Re: Breaking backwards compatibility - good or bad?

Wayne wrote: 
On Sun, 1 Jan 2006 10:49:17 -0000, "Tony Marston"
<tony@xxxxxxxxxxxxxxxxxx> wrote:


Not in C, C++, Java, etc. they aren't. And we're talking COMPUTER languages here - not HUMAN languages. Or can't you figure out the difference?

But the point of this argument is that variable names such as 'box', 'Box' and 'BOX' should all indicate the same variable, not different variables. If you really want to have three different boxes you would use names such as 'box1', 'box2' and 'box3' to avoid any confusion. 


I'd like to point again (although not necessarily for you Tony, more
for Colin) that Human programmers, in case-sensitive languages, do use
case to distinquish meaning.

In Java, classes begin with an uppercase letter and
variables/functions do not.  And "box" and "Box" are commonly used
together.  In C, all capitals signals constants or types and BOX and
box are commonly used together.

Humans do these sorts of things.  I'm not saying it's good or bad --
I'm just saying that's how it's done.

I don't deny it. I'm not saying that humans don't or can't do these things.
Case distinction is but one of the many distinctions and pernickity rules that we have to learn when we learn to program in a language or system that uses it. We can of course learn it, just as we can learn that letter 'o' must be distinguished from zero. And because it's there, some people make use of it.


But, to reiterate and expand my previous argument, the problem comes when we make up the letters into words - which, most of the time we do, when we are naming things - variables, classes, files.
Those of us who grew up with English (and other European languages which have case in their alphabets) are used to thinking of the word 'box' as being the same word whether written 'BOX' 'box', 'Box' or even 'bOX'. It's not that we can't distinguish these, but we don't normally need to, and more significantly, we don't have very convenient ways of expressing the distinction.
This means first that (as I said before) if we need to dictate a line of code over the phone, we need in principle to express the case of every letter. There are ways of shortening this of course, and if there is a relevant convention (such initial cap, or camel case) we may be able to express it relatively easily, provided the hearer knows the convention.

But actually I think there is a more severe issue, to do with remembering and comparing names. The variable I was just using in that other routine - was it called 'textbox', 'TextBox' or 'textBox'? For me anyway, remembering a name made up of ordinary words or familiar abbreviations for them is trivial (I probably vocalise it). But remembering the case, if it is not all a single case, is not trivial - it takes more effort and is prone to error. I can't believe that I am the only one to have this experience.
It is true, of course, that conventions for use of case assist in this. If I know that class names will always have an initial capital, I don't have to remember this for each name. But I find that in practice there are lots of different conventions and I'm often switching from one to another as I work on different scripts written by different people at different times, even on the same system. A particular bugbear (because particularly difficult to remember) is when using CamelCase, are little words in the middle of names capitalised or not? (i.e. is it WriteToFile or WritetoFile?) Again you can, and probably should, have a convention, but what if there are different conventions?


Somebody will probably come back and say, well you should have conventions on this and then you won't have a problem. And I say, yes, but if the language weren't case sensitive in the first place, there wouldn't be a problem.

Despite Jerry's insistence, I have yet to see an argument in favour of case sensitivity that doesn't come down to either 'everybody uses it' or 'for consistency with XXX system that uses it'. I recognise that as a practical fact it is here to stay, and yes, I use it. But that doesn't stop me from arguing its demerits, and I resent being patronised and insulted for doing so (by Jerry, rather than by you Wayne).

Colin
.

Relevant Pages

  • Re: Who uses clapack?
    ... > Interoperability between languages is a gray area where one ... > 2 Knowledge of linker conventions such as upper case conversion ... > 3 Ability for the programmer to specify parameter passing by ... > 5 Conventions on stack based parameter passing notably padding ...
    (comp.lang.fortran)
  • Re: Who uses clapack?
    ... > Interoperability between languages is a gray area where one ... > 2 Knowledge of linker conventions such as upper case conversion ... > 3 Ability for the programmer to specify parameter passing by ... > 5 Conventions on stack based parameter passing notably padding ...
    (sci.math.num-analysis)
  • Re: Case sensitivity in programming languages.
    ... These conventions are useful for languages that have no ... disputing their usefulness. ... universally followed by all professional programmers. ... Those conventions are only used in those languages whch do not have ...
    (comp.lang.php)
  • Re: Case sensitivity in programming languages.
    ... These conventions are useful for languages that have no ... universally followed by all professional programmers. ... Case sensitivity was introduced as a programmer ...
    (comp.lang.php)
  • Re: schoen
    ... but I have been limiting the Central/Middle distinction ... I've been talking about, e.g., MIDDLE English, which refers only to ... I'm talking about FUCKIN' *LANGUAGES AND DIALECTS*, ... Swedish" refers to a stage of Swedish between Old Swedish and Modern ...
    (sci.lang)