there were albergues well within the cost distance and I would not have been pushed over my absolute maximum distance on that or subsequent days. In some places, it put me in non-albergues where albergues are present (e.g. Ponte de Lima, which has an albergue but it put me in apartments).

I have a feeling it is weighting my preferred (rather than absolute) maximum much higher than my other preferences and is not willing to let the "cost distances" drive it to exceed that preferred maximum. My thinking was that by filling in the cost, I had said I was ready and willing to walk that much more than what I would otherwise walk.

That's my feedback. I hope it helps.

4km cost for albergues means you'd rather walk 4km than stay in an albergue.

1km cost for banking at the end of a day means you'd walk 1km to go to a bank. Every day, apparently, because the penance of not having access to a bank is unbearable. (I believe this is how the app is programmed)

The icon means you avoid it at all costs. Or it's required at all costs.

A real world problem, as an avid user of buen camino, is you cannot depend on bars to be open. You cannot depend on the grocery store to be open more than 2 hours in the morning. You probably don't want to pay €7 to withdraw cash from one of those dreadful euro net atms.

The problem, and i hope the op takes note, is the data from buen camino is incomplete - but laid out in such a way where you can say "well, if this bar is closed we can buy something in the supermarket", but a bigger problem is how the logic coded into creating the stages does not represent what a user is looking for.

As an actual walker of the camino, i like to know if there is a gap between services. I'd rather know there's no atm for 100km, rather than an app trying to find me somewhere to sleep at night 250m from an atm! On Saturday I want to know if the supermarket will be closed by my 5pm arrival and not reopening til monday morning so i can buy stuff on the way as there is no benefit to the app having me sleep beside a shuttered shop. I want to stay in the must-stay monestary, instead of walking 5km to avoid every albergue and end up paying €60 for a rundown hostal. Even gronze and buen camino integrate with booking.com for prices/availability and have their own ratings systems to help users make informated decisions.

So if you're using incomplete data, producing routes using variables that don't have much sense, and relying on help screens and how tos, there's not really a product/service here, free or otherwise. Just somewhat random points +/- 10km from another random point.

Genuinely not trying to be mean. It's just there are so many variables that your app doesn't consider, and so few that it places too much emphasis on.

It might work better as a daily app rather than the unwieldy task of creating an itinerary out of variables that aren't suitable to create such an itinerary.

Now, I haven't used it and your best feedback will come from the people who have walked the exact camino your app currently provides who will be best positioned to tell you if they would enjoy the itinerary your app creates, or would it have made their experience worst. Let's call them your testers.

Your app has some usability issues (terminology, confusing descriptions, relying too much on how-tos) so you should probably help your testers enter the variables initially.

 

[charvolant]

That's my feedback. I hope it helps.

It certainly does. Thanks very much, and on to version 0.2

 

[charvolant]

Well done to create such a useful resource. What I'd like to see is an app that helps pilgrims to find Catholic mass on the Camino. I wish I had the skill to create one myself. It'd make a big difference to bible-bashers like myself

I can't promise anything but, theoretically, I think that it's possible in the model I've built. If you could add something to https://github.com/charvolant/de-calixtinus/issues describing how it would work I'll have a think about it.

 

[charvolant]

Genuinely not trying to be mean. It's just there are so many variables that your app doesn't consider, and so few that it places too much emphasis on.

Oh, I'd agree. I'm painfully aware of the Sundays and not open problem myself. But I have to (well, want to) start somewhere. As you said elsewhere, it's a rabbit hole but it's an interesting one to go down, full of doors and tea-parties.

As a project, this grew out of us realising that we'd painted ourselves into a corner over several days based on decisions made a few days earlier, when in-turn we realised that we weren't up to our initial plans. There was a stretch between Coímbra and Porto where nothing seemed to line up properly. So part of the motivation is to try to globally optimise (and then re-optimise as plans change, and then re-optimise) across the entire trip so that there's less future-bite in the choices.

Playing with gronze, I found myself falling into the trap we'd set ourselves by chunking distances into what looked like acceptable blocks and then finding that I had some ludicrously short or long segment nearing a major stop. So, while it may just be a failing on my part, that's the process I want to allow people to avoid.

Like any optimisation solution, it only tells part of the story. But I'm hoping that it generates a useful first cut for people to bend and shape to their needs.

 

[charvolant]

And the fact that it's in Haskell is... intriguing.

I have an interest in programming languages. Haskell has always been on my to-do list and this looked like an excellent project to get me over the initial hump. There's a lot of stuff where I really need to go back and re-engineer some stuff based on me getting to the point where things like applicative programming have become more natural. (Monads still look to me as if the response to the old adage, "if you're in a hole, the first thing to do is stop digging" was "but if we keep digging, we'll come out on the other side of the world.")

I've been pleasantly surprised in how sprightly the result has been, with a pleasingly low memory footprint.

 

[charvolant]

Based on the feedback I received from my previous posting I've been working away at producing an updated version, with the results available at https://de-calixtinus.org

You can view the release notes at https://github.com/charvolant/de-calixtinus/releases/tag/v0.2

The major change is the way accommodation preferences are treated. Now, a higher value = better, consistent with the way services are handled. There isn't a neat way of expressing this, so the results from these preferences lead to a rather complex calculation. Essentially, the planner looks at nearby locations to see if there is a better option close enough to be tempting. As a consequence, the penance cost for most accommodation appears as zero, meaning "under the circumstances, this is good enough." However, the planner will try to push for nearby accommodation more to your tastes, if it looks convenient.

I've also added a comfort level that conditions preferences around the things you might want on your trip and how willing you are to put up with

For those who wanted to stay in municipal albergues but couldn't. That was a bug and has been fixed.

For version 0.3, I hope to include the Camino Frances, although that's quite a daunting amount of data, and include some handling of points of interest. For version 0.4, I plan to start looking at including rest days and managing the Sunday famine.