re: Questions re: Optica
- To: mathgroup at smc.vnet.net
- Subject: [mg5089] re: [mg5006] Questions re: Optica
- From: D.Barnhart at lboro.ac.uk (Don Barnhart)
- Date: Wed, 30 Oct 1996 22:03:37 -0500
- Sender: owner-wri-mathgroup at wolfram.com
Hello, this is Don Barnhart, Optica's developer. Yes, it is true that Optica is probably one of the slowest optical packages on the market. This is partially due to the fact that it is written entirely in the Mathematica language. Of course the processor speed will effect the ray-tracing time. I recommend using either Pentium or PowerPC based machines. However, as a partial trade-off for its lack of speed, Optica does offer a flexability to the user which is entirely unknown to other packages. By being part of the Mathematica environment, Optica allows you to quickly build up new analysis tools from scratch in very short period of time. You are not limited by what the ray-tracing package has chosen to include as built-in functions. In addition, Optica's excellent graphical representations of optical systems is probably unsurpassed by any other package on the market. The upcoming versions of Optica will indeed run much faster than the Optica's present version. I have already created a new ray-tracing engine that runs nearly 10 times faster than the Optica's present version for seqential ray-tracing operations. I am also working on a C-code kernal for Optica that should run 1000 times faster. Ultimately, its not the ray-tracing time that counts, but the overall development time for a new project. I originally wrote Optica to help with the design of a holographic stereo camera system in which I needed to determine the dimensions and placement of a set of lenses and prisms in three-dimensions. Although the ray-tracing speed of Optica was slow, the project was completed in a fraction of the time it would have required without the help of Optica. Because Optica is extremely flexable and makes no approximations or assumptions (it ray-traces non-paraxial solutions in three-dimensions for nonsequential surfaces), I find that Optica also helps to give me an understanding of the optical system rather than simply an answer to a particular optimization question. Regarding the RAM requirements of Optica: Optica and in general Mathematica does use up a great deal of RAM. In general, I prefer to work with 32 Megabytes built-into my machine. While this may sound like alot of RAM, this amount is not particularly expensive to come by these days. Regarding the Optica User's Guide: I do appologize if the writing is terse. It is difficult when the author of a software package also attempts to explain his own package to a beginning audience. I am employing a professional technical writer to produce the next version's User's Guide. Regarding CircleOfRays as well as any of the other ray source functions, you can use any of the 30+ Ray options within CircleOfRays in addition to the particular options specific to CircleOfRays. I guess that I had assumed that the user could easily type into Mathematica: Options[CircleOfRays] and Options[Ray] to learn what is possible. In general, many of Optica's functions have overlapping options and I did not think that the extra printed book space was justified. Perhaps I will change this in the next release of the User's Guide. In general, if you have any questions or comments regarding the use of Optica, you should send your questions to optica-support at wolfram.com. That's what I am here for. Best regards, Donald >From: Fisherman <halasz at sprynet.com> To: mathgroup at smc.vnet.net >To: mathgroup at smc.vnet.net >Subject: [mg5006] Questions re: Optica >Organization: Sprynet News Service >X-Mozilla-Status: 0001 > >Anyone running Optica out there? > >Do I have things screwed up; or is this the slowest optical package on >the market? > >For a 100 ray trace function, the kernel keeps asking for more RAM (I've >got it up to 12MB), and after fifteen minutes or so, I lose my patience. > >Oh, it's using one of the package demos, a cylindrical light source >reflecting into a parabolic mirror. > >While I'm on my rant, who or what writes the booklets? For function >calls, say CircleOfRays, they list the parameters in brackets as >'options', even in the glossary of terms. Couldn't they list what those >options _are_, or should I put in something like >CircleOfRays[air_conditioning, corinthian_leather]? > >Any help is much appreciated (can't get to a phone for the next two days >during business hours). > >TIA; > >Fisherman. ****************** Donald Barnhart Department of Mechanical Engineering Loughborough University Loughborough, Leicestershire LE11 3TU, UK Tel:(011 44 1509) 223177 Fax:(011 44 1509) 223934 ****************** "Life is about being complicated as simply as possible"