GMSH as a front end for our EM solver engines...

[Eric W. Lucas]

	Dear Dr. Geuzaine,

	Here is an e-mail that I sent to your partner Jean-Francois. I don't know
if the posted e-mail address for him is valid or not. I thought you should
read it as well and feel free to forward it if you see fit. Is this
something that you guys have any interest in at all?

	Regards, Eric Lucas

----------------------------------------------------------------------------
----------------------------------------------------------------------------
-----------------------------------------------
	Dear Sir,

	My name is Eric Lucas and I work for a small start-up company called Alpha
Omega Electromagnetics (' AOE '). I have sent so many e-mails over the past
few days,  I may have already sent you an e-mail in which case please
forgive my repetition. I'll try to keep this as short as possible. My
partner and I have a very powerful 3D H1(Curl) TVFEM electromagnetic high
frequency simulation code. I have developed it over the past 10 years while
working at Westinghouse (now Northrop Grumman) and have published the basic
technique (IEEE Trans. Antennas and Propagat. Feb 1995 pp.145-153) and
presented many papers on its further development over the years (IEEE APS
symposia). I have used it over the past several years for the industrial
design of advanced technology phased-array antennas with really good
success. If you are familiar with the commercially available high frequency
EM simulation software packages 'HFSS' from Hewlett Packard and more
recently Ansoft Corp.,  then you will have a good idea of our capability as
well because they are very similar (S-parameter computations for 3D volumes
with 2D modal 'port' excitations). The one discriminating feature of our
solver over the years has been its 'generalized periodic boundary
condition' whereby we are able to simulate infinite periodic phased-array
antennas and surfaces in addition to the class of 'closed' EM problems. We
have successfully had this feature operational since 1989, and only very
very recently has Ansoft offered a similar (but not quite sufficient)
version of this capability. Our adaptive mesh refinement scheme and
iterative solver technology is state-of-the-art. A full complex*16
non-symmetric solution (@1 freq.)  for a matrix system of N=125,000 takes
about 5 minutes on my NT PC. The symmetric solution time is obviously even
less. I wrote all this only to establish some credibility, again please
forgive the long-windedness.
	Over the years while at Northrop we used SDRC IDEAS as our CAD/Initial
Mesh front end. Now we are seeking a more 'EM' compatible front-end. I saw
your 'GMSH' on the web and it seems like you guys have EM backgrounds as
well, is that correct? Anyway, I have the following two tasks to
accomplish:

	1. Obtain some useable front-end for our own in-house design consulting
operation. (Need this in a hurry...by far first priority)

	2. Customize a simple front-end in conjunction with our solver engines to
offer for commercial sale targeting NT-based PC                 platforms.
(This would be a truly competitive 3D capability which could be priced far
below the industry standard HFSS                 products)

What you guys have done, or are attempting to do looks allot like the piece
that we are missing. I wonder if we could discuss this a little just to
investigate the plausibility of some kind of collaboration? In any case,
could we at least talk so that I can determine whether or not GMSH would be
useable for our in-house needs?

	Thank you for your consideration,
	Eric W, Lucas
	410-750-0190
	8787 Manahan Dr. Ellicott City, Md. 21043 USA
	alphaomega at maranatha.net