12,516 reputation
53045
bio website
location
age
visits member for 4 years, 3 months
seen 22 hours ago
Very senior embedded real-time systems programmer (Texas law is very picky about the use of the term "engineer"), unsuccessful candidate for U.S. Congress, and unreconstructed Cold Warrior.

Mar
25
answered Does the GPL allow anyone to obtain GPL software without paying for it?
Mar
17
comment What is a scalable and practical way to search existence of a group of strings in a huge file
Option c is the way to go. This is what hashing is for.
Mar
13
answered When not to move a “C style” code to “C++ or Object Oriented” code?
Mar
8
revised Chained ruleset parsing and evaluation code: recommendations?
Add description of Batory LEAPS paper.
Feb
28
answered Chained ruleset parsing and evaluation code: recommendations?
Feb
27
reviewed Close Does your programming knowledge decrease if you don't practice?
Feb
26
comment Unit Testing with an Optimization Problem
The problem you have in something like this is ensuring that the algorithm is in fact doing what you wanted it to do. There were two spectacular failures in neural net applications. One, a neural net to recognize tanks in cluttered photographs, was found to have trained itself on simple lighting differences. Another, a neural net to recognize submarines in cluttered sonar returns, was found to have trained itself on differences between the microphones used to record the test samples. (The sub-present samples used one microphone, the sub-absent ones used a different mic.)
Feb
21
comment Programming conundrum - most efficient path
Does Javascript compile down to actual machine language, or are you getting eaten alive by interpreter overhead?
Feb
21
comment Why are scientific programming languages so weird?
FORTRAN had 1-based indexing. PASCAL allowed arbitrary-based indexing: you could declare an array whose index ranged over, for example, -42 to +57. (See en.wikipedia.org/wiki/… for an example where this is useful.)
Feb
20
reviewed Approve Does it make sense to write tests for legacy code when there is no time for a complete refactoring?
Feb
19
reviewed Leave Open Is there any use for a smart algorithm developer who will delegate the coding?
Feb
18
comment Will we go back(?) to fixed-point arithmetic in the near future?
@supercat, I dealt with FORTRAN on CDC 60-bit machines, DEC PDP-10, SDS 930 and Xerox Sigma 5 (very briefly), IBM 1130 (VERY briefly), Univac 1108, Harris supermini, Perkin-Elmer 3252, and TI 990. I never encountered a FORTRAN with anything like those fixed-point types. IBM might have snuck something into one of their FORTRAN compilers, I don't know.
Feb
17
comment Will we go back(?) to fixed-point arithmetic in the near future?
@supercat, the only languages I ever heard of of with support for that kind of fixed-point were PL/I and COBOL. As far as I know, PL/I never got much traction outside the IBM mainframe world, and COBOL never got traction outside of business data processing, which was (and still is) almost entirely IBM mainframes. Oh, and the correct notation is PICTURE 99V99, if memory serves me.
Feb
17
answered Is it necessary to science of DSP for a c++ programmer?
Feb
16
comment What is the benefit of having the assignment operator return a value?
@delnan, I got my start in this crazy racket in 1970. I've been getting paid to do this, one way or another, since 1974. In all that time, I don't think I've EVER seen code actually PORTED from one language to another. (I've seen LOTS of FORTRAN-to-FORTRAN conversions, moving something from one system to another. I've NEVER seen a FORTRAN-to-PASCAL or FORTRAN-to-C or PASCAL-to-C or BLISS-to-C, or anything else. Normally, the code is rewritten from scratch at that point, with the existing code as a guide. (I've seen ONE (1) MATLAB-to-C conversion, for a radar project. Special case.)
Feb
15
reviewed Approve Are these programming “classics” still relevant today and should I read them?
Feb
14
comment Physics Engine with Fixed-Point Positions
"With the GNU C Compiler, long double is 80-bit extended precision on x86 processors regardless of the physical storage used for the type (which can be either 96 or 128 bits)." Reference to gcc.gnu.org/onlinedocs/gcc/…
Feb
14
comment Physics Engine with Fixed-Point Positions
The moral of the story has something to do with not using Microsoft tools for serious computing.
Feb
14
comment Physics Engine with Fixed-Point Positions
References: [2]msdn.microsoft.com/en-us/library/9cx8xs15.aspx [3]software.intel.com/en-us/articles/…
Feb
14
comment Physics Engine with Fixed-Point Positions
Continued from previous comment: "The Intel C++ compiler on Microsoft Windows supports extended precision, but requires the /Qlong‑double switch for long double to correspond to the hardware's extended precision format.[3]"