Generating Random Letter(s)?

FoxhoundX

Is there a way to generate a random letter easily, without using something like

<?
$letter = rand(1,26);
if ($letter == '1'){
$letter = 'a';
}
if ($letter == '2'){
$letter = 'b';
}
?>

ereptur

Sure there is:

$alpha = array("a","b","c","d","e"...."z");
$letter = rand(1,26);
echo $alpha[$letter];

laserlight

That has to be modified slightly, array indices are numbered starting from 0, not 1.

Incidentally, this might be an interesting alternative:

<?php
mt_srand((double)microtime() * 1000000);
function randomSelect($input = "abcdefghijklmnopqrstuvwxyz") {
	$num = mt_rand(0, strlen($input) - 1);
	return $input{$num};
}

$char1 = randomSelect();
$char2 = randomSelect("random");
?>

The idea would be to use a string as an array.
Seed the pseudo-random number generator outside of the function.
Then use an input string to provide the possible characters to be returned.
To simplify in case there is no need for different input strings, we just have a default input string.

Pig

I find it easier to create a random number, then md5 it and chop it to the length you want. You can even choose a random length and starting point.

Pig

sorry, misread. thought you meant a random string. bad pig.

laserlight

even if we were generating a random string, your method would have only 16 possible characters used.

imnu

Looks like this is my day. I was going to post a question regarding generating random letters. I will try laserlight's method. Thans guys.

imnu

Lasserlight, how do I get more than one letter using your method?

FoxhoundX

Here is what I did, Using ereptur's method-

$FF = "0";
$regcode = "";
while($FF != 9) {
$alpha = array("a","b","c","d","e","f","g","h","i","j","k","l","m","n","o","p","q","r","s","t","u","v","w","x","y","z");
$letter = rand(0,25);
$regcode .= $alpha[$letter];
$FF++;
}

Replace the while line with however many letters you want 🙂

laserlight

If you're planning to repeat the use the function, then it should be modified.

My solution would then be:

<?php
mt_srand((double)microtime() * 1000000);
function randomSelect() {
	$input = "abcdefghijklmnopqrstuvwxyz";
    $num = mt_rand(0, 25);
    return $input{$num};
}

$randstr = "";
for ($i = 0; $i < 10; $i++) {
	$randstr .= randomSelect();
}
echo $randstr;
?>

jimson

i am thinking about...
how do we know that a random selected number is a good random?

is a good random means

for($i=0; $i<26; $i++) {
echo myrandomfunction();
}
will shows up 26 of a-z in different order

less than 26 characters where some of them is show up more than one

how to we define random... (a non-patterned?)
is not pattern a pattern ?

laserlight

a random number is a random number

however, it is possible to test if a particular pseudo-random number generator is good using statistical tests

jimson

it is possible to test if a particular pseudo-random number generator is good using statistical tests

so what it means... in general...

laserlight

This page has a program that does some testing:
http://www.fourmilab.ch/random/

jimson

yup, i know the test as i already studied them before.
anyway, how to u define random? (exclude the statistical tests)

imnu

Thanks laserlight. I tried your code. It works. I added 0-9 numbers to $input and changed the limit from 25 to 35, and I got mixed letters and numbers string. Simple and effective! Thanks a lot.

laserlight

how to u define random

I define a random event as an event whose outcome is determined by "chance", such that the outcome of any given event is unpredictable (within the sample space).

Since pseudo-random number algorithms must be deterministic, it is possible, given the algorithm and sufficient data, to predict future outcomes of a particular sequence.

As such, pseudo-random numbers are not random numbers.

Ravenous

After some carefully study and discussion with one of my smarter math instructors, I have come to the conclusion that nothing is truly random. There is an algorithm for everything.. and I'm talking everything in the Universe. As mere humans, we lack the understanding on most all subjects to notice the algorithms, but I contend that they do, in fact, exist.

So nuts to ever generating a truly random number.

jimson

well, i agree with u Ravenous,

if random numbers can be statistically test to check their relationship then it is already not a random thing coz there is already somekinda relationship (pattern) there.

a no pattern random numbers is already a pattern.

that is what i want to highlight.

laserlight

nothing is truly random. There is an algorithm for everything

actually, IMO we dont really know that.
it was proposed, during and before debates on quantum theory in the early half of the last century that every event might be predictable, i.e. there is an (as yet) undiscovered algorithm that can predict every and any event completely, or set of algorithms that have such a property.

however, physicists (many of whom are indeed brilliant mathematicians who I presume are as good if not better mathematicians than your "smarter math instructor") generally still believe that quantum events are truly random.
the actual reasoning behind such a conclusion probably lies in things like Heisenberg's uncertainty principle, Goedel's incompleteness theorem, Bell's theorem etc, but I'm afraid I'm not qualified to discuss the details.

If you want me to speculate as a layman, I would say that by Heisenberg's uncertainty principle, one cannot know the exact position and momentum of an electron at any given time, thus one cannot determine its exact future. As such, it is possible to conclude that there exist events such that future events pertaining to that event, or sequence of events, cannot be precisely determined from that event or sequence of events.
If that is so, then that future event is indeed unpredictable, hence randomness exists, hence random numbers exist.

if random numbers can be statistically test to check their relationship then it is already not a random thing coz there is already somekinda relationship (pattern) there

there's no point in statistically testing random numbers.
the statistical tests I refer to pertain to pseudo-random numbers.