Re: Using Mathematica
- To: mathgroup at smc.vnet.net
- Subject: [mg120949] Re: Using Mathematica
- From: Joseph Gwinn <joegwinn at comcast.net>
- Date: Thu, 18 Aug 2011 03:23:48 -0400 (EDT)
- Delivered-to: l-mathgroup@mail-archive0.wolfram.com
- References: <j2g37v$1em$1@smc.vnet.net>
Let me preface my comments by noting that I have been a Mathematica user since version 1.2. It would have been v1.0, but I waited for the initial teething problems to be resolved - never buy version 1.0 of anything. In article <j2g37v$1em$1 at smc.vnet.net>, "David Park" <djmpark at comcast.net> wrote: > There are many postings to MathGroup related to problems of exporting to or > interfacing Mathematica with other applications. Although I recognize that > there may be some cases where this is really useful (acquisition of data or > interfacing with a supercomputer say), in most cases it is nothing more than > exporting to a far inferior medium, or finding ways to organize technical > work, which could be done perfectly well and even better within Mathematica. Umm. A lot of us use Mathematica for things like data reduction, and there is no future to typing 50,000 points in by hand. > Too many users are ready to jump ship before they have even learned to sail > it. The problem with this is that it detracts from learning how to get the > most out of Mathematica. It is such a new and powerful medium that none of > us truly know how to use it to best advantage. How much time did it take us > to learn how to think and write decently in our own language? How much more > difficult is it to think about, develop and express technical ideas? If the alternative is to type 50,000 point in, can you blame them? Said another way, if there is no automated and reliable way to get data in and out of a tool, however wonderful that tool is, it's perfectly rational to find another, better-suited tool. > Mathematica has superb capabilities for computer algebra, graphics, active > calculation, dynamics and textual discussion, presentation and formatting as > coherent and elegant documents. With Workbench we have an excellent way to > generate, preserve, document and organize technical material - say within a > project or for a book. (Most day to day work can be done in Mathematica.) > WRI has reasonable purchase options for anyone interested in math and > science. Everyone should get it. If they don't want to, then it's "the happy > few". There is no reason to go to inferior and out-of-date media. There is > no reason to restrict Mathematica use to being a "super programmable > graphical calculator". Sure there are reasons. Like speed. Mathematica is very general, is interpreted, and does everything in text files. More specialized tools can be orders of magnitude faster, and with large data sets this becomes quite important, to dominant. I've written C-coded preprocessors to do the first data-reduction step, and/or to convert from native format to a Mathematica notebook. This in the days before Mathematica had an Import[] function worth the name. Now that Import[] can handle reasonable spreadsheet and database file formats, one resorts to writing such C-coded utilities only for things that are done many times. For a one-off, one just lets Mathematica grind away. Back in the day of version 2 or so, I was plotting 100,000 data points in a single 2D dot listplot. The data came from custom Ada code that generated the data in a Mathematica-compatible format from the beginning. This was all run on a dedicated Sun workstation. It took Mathematica something like 45 minutes to generate that plot, but Mathematica did it and didn't crash, and that's what counted. At the time, there were few alternatives, and we didn't have so many plots to make that we couldn't afford the ~hour it took per plot. > It's true that there are many ham-fisted features on the fringes of > Mathematica, Grid features that don't work properly, file saving in M8 that > seems to have been bollix up by the advent of CDF (which I consider > questionable and inadequate and now probably unnecessary) and Workbench has > a lot of ragged behavior. They don't seem to have fully embraced it use > outside of WRI documentation. But none of this is fatal and it can be fixed > - with a little pressure from users. Well, this user has another approach: Steadfastly ignore the new stuff until it achieves maturity. Or, more commonly, vanishes. But you don't want people to stop producing the new stuff. While 90% of the new stuff will fail, it is the 10% that succeeded in years past that is our world today. Joe Gwinn