Path: utzoo!attcan!uunet!lll-winken!lll-tis!helios.ee.lbl.gov!pasteur!ames!mailrus!tut.cis.ohio-state.edu!bloom-beacon!apple!rutgers!bellcore!clyde!watmath!watcgl!awpaeth
From: awpaeth@watcgl.waterloo.edu (Alan Wm Paeth)
Newsgroups: comp.graphics
Subject: Re: World map (now: WDBII)
Message-ID: <5831@watcgl.waterloo.edu>
Date: 14 Sep 88 16:50:25 GMT
References: <14371@hc.DSPO.GOV> <1197@cleo.SW.MCC.COM>
Reply-To: awpaeth@watcgl.waterloo.edu (Alan Wm Paeth)
Organization: U. of Waterloo, Ontario
Lines: 26

In article <1197@cleo.SW.MCC.COM> peterson@cleo.SW.MCC.COM (James Peterson) writes:
>In article <14371@hc.DSPO.GOV>, siegel@hc.DSPO.GOV (josh Siegel) writes:
>> Does anybody have a polygonal database of the world?

On a related note, I have some C code used to compress down the World Data Bank
II (six million vertices) tape set into a more managable size. The original
data is encoded in EBCDIC on five (5) 2400' mag tapes at 1600bpi, using FORTRAN
style card image records of twenty bytes. Records encode vector chains, which
are not polygons (unlike WDB I).

My C code generates a companion bounding box entry for each chain, and reduces
the vector data to just over two bytes per vector (one byte each for delta-x
and delta-y vector motions, with the occasional long chains fragmented) With
the ~10:1 data reduction the entire set can now live on one volume, or at
6250bpi on a runt 600' tape.

Although the WDBII data is detailed, it is clearly the work of lots of monkeys
sitting at digitizing tables -- the chains have no rhyme or reason, and even
single vector chains which form the boundaries of small lakes and islands are
not guaranteed to close (make that "are guaranteed not to close").

If there is enough interest I will post the software; I have no desire (nor
liscensing authority) to send off the data _per se_, excerpted or otherwise.

    /Alan Paeth
    Computer Graphics Laboratory
    University of Waterloo