NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/lang/c++-part1/6743

Internet Message Format | 1996-08-05 | 4.1 KB

Path: nut.trin.cam.ac.uk!jad1002
From: [e-mail] (James Dennett)
Newsgroups: comp.lang.ada,comp.lang.c++,comp.lang.c,comp.lang.modula3,comp.lang.modula2
Subject: Re: Hungarian notation
Date: Thu, 8 Feb 1996 10:26:08
Organization: University of Cambridge, England
Message-ID: <[e-mail]>
References: <4cud8f$[e-mail]> <[e-mail]> <[e-mail]> <dewar.822407363@schonberg> <Pine.HPP.3.91.960124References: <[e-mail]> <4bd <4cc2b2$11jq@navajReferences: <4cud8f$[e-mail]> <[e-mail]> <[e-mail]> <dewar.822407363@schonberg> <Pine.HPP.3.91.960124References: <[e-mail]> <4bd <4cc2b2$11jq@
rganization: Trinity College, Cambridge
NNTP-Posting-Host: nut.trin.cam.ac.uk
X-Newsreader: Trumpet for Windows [Version 1.0 Rev A]

In article <[e-mail]> [e-mail] (Richard Wesley) writes:
>From: [e-mail] (Richard Wesley)
>Subject: Re: Hungarian notation
>Date: Tue, 06 Feb 1996 07:45:26 -0800

>In article <[e-mail]>,
>Douglas Evan Cook <[e-mail]> wrote:

>>On Wed, 31 Jan 1996, James Dennett wrote:
>>
>>> >No, but somebody must have written those methods, no?  And that somebody
>>> >needed to know how many bits each data type is, no?  When we changed from
>>> >16 bit compiler to 32 bit compiler, the data files remained the same.
>>>
>>> Herein lies your problem - you broke the ADT. If the ADT didn't specify the
>>> size of the components then your code can't rely on them...
>>> so the data files would not necessarily remain the same if
>>> the implementation of the ADT changed.
>>
>>Ah, but the ADT *did* specify that the file format would remain the
>>same.  And not all users of the applications are using the same platforms
>>- some are running on 386's with 386 DOS extenders, some are running with
>>286 extenders, some are running on overlaid versions of the program done
>>for straight DOS, some are running on Windows NT, and some are running on
>>non Intel based machines.  The users expect to be able to transfer files
>>between each platform.  And so we compile, but in each platform we don't
>>want to have to modify the ADT's because of compiler dependancies.  So we
>>have the ADT's depend on typedef equivalents, so that when we need to
>>compile for a new platform we only need to change 10-12 lines in the base
>>definition file.

Do we have a problem here or not?

So, your ADT was relatively concrete and specified the implementation in
some detail. (Is there a difference between an ADT and a data structure?)

In that case, there's no problem at all; you can of course hide the machine
dependencies in a module (or just a file) where you would list your typedefs
if you are happy to forego the optimisations offered by using native types
for one type of increased portability.

Alternatively, you could use the native integers internally and just write your
own I/O code. This approach would then be portable to machines of varying
endian-ness, or even those legendary machines which aren't based on 8-bit
chunks of memory. The encapsulation can also be made stronger (but then,
where did encapsulation enter the C philosophy) as these details are all
within the implementation of the ADT, rather than in a "System Independence"
module somewhere else.

Each solution  may have its place.

>So out of curiosity, how did you handle endian issues (x86s do not use
>network byte order, unlike most of the rest of the hardware in the known
>universe)

(Thanks for this point!)

James.

>- rmgw

>http://www.punchdeck.com/hawkfish/PunchDeck.html

>----------------------------------------------------------------------------
>Richard Wesley             | "I don't know about your dreams
>[e-mail]     |  But mine are sort of hackneyed"
>[e-mail]  |   - Laurie Anderson, "Talk Normal"
>----------------------------------------------------------------------------