Re: Other areas that haven't been discussed.

New Message Reply About this list Date view Thread view Subject view Author view

From: Ralph Hartley (hartley_at_aic.nrl.navy.mil)
Date: Sat Feb 10 2001 - 14:42:22 CET


Received: (from mdom_at_localhost) by karto.ethz.ch (8.9.3/8.9.3/SuSE Linux 8.9.3-0.1) id OAA14452 for cavexml-outgoing; Sat, 10 Feb 2001 14:28:06 +0100
Received: from spartan.spruceMt.com (spartan.sprucemt.com [209.70.114.2]) by karto.ethz.ch (8.9.3/8.9.3/SuSE Linux 8.9.3-0.1) with ESMTP id OAA14448 for <cavexml_at_cartography.ch>; Sat, 10 Feb 2001 14:28:04 +0100
Received: from aic.nrl.navy.mil (ppp2 [209.70.114.12]) by spartan.spruceMt.com (8.8.7/8.8.7) with ESMTP id IAA02210 for <cavexml_at_cartography.ch>; Sat, 10 Feb 2001 08:37:57 -0500
Message-ID: <3A85453E.3010506@aic.nrl.navy.mil>
Date: Sat, 10 Feb 2001 08:42:22 -0500
From: Ralph Hartley <hartley_at_aic.nrl.navy.mil>
User-Agent: Mozilla/5.0 (Windows; U; Win95; en-US; m18) Gecko/20010205
X-Accept-Language: en
To: cavexml_at_cartography.ch
Subject: Re: Other areas that haven't been discussed.
References: <Pine.GSO.4.05.10102091347540.11013-100000_at_cor.oz.cc.utah.edu>
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-cavexml_at_karto.baug.ethz.ch
Precedence: bulk
Reply-To: cavexml_at_cartography.ch


> If a measurement is marked reliability="godgiven" any program should 
> treat it as fixed, and apply no corrections, adjustments, etc. to it.
> Loop closures should consider its variance to be zero. It is the users
  If you have more than one point with variance zero, then the least
squares adjustment has a singular matrix, and the adjustment is
therefore invalid.

I don't think you understand the ramifications of zero variance.
I think I do. I don't think there are any problems unless there is at least one loop. I should have said that loops specified as goodgiven should not be closed. A program is free to do anything if there is such a loop with non-zero error. Also, "should" is weaker than "must". Not all programs need understand this at all.

Seting reliability="error" is much better, from a data safety point of 
view, than deleting bad measurements from the file, because it is not so
irreversible.

> Survey books cover the correct ways of dealing with the problem.
> Let's not put a hack in to do things some other way.

What is the correct way to deal with data known (not suspected) to be incorrect? You have to mark it in some way. It may be consistant with all other data, so you can't expect the program to detect it (and even if not the program can't tell which data is wrong). You can't delete it because of the data saftey issue, and because it may contain information other than measurements (names of participants etc) that need to be preserved.

Ralph Hartley



New Message Reply About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b30 : Thu Mar 01 2001 - 18:00:00 CET