Seattle Prototype Publicity?
Daniel Carleton
To fill you in, this is Richard and I re: spreading the prototype around more
broadly to get some feedback from Seattleites.
> I think a lot of people could get some use out of it, but, I also
> know that if the link gets out it will generate a /lot/ of interest
> quickly, and I"m not sure how interested you are in being potentially
> inundated with praise, bug reports, and various flavors of internet
> gripes. If you feel ready to handle (or ignore) all that, then I'm
> sure people would like to start playing with it.
What do you think, Wyatt? Should we load test ralph? = )
> Perhaps a forum could be setup for feed back on Seattle specific issues?
Yes, and we should couch the whole thing pretty carefully, I think. Some kind
of obvious indicator that the current instance is a prototype, and that we're
looking for input on what the pathfinding algorithm currently does "wrong."
Wyatt: This could be an immediate lead into our ideas about user reported routes.
First, we get users to go to the site, calculate a route between two points
they ride between every day, and give us feedback like "the trip planner has
you go over this huge hill, but it's better to go around." Next, we do
further data massaging and imports, and then use the user reports like
functional tests. e.g. "Ok, I got elevation data in there and tweaked some
edge weighting in the algorithm; does the trip planner go around the hill
now?"
In fact, we could turn user reported routes into automated tests using an
initial version of the route drawing tool (the one that quantizes map clicks
to the grid). From this we could build a suite of sanity / regression tests
for algorithm development, in addition to tweaking things to get as many to
pass as possible.
These are all very agile-centric ideas. We have the first iteration of
implementation, and now it's time to demo it to the customer and get some
feedback. In this case, our customer is the public. Also, by getting their
feedback on specific routes, we create a means to do behavior / test driven
development. i.e. The users tell us what they want the planner to do, and we
code towards satisfying as many of them as possible.
This feels more pragmatic than some of the more visionary ideas we've had, and
yet builds towards them, I think.
So yeah, I think a little development would need to be done in prep for this.
We need to give the trial users a workflow where they're asked to calculate a
familiar route and then report differences. It could be pretty rudamentary at
first; we'd just need some volunteers to interpret responses and create route
data. = )
Anyway, this all just popped into my head. Let me know what you think.
Cheers,
--
Daniel Carleton <daniel....@pobox.com>
Wyatt Baldwin
> Hey Wyatt and Team,
>
> To fill you in, this is Richard and I re: spreading the prototype around more
> broadly to get some feedback from Seattleites.
>
> > I think a lot of people could get some use out of it, but, I also
> > know that if the link gets out it will generate a /lot/ of interest
> > quickly, and I'm not sure how interested you are in being potentially
> > inundated with praise, bug reports, and various flavors of internet
> > gripes. If you feel ready to handle (or ignore) all that, then I'm
> > sure people would like to start playing with it.
>
> What do you think, Wyatt? Should we load test ralph? = )
If it's only being shown to "insiders" and it's made very clear that it's a
pre-alpha demo, I think it would be good to test the new server AND get some
initial feedback. Other than that, I'd say it's up to you whether you think
the Seattle version is ready for something like that.
(PS to everyone else: "ralph", named after one of our cats, is the name of a new
server we recently started leasing. The Seattle prototype can be accessed at
<http://prototype.bycycle.org/seattle>.)
> > Perhaps a forum could be setup for feed back on Seattle specific issues?
>
> Yes, and we should couch the whole thing pretty carefully, I think. Some kind
> of obvious indicator that the current instance is a prototype, and that we're
> looking for input on what the pathfinding algorithm currently does "wrong."
What kind of forum are you thinking of? I would like to hereby specifically
unvolunteer for setting up such a forum. ;) We could just change the the
feedback/contact links to point somewhere else and create a new mailing list
for Seattle.
> Wyatt: This could be an immediate lead into our ideas about user reported
> routes
<snip>
> So yeah, I think a little development would need to be done in prep for this.
> We need to give the trial users a workflow where they're asked to calculate a
> familiar route and then report differences. It could be pretty rudamentary at
> first; we'd just need some volunteers to interpret responses and create route
> data. = )
>
> Anyway, this all just popped into my head. Let me know what you think.
I think that all sounds great. The only thing I think needs to definitely be
done before announcing the prototype is removing anything that's not going to
work, like selecting other regions and the Portland route preference --
basically strip it down as much as possible.
Aside: In the next version of the trip planner, each region will have its own
template so everything won't be all scrunched together and we can add
region-specific content, links, etc.
~wyatt