Re: Cross Section libraries

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

From: martinl_at_talk21.com
Date: Sun Jan 21 2001 - 01:22:59 CET

Next message: Michael Lake: "Re: LRUD: 'From' or 'To' ?"
Previous message: Andreas Neumann: "National delegates"
Maybe in reply to: Julian Todd: "Cross Section libraries"

From: martinl_at_talk21.com
To: cavexml_at_cartography.ch
Date: Sun, 21 Jan 2001 00:22:59 GMT
Subject: Re: Cross Section libraries
Message-Id: <20010122024654.NIZS24947.t21mta03-app.talk21.com@t21mtaV-lrs>
Sender: owner-cavexml_at_karto.baug.ethz.ch
Precedence: bulk

Martin Sluka wrote: ".. Try to think about something as "library of cross section" .. In real cave there are not endless kinds of profiles."

Julian Todd, having provided this facility in Tunnel, was a bit chastened by the experience and pointed out that this was valid as a timesaver for data input, but that the resultant models should be interpreted with caution.

Ken Grimes fulminated: "Disagree strongly !!! As a geomorphologist, the thought of a proscribed" (sic - he meant the opposite: prescribed) "set of 'standard' cross-sections is a can of worms. The possible variations are infinite! ... "

--> Yes, but the verbal shorthand descriptions of normally adequate approximations are quite limited. There are certain cross sections which seem eminently suitable for inclusion in a library and are scalable viz: borehole/tube (elliptical morphology, generally phreatic development); parallelogram (rectangle or rift morphology, often breakdown and/or vadose and/or paraphreatic development); with parameters common morphologies like keyhole, figure of eight, and laterally grooved should be tractable - at least for use by the general caver and for directing the specialist to places of potential interest. These shapes often persist for a significant number of survey legs as is reflected by guide book descriptions and in most drawn-up surveys where the cross-sections are shown only for a small percentage of the 'stations' surveyed yet represent the cave adequately (actually I have often drawn sections for the passage part way along the leg being surveyed where it is best seen, and Julian Todd has pointed out case
s where it may be more convenient to draw sections along legs which cannot actually be surveyed, as when you doglog around an obstruction which bisects the passage).

Ken worries that "Non-geological Cave mappers who assign a "standard" section "type" instead of sketching what is really there could confuse things badly..."

--> Only if the users of the survey are incautious enough to check the claimed grade of the survey and assume that it was drawn for detailed research use.

He continued that the "ultimate objective of all this is to draw actual MAPS for peole to LOOK at - not to generate lots of computer files. The data files should always be accompanied by sketches of the details (scanned if needs-be, but it must be an true image not an over-simplified or 'standard' approximation!)..." and then says "views using LRUD etc are only usefull for SMALL-SCALE maps giving an overview of a whole cave. Now, this IS a usefull end (especially for large caves), but it needs little more than LRUD for the plot scales involved." which seems unexceptionable and consistent with "auto-rendered plots can never be used to show the fine details of wall-nitches, alcoves, bell-holes, spongework etc... and we should not pretend they can... "

--> I doubt if many, if any, cave surveyors would dream of claiming that they though the routine depiction of any such detail possible other than by textual annotations or in a cave description.

Continuing "Sketching of wall detail (including sections) and contents is still the essential part of any cave mapping exercise - the numeric data is just a positional framework that we hang those details onto. "

--> Whoa! I have noticed that many Australian caves are on the small side and have meticulously detailed surveys, but I suspect most (caver's) cave surveys (archaeological, geological and geomorphological surveys are clearly more specialised) are done mainly for the centre line skeleton and the flesh of wall detail is a nicety.

Ken concludes by suggesting that: "..XML data should stick to a simple LRUD set (for use in small-scale renderings), possibly with pointers ... to the locations of detailed sketches ... which are held in separate (scanned?) files or paper-drawings."

I would like to repeat my view that a CrossSection construct allowing an arbitrary number of points, or a reference to a standard scalable shape giving such a set of points, with optional scale and orientation parameters is better as it is capable of accomodating both 'legacy' LRUD data (which should, logically, also have an indication of the shape it intends to convey from diamond or rectangle) and detailed data such as he would like to have for his geo work.

Finally, Ken concludes "The best way of getting the sketched details into a computer file is via CAD software
..perhaps it *might* be possible to write programs (specific to each CAD package!) to auto-generate an XML version of the CAD image ..Most people would not have the time to do it."
This is unduly pessimistic; I don't think we NEED CAD programs at all. You can, today, take a curve scanned as, for example, a jpg or gif and have a text representation of it (not quite in the format I suggest, but close) generated automatically in the XML dialect known as SVG. A tool which does this is CR2V, freely available from http://www.celinea.com/

--------------------
talk21 your FREE portable and private address on the net at http://www.talk21.com

Next message: Michael Lake: "Re: LRUD: 'From' or 'To' ?"
Previous message: Andreas Neumann: "National delegates"
Maybe in reply to: Julian Todd: "Cross Section libraries"

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

This archive was generated by hypermail 2b30 : Wed Feb 14 2001 - 00:03:52 CET