|
|
Posted by Tony Marston on 11/23/41 11:55
"Jerry Stuckle" <jstucklex@attglobal.net> wrote in message
news:O4udnUk-gOCom0zZnZ2dnUVZ_oWdnZ2d@comcast.com...
> Tony Marston wrote:
>> "Shelly" <sheldonlg.news@asap-consult.com> wrote in message
>> news:7BGzg.6578$gF6.716@newsread2.news.pas.earthlink.net...
>>
>>>"Tony Marston" <tony@NOSPAM.demon.co.uk> wrote in message
>>>news:ean53d$rlk$1$8300dec7@news.demon.co.uk...
>>>
>>>>Case sensitivity can lead to unmaintainable code. For example, if there
>>>>is a variable called $foo I expect $Foo and $FOO to mean the same thing,
>>>>not different things. The fact that they are not can lead to unexpected
>>>>results.
>>>
>>>You just said it here right now. It can lead to unexpected results
>>>....for YOU....because ...YOU...expect them to be the same when, in fact,
>>>they are different thingsp
>>
>>
>> And so does everybody else with a background in case insensitive
>> languages.
>>
>
> You cannot speak for "everybody else with a background in case insensitive
> languages" - and you certainly do not speak for me. In fact, you don't
> speak for most programmers I know with that background.
>
> And yes, I did start with case insensitive - FORTRAN II back in 1967, and
> several other languages since then, including some you've probably never
> heard of.
>
>>
>>>>Programmers who deliberately create different variables with the same
>>>>name but different case are bad programmers, but a *proper* language
>>>>(such as COBOL) removes the possibility for such bad programming by
>>>>ignoring case and treating all the variables as a single variable. Any
>>>>language which
>>>
>>>IMO COBOL is the worst language I have ever seen. I have not programmed
>>>in that language for almost thirty years. At that time it treated all
>>>variables as globals -- talk about you UNMAINTAINABLE code!!!!
>>
>>
>> It is possible to write mantainable code in COBOL just as it is possibe
>> to write unmaintable code in Java. It is the programer and not the
>> language which is the deciding factor.
>>
>>
>>> (I think they changed that somehow, but am not sure). If you happened
>>> to use the same name in two different (were the "subroutines" called
>>> paragraphs or procedures? -- I forget), then a change in one changed the
>>> value in the other -- and that was not convention; it was the language
>>> itself. Ugh and double-ugh!!!
>>
>>
>> That is not the case as I remember it. Each subprogram has its own
>> working storage section, so anything declared in one subprogram cannot
>> possibly affect anything in another subprogram. It is only when you get
>> to shared storage areas, such as common storage or the linkage section,
>> that you may have problems. Yet agin, as COBOL is a compiled language,
>> anything which is actually declared more than once can be flagged as an
>> error at compile time, so the problem is easily spotted and fixed.
>>
>>
>>>>deliberately allows programmers to wrte unmaintainable code is a bad
>>>
>>>Oh, I'll totally agree with that statement. That is why COBOL (at least
>>>circa 1975) royally sucked wind.
>>
>>
>> I agree. COBOL 85 was much, much better.
>>
>>
>>>>language. After all, that is why most modern languages do not include
>>>>GOTO because f the problems it can cause.
>>>
>>>Uh, in fact that is not correct. Java, C, C++ and any other I can think
>>>of all allow the goto. It is just that good programmers don't use it.
>>>What do we call that now? I think the word "convention" comes to mind.
>>
>>
>> Programmer conventions are still not the same as language rules, and I
>> dislike the idea of certain conventions in some languages being promoted
>> to rules in other languages.
>>
>
> No, but programmer conventions are built around language rules, and take
> advantage of those rules.
I disagree. The convention of using different case to differentiate between
constants, variables, functions and methods grew about because of a
deficiency in the language. There was no rule in the language which made
such a convention obligatory.
--
Tony Marston
http://www.tonymarston.net
http://www.radicore.org
Navigation:
[Reply to this message]
|