Page 2 of 2 FirstFirst 12
Results 16 to 17 of 17

Thread: Android Hiking App

  1. #16
    Senior Member DougPaul's Avatar
    Join Date
    Jan 2005
    Location
    Bedford, MA; Avatar: eggs anyone?
    Posts
    10,608
    Quote Originally Posted by Trail Boss View Post
    To say there is "no fundamentally correct value" is fine for philosophers but not cartographers. At the end of the day, "no correct value" is replaced by "good enough for all practical purposes".
    Cartographers and surveyors are very aware of accuracy issues. For instance, there are published accuracy specs for USGS topos. And a surveyor should be able to tell you the accuracy of any point that he has located.

    I agree with you in principal but in practice there is a point where the fractal aspect makes no appreciable difference for hikers. It may not be possible to agree on a trail's distance or ascent with millimeter accuracy but it is possible to arrive at something to within (tens of) meters ... which is "good enough" for the purposes of hiking.
    My point is that one should be aware of the errors whatever they are.

    IMO, a few tens of meters would be pretty optimistic for a full day's hike with lots of twists and turns and ups and downs.

    You have presented a method for determining total climb that has at least one obvious weakness. If you are happy with it, fine, but IMO your case is weak. In some terrain, it is probably good enough, but in some other terrain it is likely not very good.

    In the absence of a fundamentally correct value or baseline data with a proper error analysis showing errors that are "small enough" it is very hard to show that one proposed method is more accurate than another proposed method. You may only be able to show consistency between two methods--consistency is necessary, but not sufficient to show accuracy.

    Other items:
    * Many GPS users don't know their limitations or how to use them effectively.
    * In ideal conditions (good skyview, good satellite constellation etc) most GPSes will give similar results. In less than ideal conditions (eg when hiking) the poorer units become less accurate. (For example, I have made a direct comparison between an Iphone 4s and a Garmin 60CSx on a hike--the 4s track showed some unreasonable large excursions from our route, the 60CSx track did not.) Many of those unrealistic trip computer numbers likely come from such excursions.

    Yawn,

    Doug

  2. #17
    Senior Member Trail Boss's Avatar
    Join Date
    Sep 2003
    Location
    Montréal, QC
    Posts
    214
    You stated there's "no fundamentally correct value" (here and in other posts) like it's the unattainable pursuit of the Holy Grail because, fractals. There is an acceptable value as far it pertains to hiking and, for the average hike, it only needs to be accurate to within tens of meters (arguably even a half klick) to be more than acceptable to the vast majority of hikers. Certainly not good enough for other activities, like sprinting or surgery, but more than adequate for a 15 km hike.

    ... IMO your case is weak. In some terrain, it is probably good enough, but in some other terrain it is likely not very good.
    I welcome the opportunity to learn about the methodology and data you collected that supports your claim.

    In the absence of a fundamentally correct value ... it is very hard to show that one proposed method is more accurate than another proposed method. You may only be able to show consistency between two methods--consistency is necessary, but not sufficient to show accuracy.
    Sure but in this case the "fundamentally correct value" has traditionally been extracted from a topographic map. Land surveys, photogrammetry, radar altimetry, etc helped produced the topos that continue to be used by hikers to this day. It serves as the long-standing baseline and used to calculate ascent values published in (older) guidebooks (... unless they employed an altimeter).

    If a barometric altimeter reports elevation and total ascents consistent with topo then it has established a degree of trust. If a consumer GPSr tells me the total ascent is substantially higher (like by 25-50%) then I have good reason to doubt it. Especially when it does that for even level stretches of terrain. The route is self-evidently flat yet the device records an appreciable ascent. Well, well, that's a nasty little characteristic that doesn't inspire confidence.

    This whole thing started when I hiked to Seymour and the app pulled an extra 300 meters of ascent out of thin air. Most of that was accumulated over the initial 8km of level/gradually-ascending terrain. Completely unsupportable by topo map and, especially, by the amount of expended effort. Screw it. I switched to a DEM because it produces ascent values consistent with topo and altimeters.
    Last edited by Trail Boss; 06-15-2017 at 08:26 PM.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •