 |
BorlandTalk.com Borland discussion newsgroups
|
| View previous topic :: View next topic |
| Author |
Message |
Colin B Maharaj Guest
|
Posted: Sat Dec 04, 2004 1:47 pm Post subject: Defining the Death of a Compiler or Development Environment |
|
|
If I did not see it I would not believe it. I am doing an installation
of the telecom software for Call Accounting from...
http://www.avotus.com/site/sub_prodservices/icm-enhanced_call_accounting.html
....at the new Marriot Hotel here in Trinidad.
What is interesting is I am dealing with the Property Management
Interface guy whose function (among others) is to tie in the card
readers, the commands to open and close rooms to the PABX and (my part)
to send telephone call records made by the guest, from the Package to
the PMS system. All this is done over a serial interface. And what I see
is incredible, it looks like most of the interface modules are MS-DOS
serial applications using Clipper files.
This apparently is the non-windows version of Fidelio that still manages
hotels like the Marriot.
What I have realised over the years too is that Windows 2000 and XP
actually have better support for 16 bit apps than say NT or ME
Thing is, if it works , why fix it. My Dev environment is still Win 2000
and I moved up from Win95 - I did not even spend much time in 98.
Microsoft have documented via their LOGO program, specific ways to have
apps run in their OSs, with code snippets on detecting OS version,
Service packs etc. and with stuff like LoadLibrary etc... you can still
call Win200 / Win2003 or XP specific code from BCB 6.
My Dev environment is still BCB3 and I have created apps working in Win
95B all the way up to Win2003 Server.
Like I said many moons ago, when the PE file format has changed, then we
are in trouble.
http://tinyurl.com/4b53
|
|
| Back to top |
|
 |
Randall Parker Guest
|
Posted: Tue Dec 07, 2004 12:56 am Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
Colin,
Why do you use such an old version of BCB? I'd hate to go back even as far as v4.
Colin B Maharaj wrote:
| Quote: | My Dev environment is still BCB3 and I have created apps working in Win
95B all the way up to Win2003 Server.
|
|
|
| Back to top |
|
 |
Colin B Maharaj Guest
|
Posted: Tue Dec 07, 2004 11:20 am Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
Sadly, I have not gotten the time to 'port' everthing to BCB 6, which I
did buy about a year ago. I did start certain utility applications and
simple communication applications using BCB6 which are essential for my
project, but the largest EXEecutable is still in BCB3.
Thing is, I am doing two jobs, one as a full time telecoms technician
here in Trinidad and the other as a Programmer - part time. I am hoping
to leave the current full time job to give myself time to do the right
thing. But it is still interesting that somehow - BCB3 can create fresh
applications that would work on both Legacy and modern OSs - with some
effort.
Like I said when the PE format is gone then we are in trouble.
Randall Parker wrote:
| Quote: | Colin,
Why do you use such an old version of BCB? I'd hate to go back even as
far as v4.
Colin B Maharaj wrote:
My Dev environment is still BCB3 and I have created apps working in
Win 95B all the way up to Win2003 Server.
|
|
|
| Back to top |
|
 |
Henry Hansen Guest
|
Posted: Wed Dec 08, 2004 8:54 pm Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
Colin B Maharaj <noreply (AT) myhost (DOT) com> wrote in
news:41b1c007$1 (AT) newsgroups (DOT) borland.com:
| Quote: |
My Dev environment is still BCB3 and I have created apps working in
Win 95B all the way up to Win2003 Server.
|
In 20-20 hindsight, I wish I would have spent far more time simply
improving my programs, learning new algorithms and writing C++ code than
upgrading and upgrading and upgrading ...
A carpenter can build a house that stands for decades and have something
to show for his work. A programmer is lucky if his program even
compiles and runs in a few years. I don't think this culture of
throwing away our work in a constant treadmill of language changes and
upgrades is very wise.
|
|
| Back to top |
|
 |
Duane Hebert Guest
|
Posted: Wed Dec 08, 2004 9:45 pm Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
"Henry Hansen" <hh1000 (AT) h-o-t-m-a-i-l (DOT) com> wrote
| Quote: | Colin B Maharaj <noreply (AT) myhost (DOT) com> wrote in
news:41b1c007$1 (AT) newsgroups (DOT) borland.com:
My Dev environment is still BCB3 and I have created apps working in
Win 95B all the way up to Win2003 Server.
In 20-20 hindsight, I wish I would have spent far more time simply
improving my programs, learning new algorithms and writing C++ code than
upgrading and upgrading and upgrading ...
A carpenter can build a house that stands for decades and have something
to show for his work. A programmer is lucky if his program even
compiles and runs in a few years. I don't think this culture of
throwing away our work in a constant treadmill of language changes and
upgrades is very wise.
|
Agreed. But that's the nice thing about standards and compliance. Most of the
code that I've thrown away in the last several years is GUI stuff and that's usually
due to clients (or management) expecting a new style/flashier interface. I'm still using
classes to do fuzzy calculations and other control stuff that I wrote a few years
ago. Most of it is the same as it was with BCC except that it was updated to
use std containers etc. It's gone through all the BCB iterations, MSVC and
still works on Linux boxes.
I think this is a compelling reason to get a good compliant
compiler, have a strict separation between GUI and non gui stuff
and write compliant code without depending on non-standard
extensions when not imperative to do so.
|
|
| Back to top |
|
 |
Ed Mulroy [TeamB] Guest
|
Posted: Thu Dec 09, 2004 1:32 am Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
That's been my experience as well. I did a little job for
someone last year which contained code that I first wrote
for DOS in the late 80's.
In the late 80's the UI was DOS code. In '92 it received a
Windows 3.1 EasyWin interface (yeah, I know, but that's
what they wanted). In '97 it was a Win32 GUI with
the DOS version as the stub (so the same binary ran
on both his DOS and Win95 machines). Now it's a
Win32 UI only (still simple - I dislike a UI that sacrifices
clear, easy operation on the altar of "looks pretty").
All those programs have the same inner workings.
Depending upon what you mean by a "good compliant
compiler", I might dissagree. Some must write using the
very latest thing, but I've found that a stable, maintainable
code base is not the most likely place to find code from
the the Lambda libraries. I wouldn't push the envelope
in what I charge money for unless the customer
requested it.
.. Ed
| Quote: | Duane Hebert wrote in message
news:41b775e9 (AT) newsgroups (DOT) borland.com...
I think this is a compelling reason to get a good
compliant compiler, have a strict separation between
GUI and non gui stuff and write compliant code without
depending on non-standard extensions when not
imperative to do so.
|
|
|
| Back to top |
|
 |
Hendrik Schober Guest
|
Posted: Thu Dec 09, 2004 2:19 am Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
Ed Mulroy [TeamB] <dont_email_me (AT) bitbuc (DOT) ket> wrote:
| Quote: | [...]
Depending upon what you mean by a "good compliant
compiler", I might dissagree. Some must write using the
very latest thing, but I've found that a stable, maintainable
code base is not the most likely place to find code from
the the Lambda libraries. I wouldn't push the envelope
in what I charge money for unless the customer
requested it.
|
Depending on what you call "the very
latest thing". <g> Some might argue that
AOP is it, and that TMP is out already.
I wouldn't use the Lambda lib nowadays,
because I am expecting to many problems
with some of the compilers we use. OTOH,
I do use some of boost (I think you do
despise us weak guys needing smart ptrs
to clean up after us >, but how about
'boost::array<>' to have a plain old C
array play nicely with the STL?) and I
did write some TMP stuff myself.
Latest outcome, still hot from just
being typed last night, is a logging lib
that will fold this calls
log<Debugger,Warn>( ORIGIN_HERE, "Live is @(liv), mana is @(man)", ERROR_PAR(liv)+ERROR_PAR(man) );
to a single 'if' at runtime, if there is
no log target installed which takes the
message. Only if there is one, the error
parameter list will be created, and
everything passed to it. If I disable
logging for some targets (e.g. the
'Debugger' target from above) or some
severities (like 'Warn') at /compile-time/,
VC7.1 will optimize the /whole line/ to
nothing. Of course, this is demanding on
compilers, but if I get this to compile
and run on all that we use, it will be
great. (We had found logging to be a
performance bottleneck in some places
even if no log targets were installed due
to all the parameter lists being created
and thrown away for nothing.) And with
compilers catching up on std conformance,
porting this to newer ones will only become
easier every year.
So, while this probably might use "the
latest thing" according to your definition,
it does give some real value and hopefully
will still compile and do what it is
expected to do in a few years. (Well, at
least on Windows it will compile in ten
years, when VC is fully conformant. <g> )
Schobi
--
[email]SpamTrap (AT) gmx (DOT) de[/email] is never read
I'm Schobi at suespammers dot org
"The presence of those seeking the truth is infinitely
to be prefered to those thinking they've found it."
Terry Pratchett
|
|
| Back to top |
|
 |
Andrue Cope [TeamB] Guest
|
Posted: Thu Dec 09, 2004 9:17 am Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
Henry Hansen wrote:
| Quote: | A carpenter can build a house that stands for decades and have
something to show for his work. A programmer is lucky if his program
even compiles and runs in a few years.
|
Speak for yourself :)
Our run time library originally started as a .lib for MSDOS
applications built using TC3.1. After a while it got converted into a
lib/dll for Windows 3.x using BC4.5. Then it was ported to Win32 using
BC5.01. Then BCB3, then BCB4, then BCB5 now BCB6.
There are parts of it that are not as well coded as others but we are
removing most of them now. This year we proactively removed/rewrote
anything that looked ugly. Next year we are going to review every one
of the one hundred and ninety units. It's a witch hunt!
I leave you with this comment from one of the header files:
/*
*START
FILE NAME : HSP.cpp
CREATED BY : Andrew Cope DATE : 15.1.92
COPYRIGHT : Vogon Intl. Ltd. FILE TYPE : C++ source
DESCRIPTION : Code for HSM (handle sharing module)
:
: 05/01/99 - Updated to compile under WIN32
Ah it brings back memories, that does. It also reminds me that I've
been working on the same project for over twelve years. Different
versions, thankfully and it has grown a lot but still :-/
--
Andrue Cope [TeamB]
[Bicester, Uk]
http://info.borland.com/newsgroups/guide.html
|
|
| Back to top |
|
 |
Hendrik Schober Guest
|
Posted: Thu Dec 09, 2004 10:50 am Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
Andrue Cope [TeamB] <no.spam (AT) not (DOT) a.valid.address> wrote:
| Quote: | [...]
I leave you with this comment from one of the header files:
|
Wow, you do have a brother Andrew???
You can't leave us with that without
telling the whole story! <g>
Schobi
--
[email]SpamTrap (AT) gmx (DOT) de[/email] is never read
I'm Schobi at suespammers dot org
"The presence of those seeking the truth is infinitely
to be prefered to those thinking they've found it."
Terry Pratchett
|
|
| Back to top |
|
 |
Andrue Cope [TeamB] Guest
|
Posted: Thu Dec 09, 2004 11:10 am Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
Hendrik Schober wrote:
| Quote: | I leave you with this comment from one of the header files:
Wow, you do have a brother Andrew???
You can't leave us with that without
telling the whole story!
|
Andrew is my real name. I use 'Andrue' for online purposes. It's a
great way of ensuring you can have the user name you want :)
It did cause a bit of confusion at Manchester airport when I went to
pick up my tickets for BorCon but I just told the guy they were bought
for me by an American company and Americans can't spell :D
--
Andrue Cope [TeamB]
[Bicester, Uk]
http://info.borland.com/newsgroups/guide.html
|
|
| Back to top |
|
 |
Duane Hebert Guest
|
Posted: Thu Dec 09, 2004 3:20 pm Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
"Ed Mulroy [TeamB]" <dont_email_me (AT) bitbuc (DOT) ket> wrote
| Quote: | In the late 80's the UI was DOS code. In '92 it received a
Windows 3.1 EasyWin interface (yeah, I know, but that's
what they wanted). In '97 it was a Win32 GUI with
the DOS version as the stub (so the same binary ran
on both his DOS and Win95 machines). Now it's a
Win32 UI only (still simple - I dislike a UI that sacrifices
clear, easy operation on the altar of "looks pretty").
All those programs have the same inner workings.
|
Our first "gui" was in dos using BCC. It's gone through
BCB 3->6->VC/Qt. Most of the non gui stuff hasn't changed.
| Quote: | Depending upon what you mean by a "good compliant
compiler", I might dissagree. Some must write using the
very latest thing, but I've found that a stable, maintainable
code base is not the most likely place to find code from
the the Lambda libraries. I wouldn't push the envelope
in what I charge money for unless the customer
requested it.
I'm not talking about the very latest thing but standard lib/STL |
stuff (including boost) needs to compile. The problem's that
we've had going from BCB to say G++ have been minimal
(things like Borland's math.h that aren't compatable) I can
currently build the majority of my non GUI stuff on BCB/G++/VC7.1
with only a couple of ifdefs. This makes debugging a bunch
more effective.
If CBX gets updated and works with Qt I should
be able to do nearly the whole thing on all three.
|
|
| Back to top |
|
 |
Henry Hansen Guest
|
Posted: Thu Dec 09, 2004 4:58 pm Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
"Andrue Cope [TeamB]" <no.spam (AT) not (DOT) a.valid.address> wrote in
news:41b8182e$1 (AT) newsgroups (DOT) borland.com:
| Quote: | A carpenter can build a house that stands for decades and have
something to show for his work. A programmer is lucky if his program
even compiles and runs in a few years.
Speak for yourself :)
Our run time library originally started as a .lib for MSDOS
applications built using TC3.1. After a while it got converted into a
lib/dll for Windows 3.x using BC4.5. Then it was ported to Win32 using
BC5.01. Then BCB3, then BCB4, then BCB5 now BCB6.
|
Yes ... the intent of my original comments was that I wish I'd used more
C/C++ over the years (as I too have 10 year old C code producing
mission critical results). Its the other dead end programming languages
that I've wasted time with such as VB6, ASP, Clipper, FoxPro ... and
even possibly Delphi, that make me think I should have done all my work
in C/C++, even if it required more work. Heck my C programs have
outlived complete application life cycles for VB6 apps, that have long
since been forgotten.
|
|
| Back to top |
|
 |
Mike Margerum Guest
|
Posted: Thu Dec 09, 2004 9:23 pm Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
| Quote: | I think this is a compelling reason to get a good compliant
compiler, have a strict separation between GUI and non gui stuff
and write compliant code without depending on non-standard
extensions when not imperative to do so.
Anyone have any good documentation/books to read for approaches on doing |
this?
|
|
| Back to top |
|
 |
Chris Uzdavinis (TeamB) Guest
|
Posted: Thu Dec 09, 2004 10:07 pm Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
Mike Margerum <mike (AT) junk (DOT) com> writes:
| Quote: | I think this is a compelling reason to get a good compliant
compiler, have a strict separation between GUI and non gui stuff
and write compliant code without depending on non-standard
extensions when not imperative to do so.
Anyone have any good documentation/books to read for approaches on
doing this?
|
The technique goes by many names, but one of the most prominent is
"Model-View-Controller (MVC) pattern".
The first link I found on google documenting this (briefly) is by
Apple:
http://tinyurl.com/5p98v
(http://developer.apple.com/documentation/Cocoa/Conceptual/AppArchitecture/Concepts/MVC.html)
--
Chris (TeamB);
|
|
| Back to top |
|
 |
Colin B Maharaj Guest
|
Posted: Fri Dec 10, 2004 11:07 pm Post subject: Re: Defining the Death of a Compiler or Development Environm |
|
|
Can't believe this, for the past week I have been using my 'old' laptop
I got in '96 with a Terminal Emulator I did in DOS back in '97. I/We
have been using this emulator to administrate the PABX at the Marriot,
Trinidad. I had also build a scripting engine in the terminal and still
using it to make tedious/monotonous, changes to the 100+ extensions in
the Hotel.
I have not used the terminal in years (or the laptop) but the VT220/420
emulation works great and we use it for the Voice Mail Admin and works
great! - Still!
Wrote it in Borland C++ 3.0 for DOS.
Heck my C programs have
| Quote: | outlived complete application life cycles for VB6 apps, that have long
since been forgotten.
|
|
|
| Back to top |
|
 |
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|