|
|
Posted by A Bit Narked on 01/08/08 13:49
On Tue, 08 Jan 2008 08:00:23 -0500,
Jerry Stuckle <jstucklex@attglobal.net> wrote:
>A Bit Narked wrote:
>> On Sun, 6 Jan 2008 23:19:45 +0000,
>> Toby A Inkster <usenet200712@tobyinkster.co.uk> wrote:
>>
>>> A Bit Narked wrote:
>>>
>>>> And yet you can echo true. The output is '1'.
>>> No -- you can't echo a boolean in PHP. Why you try, the boolean is
>>> implicitly cast to a string.
>>>
>>> This:
>>>
>>> echo $bool;
>>>
>>> is actually executed as if you coded this:
>>>
>>> echo (string)$bool;
>>>
>>> If you look up the manual on type casting, you'll find that FALSE is cast
>>> to an empty string, and TRUE is cast to a non-empty string.
>>>
>>> If you cast to an integer instead of a string, then you'll find that FALSE
>>> is cast to 0, which seems to be what you want.
>>>
>>> echo (int)FALSE;
>>
>> Apologies for imprecise expression :-) PHP provides free
>> casting (e.g. to string in echo) so I ignore it.
>>
>> My real point is that there is no logical relationship between a
>> boolean value of TRUE and a string value '1' _except_ where the
>> same relationship yields a string value '0' for boolean FALSE.
>> I.e., the C precedent.
>>
>> I'm not fashed by the idea that there's a special boolean type
>> whose values are TRUE and FALSE. But logically, the string
>> representations of such sui generis values should be 'true' and
>> 'false'. There is no logic whatever in string representions of
>> '1' and ''. No one has articulated a relationship such that
>> '1' and '' makes any sort of sense.
>>
>> One of the basic principles of good design is predictability.
>> 'From one thing know ten thousand things' as the Chinese stated
>> it ages ago. So if there's a well-known precedent, we should
>> follow it unless we have a powerful reason for not doing ('Ensure
>> that everyone after us makes the same mistake we did', as Maurice
>> Wilkes put it, to appreciative laughter). And whatever we do
>> should be internally consistent. So the calling sequence for
>> some new function should be guessable by anyone who knows the
>> sequences of similar functions, one value should tell us about
>> neighboring values, and so forth.
>>
>> The PHP developers' choice for the string reps of true and false
>> violate this principle all to buggery, which is what got up my
>> nostril.
>>
>> I thought everyone here would immediately see how daft it is, but
>> I expect we're circling the drain as a species precisely because
>> so many people -even very smart ones- go through life accepting
>> without examination that 'whatever is, is right'.
>>
>
>Sorry, but it makes perfect sense.
>
>C does not have real boolean values "true" or "false". Rather, it
>gimmicks them up by saying anything with a non-zero value is true, and
>anything with a zero value is false. One of the problems is that both 1
>and 2 are considered true - but in this case true != true (1 != 2).
>
>PHP does have a real boolean "type" with the values "true" and "false".
> So true == true all the time.
>
>PHP can also do conversions. Like with C, anything with a non-zero
>value is true, and anything with a zero value is false. This includes
>an empty string.
>
>And if we kept doing things the same old way, we'd still be rubbing
>sticks together to keep the cave warm. Progress always means change.
Progress is a type of change, but change is not a type of
progess.
This seems to be a very easy error to make. You've made it 3
times in your reply!
Navigation:
[Reply to this message]
|