From mboxrd@z Thu Jan  1 00:00:00 1970
Received: (from weis@localhost) by pauillac.inria.fr (8.7.6/8.7.3) id TAA03696 for caml-red; Thu, 27 Jul 2000 19:36:40 +0200 (MET DST)
Received: from concorde.inria.fr (concorde.inria.fr [192.93.2.39]) by pauillac.inria.fr (8.7.6/8.7.3) with ESMTP id MAA05241 for <caml-list@pauillac.inria.fr>; Wed, 26 Jul 2000 12:35:01 +0200 (MET DST)
Received: from cepheus.azstarnet.com (cepheus.azstarnet.com [169.197.56.195])
	by concorde.inria.fr (8.10.0/8.10.0) with ESMTP id e6QAZ0D09578
	for <caml-list@inria.fr>; Wed, 26 Jul 2000 12:35:00 +0200 (MET DST)
Received: from dylan (dialup03ip036.tus.azstarnet.com [169.197.31.36])
	by cepheus.azstarnet.com (8.9.3/8.9.3) with SMTP id DAA06172
	for <caml-list@inria.fr>; Wed, 26 Jul 2000 03:34:57 -0700 (MST)
X-Sent-via: StarNet http://www.azstarnet.com/
Message-ID: <001701bff6ed$6a9f73a0$210148bf@dylan>
From: "David McClain" <dmcclain@azstarnet.com>
To: <caml-list@inria.fr>
Subject: New Major OCaml App
Date: Wed, 26 Jul 2000 03:36:52 -0700
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.3110.5
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3
Sender: weis@pauillac.inria.fr

I have just completed another major mathematical analysis program written as
a hybrid OCaml and NML/IPLib program. The OCaml is essential for robustness,
rapid development, near C speeds, and for prefiltering data for the
inherently loosely typed and unsafe NML/IPLib math core.

The program solves for optical aberrations in a sensor based on a random
collection of observed point targets at various sub-pixel phasings. This is
a very difficult, high dimensional, very sparse, optmization problem based
on a Levenberg-Marquardt optimization algorithm. Each iteration of the
program requires typically 900-1000 two dimensional FFT's at anywhere from
128x128 to 512x512 computation cells. It replaces a previous program
developed by another on our team that takes about 40 hours to compute the
solution with one that computes on your average desktop Pentium in about 5
minutes.

You can download a postscript document that describes the math and the
program at www.azstarnet.com/~dmcclain/blurfit3_ps.zip

Thanks again to all the wonderful team at INRIA for such a fabulous
language!

- DM