RandomNum - 2.1generate random numbers from uniform & normal distributions |
|
||||||||||||||
|
|||||||||||||||
Feedback Summary:
| This Version: | |||||
| Overall Rating: | Features: | Support: | |||
| Ease of Use: | Quality / Stability: | Price: | |||
Key to Types of Feedback:
Reviews
Troubleshooting
Usage Tips
Developer Notes
Commentary
Featured Reviews
Simple, easy to use, with a nice dash of extra functionality 



- Version: 2.1, 6/7/2007 08:26AM PST
(1 of 1 users found this comment useful)
pohld
RandomNum's handy extras include being able to generate a list of random numbers rather than just one at a time, but this list is unnecessarily limited to a maximum of 100 entries.
The user can set both a minimum and maximum value for the range from which random numbers are generated, but the results are limited to whole numbers, so if for some reason you want non-whole random numbers, you have to divide the results by the appropriate multiple of 10, which could be a nuisance if generating lists of more than just a few members.
One can also set an option for "Gaussian" results, which generates random numbers between 2.5 and -2.5 with six decimal places of precision (i.e. six numerals to the right of the decimal point) in a non-uniform distribution (i.e. in some sort of bell curve with most values near zero and progressively fewer toward the positive and negative extremes), serving some undoubtedly useful statistical purpose.
There is no documentation or help text, which along with the limitations mentioned above is why I'm not giving RandomNum an overall five-star rating. While it is true that no documentation is really needed to use the program (assuming anyone using the Gaussian option would know what that's all about without further elucidation), the topic of random number generation is not a simple one and it would be a nice extra touch to include some brief discussion to educate curious users, particularly regarding the fact that true random number generation is not actually possible through a software-only application like this, but requires a hardware component designed to provide truly random seeding.
The Wikipedia entry for "Random Number Generator" provides further details for those interested, and tells us that software-generated pseudorandomness is close enough for all but the most demanding technical purposes (e.g. cryptology). However, Wikipedia also says that some software uses quite poor pseudorandomization algorithms, so again it would be both reassuring and educational if RandomNum included some documentation describing its own pseudorandom-generation methodology and relative validity.