latitude and longitude designation and collection data

[Konstantin Savov]

On Oct 2, 21:08, Doug Yanega wrote:
k> Subject: Re: latitude and longitude designation and collection data
k>
k> Konstantin Savov <kps at florin.msk.su>
k> made many good points, but when he says:
k>
k> >1.  You always should find a proper sources to get coordinates of
k> >localities (gazeteers, maps, etc.).  For example, ten years ago I had a
k> >lot of problems with collections even from central part of Russia
k> >(proper maps was not available here, anybody had to get special
k> >permission to work with true maps).  When I received good maps, it
k> >became much easier to make labels for my own collections and find
k> >coordinates for old herbarium.
k> >
k> >It's very useful to get a digital map with settlements, rivers, etc.
k> >But it is more important to have good maps and gazeteers in any form,
k> >from my point of view.  In any case, you should spent a lot of time
k> >searching for a point needed on a map or checking gazeteers - it is very
k> >hard to automate the process.
k>
k> this assumes the luxury of a manageable amount of material. For a

Unfortunately, it's only possible to deal with manageable amount of
material. :-)  If you have very large collection, you can work with
relatively small parts of it (e.g., involved in some projects).

k> collection like ours, with over 6 million insects, using maps is
k> impractical - it would almost be better, faced with this, to wait a decade
k> or two for fast, "smart" software and terabyte memory CD-ROM maps than to

Have you over 6 million _insects_ or 6 million of _localities_?  In
order to find lat/long you should work with localities.   Number
of localities is usually less than number of specimens.  So, it's
very important to use lat/long found for locality each time you have
specimen collected there.

Is digital map supported by GIS software better than printed one?  It
highly depends on software quality and hardware resources available.
That's another story.  Actually, you can have just now very powerful and
"smart" system.  You should get powerful database server (e.g., HP or
DEC) and graphical workstations (e.g., Sun or SGI) with both database
and GIS running under Unix.  I have software solutions for such hardware
and operation systems.  You can get or prepare yourself a digital map
needed (incl.  settlements, roads and topography and some other
information you need).  Your idea concerning measuring distances along
curved lines is very interesting and can be realized in GIS-application
in rather short time.

As a result, you'll have a computerized version (rather expensive one)
of the same map you are using now in paper form.  Plus some additional
tools.  You'd save a lot of time using those tools, but you'll have to
perform just the same operations while searching for lat/long of a
locality.  You'll just work with different media.

k> spend that time finding data manually and be less than 1/5th finished (if
k> we assume 1 technician working 40 hours a week, able to look up 20
k> localities per hour - which is generous - it would take 144 years to finish
k> the job). Until the operation can be performed in a matter of seconds per

So, your only hope is a set of algorithms for finding a proper
geographical object (e.g., point with lat/long) on a map using text
written on a label.  If I understand you correctly, you need automatic
text recognition, lexical analysis and very intellectual search through
huge databases of geographic names.  All options mentioned should be
included in one package.

Well, you'd get it five or ten years later.  I'm not sure such a system
will help you significantly.  Text recognition wouldn't work with
handwritten labels well enough, analysers will have problems with
different languages, absence of plain text on labels, shortcuts and so
on.  You'll also have a lot of problems with old names, alternative
forms of the same name, etc.  while searching for geographic object
needed in a database.  Surely, you'll have to check your data manually
in any case.  Any software tools will help you with rather simple
operations.  However, it's not bad.

k> locality, we're probably better off not bothering looking up lat/long data
k> on old material as a general practice, but rather leave it (as we tend to
k> now) to the specialist who is doing a biogeographic study with a given
k> taxon.

That's right!  This point is closely related to the recent discussion
concerning curating and cataloging.  I think it's only possible to treat
huge amounts of material step-by-step.  Electronic catalogs and
lat/long data would be a result (and not the main one) of other
projects.
If any project involving data
from your collection will add some information to collection database,
you would have rather large database in several years.  It'd be also
workable to enter all new collections in a database obligatory.  It
would be acceptable, if you have the following at your site:

        - _multiuser_ database available for all specialists working
        with your collection (incl. Internet access via online,
        Mosaic, etc.),

        - software, which helps you in collection management
        (printing labels, loans registration, etc.), so, it returns
        time needed for entering data in a database;

        - software, which is useful for every-day work of different
        kinds (simple system designed for cataloging or lat/long data
        only is not suitable for other tasks, as a rule),

        - data exchange.

Coordinates are only a small part of this problem.  In order to work
with lat/long, you should have a database with localities and
specimens stored.

k>         As for new material, we've obtained a GPS unit here, and folks
k> occasionally remember to bring it with them in the field, but it's going to
k> take some time before it becomes routine. At least we can print labels
k> small enough now that adding lat/long data won't make the labels too
k> cumbersome (there's just so much you can squeeze in 10 X 15 mm)...all sorts

That's a problem.

k> of practical difficulties arise with insects that don't quite plague other
k> taxonomists, at least not to the same extent.
k> Sincerely,
k>
k> Doug Yanega       Illinois Natural History Survey, 607 E. Peabody Dr.
k> Champaign, IL 61820 USA      phone (217) 244-6817, fax (217) 333-4949
k>  affiliate, Univ. of Illinois Urbana-Champaign, Dept. of Entomology
k>   "There are some enterprises in which a careful disorderliness
k>         is the true method" - Herman Melville, Moby Dick, Chap. 82
k>-- End of excerpt from Doug Yanega



--
+------------------------------------------------------------------------------+
|        Konstantin Savov        |     DataX/FLORIN, Inc.    Moscow, Russia    |
|                                |       Advanced Data Management Systems      |
|   E-mail:  kps at florin.msk.su   |   Voice: (095)158-9520  Fax: (095)158-5700  |
+------------------------------------------------------------------------------+