Let pedestrians define the walkways | Hacker News Reader

Let pedestrians define the walkways

sivers.org

60 pointssivers24 comments

As an aside, here is a post about urban "pathways of desire" with some cool photos:

http://www.sweet-juniper.com/2009/06/streets-with-no-name.ht...

and a mathematical model for trail formation:

http://www.jamesgriffioen.net/trailmode.png

In every business I've been involved with, the execs made decisions that are really the customers' to make. While the execs were usually pretty accurate, they didn't see that they were predicting the customer (and unnecessarily so), rather than "making executive decisions".

It must be mythical, it doesn't include anyone suing over access for disabled people during the year of sludgy grass paths, or slipping and falling, or being late for class because they didn't realise they could walk over the grass.

Wait, anyway, that's not a decision made at the last minute as late as possible, that's a decision made very quickly that no paths would be fine.

An unmade decision would be more like "where shall we put these paths?" "I don't know", week one has arrows stuck to wooden stakes pointing over the grass, week two has mowed paths in a different layout, month two has borrowed festival queue tape things in yet another path, staff are bored of hearing about it at meetings, students are annoyed that it keeps changing, landscapists are annoyed that it's ugly, gardners are annoyed as it's intrusive, it's wasting time, money, brainpower, cognitive space...

Prexy's Pasture at the University of Wyoming was designed like this. http://www.uwyo.edu/tour/Prexy.asp

The only difference being, they used the trails in the snow across it to decide where to put the walkways. :)

Good. Finally. I realised this a long time ago, when I used to walk to school and got told off for taking short-cuts over the grass/plants - the truth is I wasn't taking a short-cut - the plants were just in the wrong place!

People will not be incovenienced by gardeners!

(BTW: the analogy did not allude me! :-)

In fairness, you'd have to be fairly stupid not to know where would be a good place to put paths - People tend to walk in straight lines from A-B (or door-to-door on a campus)

When you only have two buildings, only one door on each building, no other points of interest, a completely flat landscape, and no worries about ADA laws, it's a pretty easy problem. Then again, anything is easy when you control the variables.

I've always been frustrated by this. I realize that as a landscape architect there are aesthetic choices and perhaps for park lovers those shortcut trails can ruin a scene; personally I think there's a certain, non-dismissible, beauty in those natural pathway formations. At the end of the day -- well, most days -- people just want to pass as fast as possible and I think it's insane to expect anything different. At the local park where I live they built small fences (about a foot tall) to guide people around the grass, and what they got instead was stomped down broken metal bars, effectively adding trash to the park.

I'm pretty sure that whole "pedestrians defining the sidewalks" thing is actually a myth. It may happen sometimes nowaways since the myth is now so widespread but it's associated with so many schools now that I can't possibly tell if any of them actually ever did it. (And I'm pretty sure the majority didn't.)

My favourite example is when I hear that Berkley did it which doesn't seem at a plausible given articles like this: http://www.peterme.com/?p=18 that point out how they went out of their way to break the natural flow of pedestrians.

There's a comment on that page that says that Larry Wall says it was at UC Irvine. Though obviously it could still be a myth.

My elementary school was shaped like an L. And the path from the cafeteria to the playground followed the edges of the building, with a 90 degree turn at the inside corner of the L. The hypotenuse between ends of the L went through a hilly, wooded area.

Every day, at lunchtime, there were repeated admonishments to stay on the path and do not cut through the woods. And every day, immediately after lunch, someone (and oftentimes everyone) would run through the woods along the hypotenuse to get to the playground.

They tried all sorts of tricks to get us to stay on the path. They threatened instant detention for anyone caught cutting through the woods. They stationed one of the teachers aides to physically catch people who were about to stray from the path. It still didn't work, because there were so many people that broke the rules that they couldn't catch all of them.

Finally, when I returned for 3rd grade, I found that a roughly 10-foot-wide swath of the woods had been cut down, and a newly-paved asphalt path ran through it. And nobody ever took the long way around again. I think it was removed, long after I left, when the school building was enlarged.

Sometimes, it's easier to listen to your users (even if they are 5-10 year olds) than keep fighting them.

So when should you make [...] decisions? When you have the most information, when you're at your smartest: as late as possible.

Within software architecture, I've heard this called "late binding", also the Open Future Principle: "The best way to implement the future is to avoid having to predict it" [1].

1. http://www.vpri.org/pdf/rn2006001a_colaswp.pdf

I had heard this idea attributed to Christopher Alexander, from either "A Pattern Language" or "The Oregon Experiment". Possibly it originated from the University of Oregon?

Regardless of the original source, this is a great explanation of the concept..

"So when should you make business decisions? When you have the most information, when you're at your smartest: as late as possible."

I disagree. You should make decisions when you have enough data and enough understanding to draw a conclusion. For example, gather data for 1 week to figure out where the pedestrians walk (instead of waiting 1 year for the conclusion to be apparent). Gathering data is more work, but for a business it can be very valuable to decide quickly and correctly (first mover, customers won't have to wait 1 year etc.)

In the case of the school, pouring concrete during session won't work. As for "waiting as late as possible', that darned "possible" is always a tricky business.

I think that would lead to the second design, where there are diagonal paths connecting every building. In the absence of paths people will walk the most convenient (shortest) routes.

Wanting a large green space at the center is an argument against convenient paths. So the decision wasn't about where to which paths are shortest and most convenient for getting from building to building, but about whether or not the paths should be shortest.

That picture is looking down from the Valley library at Oregon State University. Spent many a Adderall fueled nights their.

The problem falls apart in Oregon though. Any paths made into grass would be mud right now - so people would take other, less efficient routes.

Is this really a winning analogy? To let everybody bumble along until patterns appear seems like a myspace strategy. Another option is to put up fences. And set folks on a finite number of paths. And call it facebook.

I don't think Facebook has been immune to this at all - they've certainly put up fences, but have knocked them down quickly as it became apparent that's not what users wnat.

Leaving your app - using the trail analogy - as an open, untamed field may not be the best way to go, but I don't think any app would survive by setting up rigid fences and refusing to adapt them as they apparently become obsolete.

You can't however do without walkways if the people have nothing to walk on, to continue your analogy. (When thinking from a business or tech sense.)

It wouldn't work for me - I just prefer walking on the grass :)

I wonder if you could use AI pedestrians to predict the paths?

You wouldn't happen to know this crazy dude named Ken Ludwig, would you? Because he told me the exact same story, except it involved a lot more rambling.

Reply on news.ycombinator.com