2012/8/25 Michael Ossipoff <[email protected]> > On Sat, Aug 25, 2012 at 5:09 PM, Jameson Quinn <[email protected]> > wrote: > > > > > > 2012/8/25 Michael Ossipoff <[email protected]> > >> > >> ... > >> > >> Are you desperately reaching, in an effort to save unimproved > >> Condorcet from a comparison to ICT and Symmetrical ICT? > > > > > > I think you're being hypersensitive here. > > Not hypersensitive--just intentionally (but good-naturedly) > aggressive. Trying to provoke some comparison between ICT, Symmetrical > ICT, and unimproved Condorcet. > > > The idea is simply to have yet one > > more check that your pseudocode is valid. > > I agree that certainly the most convincing test for code being > error-free would be to actually run it. I would, if I knew how to run > the programming language that my computers have. > > But, as I said, I've thoroughly checked the program for errors, and I > assure you that I'm confident that I've found and fixed all the > errors. > > I didn't know that those programming languages could be run online at > a website, without installing them. Of course I agree that the best > way to make a program available is to write it in a popular > programming language. Of course, on the other hand, converting a > program from pseudocode to a programming language is a very minor > task, very little work. To a person who is familiar with the language > that they're writing it in, that task amounts to nothing more than > copying. > > > Then testing as suggested would be easy. > > > > I think that python code, in particular, is even clearer for a > > non-programmer to read than your pseudocode. > > Maybe, but I don't remember if Python allows "endif", "endwhile", > etc., even though it doesn't require them.
You can always include them as comments. > I feel that, for clarity, > those statements should be included. For clarity, I like the BASIC > style FOR/NEXT loops, with "next i" at the bottom. > > And it seems to me that (in the Python version that I read about, some > time ago) Python's multi-dimensional arrays were a bit awkward, in > comparison to other languages. Not if you set them up using list comprehensions (which may not have been in the book you read, if it was old). > ...Especially since the Python book > that I had didn't specify them well. Computer language books often > don't specify a language well, especially if the author is trying to > be funny or clever. I've long felt that computer language books should > be written only by the person who devised the language, or by a > mathematician. > > >> > >> Computer-counted test scenarios can only confirm the > >> already-determined facts that I've stated. > > > > > > Yes. That's exactly what they're good for. It may have no value for you, > > because you've checked the algorithm carefully by hand, but for us a > > computerized test would save us that work. > > True. I can't expect others to take my word for it that there are no > errors in the code. Actually running the program would be more > convincing. > > But of course the criterion compliances of Symmetrical ICT are things > that can be demonstrated without using a computer program. > > Well, I might soon be saying that I've run the program in VBA, the > most widely-distributed programming language. (It's on every computer > that has Microsoft Office). > But javascript (and thus the ability to run coffeescript, which is essentially just a dialect of javascript without so many nasty gotchas) is in every computer, tablet, and phone with a modern browser... Jameson
---- Election-Methods mailing list - see http://electorama.com/em for list info
