2012.04.19 Weekly Check In

• [13:40] <FrankP> .
• [13:40] <FrankP> what do folks think about the line "Least traffic" in place of "Safest" for the bike triangle? http://maps5.trimet.org/otp.html?mode=BICYCLE
• [13:41] <mele_> YES
• [13:41] <FrankP> Should I make this an OTP project change, or local to TriMet?
• [13:41] <andrewbyrd> streets with designated bike lanes don't necessarily have the least traffic...
• [13:42] <FrankP> (for English...possibly Spanish in the future)
• [13:42] <mele_> yes but our weights already account for that
• [13:42] <mele_> secondary, primary roads have more traffic & higher penalties
• [13:42] <mele_> i think that's the description we've been looking for
• [13:43] <novalis_dt> The only difference is to the description?
• [13:43] <FrankP> It's the Marketing department asking for the change...I could honestly give a rat's ..., but...
• [13:43] <mele_> no i think it's a good idea
• [13:43] <mele_> safest has never been quite right for me
• [13:44] <novalis_dt> It certainly clarifies that you're not talking about odds of being mugged.
• [13:44] <mele_> right, without having to put a separate line in the FAQ about what "safest" means
• [13:45] <FrankP> Less traffic == dark alley (sometimes)
• [13:45] <novalis_dt> I would prefer to keep "safest" because I think otherwise people will think we are incorporating realtime traffic data
• [13:45] <novalis_dt> Maybe we should just call it "Bike lanes"
• [13:45] <mele_> but it's not bike lanes....
• [13:45] <novalis_dt> Well, it's partly that.
• [13:45] <grant_h_> what if we used "bike friendly"
• [13:46] <novalis_dt> That sounds good.
• [13:47] <mele_> works for me... now will it work for marketing?
• [13:47] <FrankP> Imagining the possible translations of "Bike friendly" makes me laugh a bit.
• [13:47] <novalis_dt> Contains hoards of helpful bicycles just waiting to carry you wherever you would like to go!
• [13:47] <FrankP> Bicellete mon amore
• [13:48] <mele_> bicycle-loving streets?
• [13:48] <novalis_dt> (sorry, hordes)
• [13:48] <novalis_dt> No, maybe I just can't spell
• [13:49] <novalis_dt> Anyway
• [13:49] <novalis_dt> I would take "bike friendly" in trunk, but I just don't want something that implies that we know anything about realtime traffic.
• [13:49] <FrankP> スーパー素晴らしいバイクのパートナーと親友
• [13:49] <novalis_dt> And I'll take anything at all if I am outvoted :)
• [13:49] <FrankP> Super terrific bike partner and best friend
• [13:49] <novalis_dt> Wakarimasen!
• [13:51] <novalis_dt> Anyway, so I've been back at area routing. Thanks so much for all the good examples
• [13:51] <novalis_dt> They're really helping me smoke out bugs
• [13:51] <novalis_dt> Especially http://www.openstreetmap.org/browse/relation/1967631
• [13:52] <grant_h_> haha, yeah I mapped some extra features in that area to make it a little more interesting
• [13:53] <novalis_dt> Fortunately, a bug in OTP meant that I redownloaded those tiles!
• [13:53] <novalis_dt> (yes, I will be fixing that bug)
• [13:55] <novalis_dt> Anyone else have anything!
• [13:55] <novalis_dt> er ?
• [13:55] <FrankP> Dave, is there anything I need to do to activate areas? My lastest /test graph seems to snap to roads (1000' feet away) in the few tests I've tried.
• [13:56] <novalis_dt> snap what to roads?
• [13:56] <novalis_dt> oh, if your starting point is inside an area?
• [13:56] <novalis_dt> The version of OTP that's on git only does area=yes ways
• [13:57] <FrankP> Ah, right ... areas have to be defined in OSM...sorry, not sure if Forest park / Zoo are defined as such.
• [13:57] <novalis_dt> Not relations with type=multipolygon area=yes
• [13:57] <novalis_dt> Try Pioneer Square
• [13:57] <mele_> that's a little iffy though because I already drew footways all through it
• [13:57] <novalis_dt> oh.
• [13:58] <mele_> ... so it's not really a pure area
• [13:58] <mele_> i drew a lot of footways in ped areas over the last several months to make sure they'd work...
• [13:59] <FrankP> yeah, that works...
• [13:59] <FrankP> (or maybe it's one of Mele'
• [14:00] <novalis_dt> You can actually route through the Starbucks in the northwest of Pioneer Square
• [14:00] <novalis_dt> There isn't a footpath there of course
• [14:09] * demory (demory@nat/openplans.org/x-essgtrxlacnenuqk) has joined #opentripplanner
• [14:09] <novalis_dt> hey demory. We were about finished up, but if you want to jump in, this is the moment!
• [14:09] <demory> hey folks, sorry, had a meeting run long
• [14:10] <demory> i'll check in quickly
• [14:11] <demory> mainly working on next stage of Deployer work, including experimenting with multiple graph OTP to allow sharing of instances between deployments
• [14:11] <demory> also working on the new website still, hope to launch that for real in the next week
• [14:12] <demory> that's really it for me -- otherwise just busy w/ meetings in NYC this week
• [14:12] <demory> novalis_dt can you send me a transcript of the full chat?
• [14:12] <novalis_dt> demory, sure.
• [14:12] <demory> great, thanks
• [14:12] <novalis_dt> And I just got what looks like a better set or area routes: http://novalis.org/bugs/vizgui-rose-quarter.png
• [14:13] <novalis_dt> (notice that there are no paths from the convex points of the multipolygon except where those have connected outer edges
• [14:14] <kpw> novalis_dt, how are things looking now on area routing? more work to be done or are you feeling good about the progress at this stage?
• [14:14] <novalis_dt> kpw, I could commit this now and we would have an improvement over what is in trunk
• [14:14] <novalis_dt> (as of about three minutes ago)
• [14:14] <novalis_dt> I have a few more things to check
• [14:14] <kpw> ok
• [14:15] <kpw> just curious what the next steps are
• [14:15] <novalis_dt> (1) Matt pointed out an area in CA that was broken; I need to check it with the new code.
• [14:15] <novalis_dt> It might be fixed by Laurent's epsilon change
• [14:15] <novalis_dt> (2) I should probably do an all-pairs SPT across only the points connected to outgoing edges and remove unused edges
• [14:16] <novalis_dt> If you look at the link, you can see that there are many edges connected to that concave areas on the rightof the larger shape.
• [14:16] <novalis_dt> Some of those are necessary, but most are not
• [14:17] <FrankP> We had a TriMet employee complain rather loudly that OTP was giving weird results…he was typing time as 133 pm … I added code to hopefully handle such situations. (Also handles military time now). http://maps5.trimet.org/otp.htm
• [14:18] <FrankP> (BTW, prior to yesterday, 133 pm would probably default to now and/or ???)
• [14:18] <novalis_dt> Saw that checkin -- excellent. Now we just need to handle "teatime" and "beer o'clock"
• [14:19] <FrankP> Beer thirty
• [14:19] <novalis_dt> I have been very patiently waiting to start hacking on the geocoder because there are so many good fictional addresses to add as easter eggs.
• [14:20] <andrewbyrd> and what we call military time in the US is what the rest of the world uses for all public transit schedules :)
• [14:21] <FrankP> Yeah, I'm surprised at all the talk about dates recently that no one's mentioned time...
• [14:21] <kpw> novalis_dt, great! do you have a sense of how much more time you'll need to wrap up those items?
• [14:21] <kpw> i'd love to see this roled into our next release if possible
• [14:22] <novalis_dt> kpw, I'm going to look at Matt's area right now. That's the one with unknown unknowns
• [14:22] <novalis_dt> The all-pairs shortest-path reduction thing will take less than a day.
• [14:23] <novalis_dt> But first I will commit all of these changes!
• [14:24] <novalis_dt> Oh, andrewbyrd, what's the command to run enunciate again?
• [14:30] <andrewbyrd> novalis_dt, it's bound to the site phase. should be "mvn site", but let me check
• [14:30] <FrankP> demory, we're going to change "Safest" to "Bike friendly" in the triangle, ala http://maps5.trimet.org/otp.html?mode=BICYCLE ... unless someone objects.
• [14:31] <FrankP> (global change to the English translation ... Spanish may follow)
• [14:31] <mele_> don't forget the drop-down too
• [14:32] <mele_> and were we ever going to change the colors? we'll talk about it later I guess. we had some feedback from sight-impaired that the yellow is very hard to see
• [14:33] <mele_> & give everything a little more contrast from background
• [14:33] <andrewbyrd> novalis_dt, yes it's, mvn site, resulting docs are in OpenTripPlanner/target/site/apidocs
• [14:33] <novalis_dt> thanks
• [14:34] <demory> FrankP, sounds good to me (re "bike friendly")
• [14:34] <FrankP> thanks...so let it be written then...
• [14:35] <demory> so the others will remain "Quickest" and "Flattest"?
• [14:36] <demory> Or should they be just "Quick" and "Flat" to go along w/ "Friendly"?
• [14:36] <novalis_dt> Maybe we should change them to "Quick" and "Flat"
• [14:36] <novalis_dt> Jinx!
• [14:36] <demory> ha
• [14:36] <mele_> yeah i think that's better too
• [14:37] <FrankP> What does our poet laureate think...Grant?
• [14:37] <demory> we are both believers in parallel construction I see
• [14:37] <grant_h_> sounds good to me!
• [14:37] <FrankP> (Grant came up with Bike friendly)
• [14:38] <FrankP> Should I also change the drop down to Quick trip ?
• [14:38] <novalis_dt> Sure.
• [14:38] <FrankP> (Applies to transit)
• [14:38] <novalis_dt> Also, you should change "fewest transfers" to a checkbox
• [14:38] <novalis_dt> Since that's what it is internally
• [14:39] <novalis_dt> (well, internally it's a penalty, so it would really be a slider, but the checkbox is saner)
• [14:39] <FrankP> Hm...so a number slide on the number of transfers you're willing to endure
• [14:40] <novalis_dt> Well, there's that, but that's a hard limit and we tend to prefer soft limits
• [14:40] <novalis_dt> The other is the transferPenalty
• [14:40] <novalis_dt> Which is how much boardings beyond the first cost (in addition to any other of their costs)
• [14:41] <andrewbyrd> ideally we should have a bank of sliders labeled "transferring is as annoying as walking for X minutes", etc. :)
• [14:41] <novalis_dt> And then you'll need a Ph.D in transportation planning to use our software...
• [14:41] <novalis_dt> :)
• [14:41] <FrankP> Phd ... push here, dummy
• [14:44] <andrewbyrd> "Advanced options: I personally find that the annoyance of transferring grows (check one) linearly, geometrically"
• [14:45] <mele_> hah
• [14:45] <novalis_dt> "In accordance with a function which I will now draw in this canvas"
• [14:45] <novalis_dt> "Let me just upload a Mathematica worksheet"
• [14:53] <FrankP> One last time...this is what I'm going to check in to the English translation: http://maps5.trimet.org/otp.html?mode=BICYCLE
• [14:54] <mele_> great
• [14:54] <novalis_dt> The sliders seem broken for me?
• [14:54] <novalis_dt> oh, they're not sliders
• [14:54] <novalis_dt> Duh.
• [14:55] <novalis_dt> Hm. I wonder how many other people try that.
• [14:55] <novalis_dt> (otherwise looks good)
• [14:56] <demory> hmm, that had never occured to me but I can see how one might make that assumption
• [14:57] <demory> perhaps something less slider-like would be better. let me give it some thought..
• [14:58] <novalis_dt> Of course adjusting one slider would have to do something to the other two, and it's not really clear what
• [14:59] <andrewbyrd> keep the combination affine while respecting the ratio between the two sliders you were not moving.