From: <HIDDEN>
Newsgroups: comp.soft-sys.matlab
Subject: Re: a huge loss for the file exchange
Date: Tue, 5 Jul 2011 23:02:09 +0000 (UTC)
Organization: Cambridge University
Lines: 24
Message-ID: <iv055h$o4s$>
References: <itnd74$g8s$> <ito8j8$h2t$> <itodpk$2th$> <itpshq$o12$> <itrot4$igd$> <itv9mh$jb6$>
Reply-To: <HIDDEN>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
X-Trace: 1309906929 24732 (5 Jul 2011 23:02:09 GMT)
NNTP-Posting-Date: Tue, 5 Jul 2011 23:02:09 +0000 (UTC)
X-Newsreader: MATLAB Central Newsreader 870214
Xref: comp.soft-sys.matlab:735006


I'd like to echo everybody else's thanks for the work you've done on the FEX and CSSM.

For me, the most valuable aspect of your work has not been the code/syntax; it has been the learning. Five years ago, I was still learning MATLAB and knew little about computational geometry. I feel strongly that your submissions (and responses when I've been stuck on the file exchange) have taught me a great deal.

I've now mastered computational geometry in <=3D and can think about problems in higher dimensions. I've constructed robust interpolants where splines were oscillating, solved nonlinear optimisation problems and fitted hypersurfaces with polynomials. Much/all of this has been achieved with your help.

I'll try and help out on CSSM a bit more than I do now - mostly looking out for the computational geometry subjects - and hopefully propagate some of that knowledge, although clearly I'm not of the same calibre!

Three of 'my' favourite achievements for which a good deal of credit goes to you:

Fixed an oscillating spline surface in 2D, which had resulted in a ship propeller blade looking like corrugated iron and vibrating like crazy as it rotated in the testing tank

I've put together multiple toolboxes (for work and PhD). I'm proud that they're easy to use, with intuitive and descriptive function interfaces which my colleagues really appreciate. Almost every one of the higher level functions contains a call to parse_pv_pairs.

Take 4 cameras, resolution 1MP, operating at 1000Hz in a vibrating environment, recording images of particles 10um in diameter. Reconstructing 3D volumetric images from the 4 cameras is impossible unless an incredibly accurate inter-camera calibration map is defined. Polyfitn helped achieve this (helped in a big way), along with several posts which described the use and theory of polyfitn.