NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/lang/c++-part1/1501

Text File | 1996-08-06 | 4.5 KB | 108 lines

Path: beach.and.nl!usenet
From: [e-mail] (Jos A. Horsmeier)
Newsgroups: comp.lang.ada,comp.lang.c++,comp.lang.c,comp.lang.modula3,comp.lang.modula2,comp.lang.eiffel
Subject: Re: Hungarian notation
Date: 11 Jan 1996 10:26:08 GMT
Organization: AND Operations Research B.V.
Message-ID: <4d2ok0$[e-mail]>
References: <[e-mail]> <[e-mail]> <4aleun$[e-mail]> <[e-mail]> <4asnkr$[e-mail]> <4ath75$[e-mail]> <4b4kij$[e-mail]> <dewar.819489496@schonberg> <4bd <4cf8hf$[e-mail]> <4cgq30$[e-mail]> <4cvu68$[e-mail]> <4d21og$[e-mail]>
NNTP-Posting-Host: klepzeiker.and.nl
Mime-Version: 1.0
Content-Type: Text/Plain; charset=ISO-8859-1
X-Newsreader: WinVN 0.99.5

In article <4d21og$[e-mail]>, [e-mail] wrote:
|In <4cvu68$[e-mail]>, [e-mail] (Warren Young)
|writes:

|>This is one of the reasons HN is useful.  By forcing the programmer to
|>visit every place that the variable is used, the programmer has an
|>opportunity to check the semantics of the variable's use.  If the type
|>of the variable has changed, chances are that the semantics have
|>changed as well.  How many times have you changed a variable's type to
|fix something, only to break something else?

|I find this is false in practice. In general, when I change the declared
|type of a variable, I'm doing the equivalent of changing an int to a long.
|The result of that in the presence of HN is a massive amount of monkey-work
|to visit and change every occurrence of the variable when the only reason
|for the change is to replace iCount with lCount and nothing more. Since I
|hate doing manual typographic changes like that and it can be difficult or
|impossible to write a script to do it automatically in all files involved,
|I find most often that it does not get done ( and I can't really blame
|the people who didn't do it ).

|Changes in semantics do happen, but usually as a side-effect of a major
|change in the fundamental way the code itself works, not as a side-effect
|of a simple type change. As such they require code changes completely
|above and beyond the variable names.

IMHO all HN tries to add to the language, is a (weak) notion of domains
of objects. I thought (but correct me if I'm wrong) that HN adds a name
prefix to an object name for the 'deeper notion' (read: the semantics or
domain); it does not prefix every object name with a short prefix
indicating the type used to implement the domain of an object.

Prefixing 'psz' to every zero terminated string is plain silly if we're
talking about, say, employee names or beer brand names; something like
'empnm' and 'beernm' should have been used instead. Only then, statements
like:

    if (!strcmp(empnmBoss, beernmMyFavBeer))

trigger the reader that something fishy is going on here ... HN, as it
is used now, doesn't add any clarity to the code at all, i.e. the statement:

    if (!strcmp(pszBoss, pszMyFavBeer))

just tells me that I want to compare my favorite beer and a boss's name,
which is silly too, but it's not the 'psz' prefix that rings a bell here;
it's just the 'Boss' and 'MyFavBeer' name parts that catch the eye.

Just adding domain name prefixes to the names of objects doesn't help
anyone; notion of domains should be present in the language itself if
one really wants domains, instead of just comparable types as implemented
in the C language. Any (decent) C compiler feels perfectly fine if one
feeds it something like:

    #include <stdio.h>
    #include <string.h>

    typedef char* empnm_t;
    typedef char* beernm_t;

    int main() {

    empnm_t empnmBoss     = "Clinton";
    beern_t beernMyFavBeer= "Grolsch";

    if (!strcmp(empnmBoss, beernMyFavBeer))
        printf("huh?\n");

    return 0;

    }

The type definitions don't help, the prefixes don't add any clarity
and still the program would run fine. If domains would have been implemented
in C, both typedefs could have been changed into something like:

    domain char* empnm_d;
    domain char* beernm_d;

and the compiler would have complained while parsing the arguments of
the strcmp() function, i.e. two different domains were compared ...

But if domains were implemented in C, we still don't need those explicit
prefixes, because the compiler would have warned us about those domain
violations, and, like you wrote above: if the semantics change, i.e. the
domains change, one still has to change those prefixes accordingly if
one used them ...

kind regards,

Jos aka [e-mail]
--
Atnwgqkrl gy zit vgksr, ug qshiqwtzoeqs!