@ChrisJesterYoung They appeared annoyed at my opinion about the task, though. Not that I care too much, honestly. The same person asked me to delete a few comments on his answer that would have helped them golf down his answer by a hundred characters more. Instead they chose to ignore them ;-)
@RebeccaChernoff I know only two people who smile while standing on their head. One is Rebecca Chernoff, the other is my gf. Since the latter is currently at work without internet access only one option remains :P
The only thing I see is that non all languages have an obvious interpretation of "from the command line arguments" (i.e. those people trying to compete using javascript in the browser). For them I often use some language like "or other mechanism natural to you environment".
Or you can just skip it. It won't hurt them to learn something new anyway.
Joey has been pushing for question authors to supply automated grading as often as possible, which would mean a script to check that all the words (1) exist in the dictionary and (2) differ from the previous one in only one letter.
Ah, fair point. To me the point is that you should be able to specify two words, or pull at random from a list. I can just leave it as simple as that I suppose.
I thought about an automated grading, but I don't have a solution, so I can't do a bunch of tests and guarantee the number of interim steps I have is the best/minimum. |: Doing just the 2 steps you mention could work.
Question: Usually the puzzle connects two words of the same length, but I have seen variation that allowed adding or subtracting letter in various ways. Perhaps you should specify which you mean.
No idea how to solve it, currently, though. But that's fine :)
As for a test script, I usually appreciate something that works on Windows, too :-), even though the vast majority of the site seems to be on Unix-likes
@Joey Don't worry, my issue will be getting a non-windows test script...I've been all Windows for the last year or so...so aside from being able to test a bash script, no doubt I'm rusty. (;
Since I don't have my own solution, short of scouring the internet for examples and hoping they're using the least number of steps...I think I'll skip that and do what @dmckee suggested, just something that checks each word is valid and only a 1 letter difference.
Though, I don't have that written, so I either post now without it and update, or delay posting.
Well, yours could probably work fine without one, I think
@dmckee: »Oh, come on. Code challenge to write a loop or comprehend a list. Whoop dee doo! – dmckee♦ 1 hour ago« – Thanks, now I don't feel bad for downvoting anymore ;-)
@Joey That'd be why I wrote the scorer for IPD in python despite being a rank beginner, who more or less writes c in any language.
Well, I can also do fortran in anything, too.
@Joey There is a sense in which I feel bad about that comment, but if there is never any negative feedback we will continue to see one trivial "challenge" after another.
The greater problem is the interface, though. Even on unix-likes spawning 200 processes per run has some impact, especially if the strategy was written in Python
@dmckee I wouldn't have expected it either. But many cross-platform scripting languages are full of such things when OSes differ. Based on my own experience Python's popen3 or so didn't work on Windows either.
Sure, process startup is worse on Windows, but the python strategies against each other took in the order of 8 times the time of two C strategies against each other :)
Admittedly, I felt bad for posting my test case generator for Nested Boxes which was written in C#, requiring Linux users to use Mono (on which it runs great, though)
If you would just mandate that the warriors directory contains no non-executable files the additional test could be dropped. I think I got it working with that.
But some strategies use header files I don't have here so I migrated over to a unixoid, too
Random question: How many trivial tasks do we tolerate, actually? As for me, I find them nice form time to time as I don't always have a few hours of time to spend for golfing, but having the site swamped with them is probably not too good as well. But I wouldn't want every task to be on the complexity level of Robot finds Kitten, to be honest.
To the matter of trivial challanges, I thinking that Airplane Navigation at least requires some "which quadrant?" logic, so I haven't voted down.
I don't think we close the trivial ones except in the worst cases, but I'm willing to put my rep where my mouth is and vote them down when they annoy me.
I agree that no one whats to write a dissertation for every challenge.