From: Paul & Eleanor (goodhill_at_xmission.com)
Date: Thu Jan 18 2001 - 08:32:20 CET
Received: (from mdom_at_localhost) by karto.ethz.ch (8.9.3/8.9.3/SuSE Linux 8.9.3-0.1) id IAA06915 for cavexml-outgoing; Thu, 18 Jan 2001 08:29:23 +0100 Received: from mail.xmission.com (mail.xmission.com [198.60.22.22]) by karto.ethz.ch (8.9.3/8.9.3/SuSE Linux 8.9.3-0.1) with ESMTP id IAA06911 for <cavexml_at_cartography.ch>; Thu, 18 Jan 2001 08:29:22 +0100 Received: from slc436.modem.xmission.com ([166.70.2.182] helo=xmission.com) by mail.xmission.com with esmtp (Exim 3.12 #1) id 14J9VY-0005AD-00 for cavexml_at_cartography.ch; Thu, 18 Jan 2001 00:29:20 -0700 Message-ID: <3A669C04.B2AE7A13@xmission.com> Date: Thu, 18 Jan 2001 00:32:20 -0700 From: Paul & Eleanor <goodhill_at_xmission.com> X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en CC: cavexml_at_cartography.ch Subject: Re: LRUD: 'From' or 'To' ? References: <3.0.1.32.20010116181528.00693dac_at_mail01.hml.aone.net.au> <00d401c08080$f7fa7750$0a3c10ac_at_gordon> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-cavexml_at_karto.baug.ethz.ch Precedence: bulk
Julian Todd wrote:
>
> I've done quite a bit of development on this topic.
Looks like good stuff. The software looks interesting.
The cross section idea is worth looking into separately.
Years ago I decided that the key point of any bit of data
for a survey was to answer whether it goes with a point
a (shot) line, the whole segment, a survey trip etc.
In this case I think we are talking about something connected to
a point, but with other points referenced to define the
data completely.
> Absolutely should not have either convention proposed.
Maybe one could be the default if nothing is specified?
After all we are talking about optional tags right? Or is
this an all or nothing thing?
I guess that certainly is a semantic question for any groups of tags.
> Obvious adjustments made to account for one or other or both of
> station_fore_location and station_back_location
> being null.
That is just like the 'from' or 'to' convention isn't it? But
requires explicit station to be repeated. Do we want to require
such forms as the only way to do it, or is there a way to spec. either
a simple alternative or a degenerate form of this?
Julian, in your case what is the orientation when neither of
the other related points are specified?
You can certainly define cross sections that match any
of the TO/FROM conventions by specifing all numbers,
but I bet others might want to see what a degenerate 'none specified'
or 'some specified' looks like in XML.
> Furthermore, LRUD defines a rectanglular cross section.
> Why not have a cross section that is a loop made from
> a series of points (like the kind you draw on your survey)?
> Surveys that start as LRUD can write this into XML as a
> rectangle or diamond: { (R, 0), (0, -D), (-L, 0), (U, 0).}
Why not? Because that is twice the values for the 'old fashion'
case. Since around the USA there are actually over 1000 miles
(1600 kms) of cave data using the 'dumb' LRCF formats, it might
be practical to service the obvious without going for twice the
number of values.
Consider those caves on long and deep cave list. We should try
for something compact if possible.
> This will not preclude people recording proper
> cross sections instead.
We certainly shouldn't stop that!
How to include the complex and not burden the simple?
That seems like a worthy goal?
Is it just handled with John Hallecks model? Simple
is for input, the "raw numbers" 1st thing that looks like XML
just specifies everything. That sounds a bit excessive to me.
Maybe lots of defaults if a value isn't present?
Two forms? Plain old LRCF, plus another tag for fully spec'ed
polygon?
Might there be something that says
'look here comes a whole bunch of shots with cross sections, if
it doesn't say otherwise, the defaults are ....' It would be
in that construct that someone somehow defines a 'from' like style,
a 'to' style, or a half way in between style, for the general case.
Then at a paricular point if there is lambda and cross section
inclination specified, those who want to can deal with that.
Note also that an XML that defines a series of stations and shots
has an order, there is often an obvious previous and next station.
This is not necessarily the case with a database of vectors
defining a set of connected points.
> The problem with the standard conventions is that they are
> unbelievably restrictive and force surveyors to
> record what is clearly nonsense a lot of the time
> (eg LRUD in a tall hading rift).
I always had that problem myself when surveying. I've seen
more than one person independantly come up with a notation for
things like "open passage in that direction", "actually
these _two_ numbers better characterize a particular dimension
of the LRCF when combined with the cross-section sketch" and
that was just in the survey book in the cave!
> Sensible cross sections need not be in the vertical plane,
> or centred on a station. They can be halfway along a leg,
> and be tilted with the general dip of the passage.
Note, in classic surveys I have often recorded extra
cross-sections, even without any desire for neat computer
maps, but mearly to help the draftsman. I'm sure this is
common enough.
I think I'll make up a strawman that defines some examples.
-Paul
This archive was generated by hypermail 2b30 : Wed Feb 14 2001 - 00:03:52 CET