From: P A Hill & E V Goodall (goodhill_at_xmission.com)
Date: Wed Mar 13 2002 - 03:23:09 CET
Return-Path: <owner-cavexml-outgoing_at_ethz.ch> Delivered-To: cavexml-archive_at_cartography.ch Received: from localhost (localhost [127.0.0.1]) by karmail.ethz.ch (Postfix on SuSE eMail Server 2.0) with ESMTP id EAABD13C25 for <cavexml-outgoing_at_ethz.ch>; Wed, 13 Mar 2002 03:20:20 +0100 (CET) Received: by karmail.ethz.ch (Postfix on SuSE eMail Server 2.0, from userid 28) id 0A58D13BDE; Wed, 13 Mar 2002 03:20:18 +0100 (CET) Delivered-To: cavexml-loopcheck_at_ethz.ch Received: from localhost (localhost [127.0.0.1]) by karmail.ethz.ch (Postfix on SuSE eMail Server 2.0) with ESMTP id 0D28313C25 for <cavexml-loopcheck_at_ethz.ch>; Wed, 13 Mar 2002 03:20:17 +0100 (CET) Received: by karmail.ethz.ch (Postfix on SuSE eMail Server 2.0, from userid 96) id 1F6E513881; Wed, 13 Mar 2002 03:20:14 +0100 (CET) Delivered-To: cavexml_at_cartography.ch Received: from localhost (localhost [127.0.0.1]) by karmail.ethz.ch (Postfix on SuSE eMail Server 2.0) with ESMTP id C5DD313C25 for <cavexml_at_cartography.ch>; Wed, 13 Mar 2002 03:20:13 +0100 (CET) Received: from mgr1.xmission.com (mgr1.xmission.com [198.60.22.201]) by karmail.ethz.ch (Postfix on SuSE eMail Server 2.0) with ESMTP id 947D4114F2 for <cavexml_at_cartography.ch>; Wed, 13 Mar 2002 03:20:10 +0100 (CET) Received: from [198.60.22.22] (helo=mail.xmission.com) by mgr1.xmission.com with esmtp (Exim 3.22 #1) id 16kyLj-00024J-00 for cavexml_at_cartography.ch; Tue, 12 Mar 2002 19:18:43 -0700 Received: from goodhill.dsl.xmission.com ([198.60.114.17] helo=xmission.com) by mail.xmission.com with esmtp (Exim 3.22 #1) id 16kyLj-0003Vq-00 for cavexml_at_cartography.ch; Tue, 12 Mar 2002 19:18:43 -0700 Message-ID: <3C8EB80D.A1F2511A@xmission.com> Date: Tue, 12 Mar 2002 19:23:09 -0700 From: P A Hill & E V Goodall <goodhill_at_xmission.com> X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en To: cavexml_at_cartography.ch Subject: Re: "Ignore" attribute References: <20020312115138.9EE4113BF0_at_karmail.ethz.ch> <032201c1ca29$903448a0$0801a8c0_at_matt> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Loop: cavexml Sender: owner-cavexml_at_karmail.ethz.ch Precedence: bulk Reply-To: cavexml_at_cartography.ch X-Virus-Scanned: by AMaViS perl-11
Matthew Ryan wrote:
> Seeing as no-one else has mentioned it, Survex has the ability (at least I
> think) to specify the estimated Standard Deviation of each value or for a
> set of values, this is then a figure which cannot be misinterpreted and can
> be used in the loop closure code.
Oh? So you think the average surveyor can make a reasonable estimate of
standard deviation? I'd be a bit suspicious.
> Survex provides a set of include files
> for the to map from standard BCRA instrument reading grades to the Standard
> Deviation of the recorded values.
That certainly makes sense and certainly should be part of an overall solution,
but after all that is pretty much the same solution as a set of named possible
choices at the shot level.
> This would seem to be much more elegant solution than anything else which
> has been suggested so far for this and would also avoid throwing away data
> from Survex files if converted.
I'm certainly for not throwing away data. This is possibly another example
of how data is represented in its original form and how it is converted to
standard units/form. A caver might be asked to identify bad and worse shots
but this really comes down to a STD based on some combination of
the basic survey accuracy and how bad the particular shot is.
> Eventual input software which generates the XML can help with converting
> confidence = word to SD = number
It seems I'd be throwing away data if I inserted a Standard Deviation
without retaining the idea that it was just a value picked because of
the assumed range of quality of the value, such an assumed quality may not
apply.
For example, play with this set of repeat readings. I've listed
STD=2.0 assuming the original input marked these shots as of less
quality than normal shot, but not extremely bad.
<azimuth value=330.5 STD=2.0 confidence="low">
<azimuth value=338 STD=2.0 confidence="low">
<azimuth value=326 STD=2.0 confidence="low">
Okay, maybe the original caver overrated his/her ability to
read under bad conditions, but don't you think this is how
real data will be presented?
Is it the author of any software that thinks 2.0 is a reasonable
guess about STD which is in error? Should "low" be a broader range
like 3.0 or 5.0? Who actually worries about 'slightly bad' readings;
it is probably the really bad that get marked in survey notes.
So maybe the categories are:
normal -- typical shot for grade of survey
poor -- obviously a hard shot
bad -- sorry but I'm sure this is off the quality of the rest of the
survey, but we couldn't do better than this poor estimate.
Should a value be calculated from the data given and if it is
bigger than the stated accuracy, a larger number for STD inserted?
Well, I guess an XML format that allows some type of marking
of not so good shots, plus a place to put a numerical value
would cover each of
1. a fixed 2.0 based on the stated quality,
2. a calculated value based on an estimate from the actual data given,
or
3. a re-stated value based on a data processor bumping the accuracy
into the next worse range based on an examination of the data.
Just playing with ideas.
Any thoughts?
-Paul
This archive was generated by hypermail 2b30 : Sun Mar 31 2002 - 23:00:01 CEST