Re: my wish list for Mathematica next major version
- To: mathgroup at smc.vnet.net
- Subject: [mg60144] Re: my wish list for Mathematica next major version
- From: carlos at colorado.edu
- Date: Sun, 4 Sep 2005 03:02:09 -0400 (EDT)
- References: <email@example.com>
- Sender: owner-wri-mathgroup at wolfram.com
Learning a language in depth in college indeed shows good intentions. But is it feasible? When I did my graduate work at Berkeley (64-66) there were only two languages: Fortran IV and IBM assembler. (Well, COBOL was out but that was for a different school.) Engineering students learned FIV in one-semester courses taught mostly by EE instructors (CS didnt exist) which went fairly deep; e.g., 2-level storage management. One interesting aside is that many of the computer-savvy faculty at the time were experts in assembler and could read octal dumps - they had started before Fortran I came out in 58. That frame of mind can be seen in Don Knuth's early books. Contrast the situation today. By edict from curriculum committees, which indirectly receive industry inputs, an engineering undergraduate must be "reasonably proficient" in a panoply of languages that include C, Excel, a matrix language, a webmaker, and a WP. Languages from a second tier (C++, Perl, Java, Mathematica, microassembler, AutoCAD, SolidWorks, Labview, etc) is taken as per major, or elective choices. (Fortran disappeared in '97.) Can they go deep in each language? For the majority, no way. No background, no time, no instructors. Few of the instructors are able to teach beyond the barest of basics. As for background & time, 30-40% of undergraduate education is remedial. Kids coming from the public HS system can barely write a complete sentence or divide two numbers by hand, let alone understand what a "dispatch table" is. And the graduate level is a similar story.