Board index » cppbuilder » Re: BDNtv: C++Builder in Delphi

Re: BDNtv: C++Builder in Delphi


2004-12-20 09:48:16 PM
cppbuilder100
Andrue Cope [TeamB] < XXXX@XXXXX.COM >wrote:
Quote
Hendrik Schober wrote:

>>My main machine at home is a notebook. It has a 3.4G P4 with 1GB of
>>RAM. That's better than my desktop at work :)
>
>
>Switch your employer.
>
>Schobi

Lol. We usually get upgraded in January [...]
I see. :) Then it's only temporary.
(I had been in a shop where the boss said
that developers should have slow machines,
so that they make faster software. I quickly
switched and never looked back -- except in
disgust.)
Quote
[...] but would you give anything back
to the IT department if they didn't ask?
I don't know. I never worked fo a company
big enough to ask that question. Here it's:
Explain the boss why you need it, and if it
pays off in a reasonable time, you'll get
it -- no matter what it is. For some things
there's no need to explain the need. (I
think I asked for halve a dozen books this
year and got them all instantly.)
Quote
[...] it's kind of cool having a 3.4 TB volume on your
personal machine.
I know enough of your job to know you guys
really /are/ a FS geeks... <g>
Schobi
--
XXXX@XXXXX.COM 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
 
 

Re:Re: BDNtv: C++Builder in Delphi

Hendrik Schober wrote:
Quote
(I had been in a shop where the boss said
that developers should have slow machines,
so that they make faster software.
Yeah I once worked for a boss who didn't let us have colour screens
because it would stop us thinking up zany schemes :-/
--
Andrue Cope [TeamB]
[Bicester, Uk]
info.borland.com/newsgroups/guide.html
 

Re:Re: BDNtv: C++Builder in Delphi

While I do not aggree with that boss, I understand what
prompted that. Often developers of tools will do it on
machines with unusually massive amounts of memory
exceptionally fast disks and very high CPU speeds while
the customers who use the products typically are using
normal setups.
The customers then find the tools to be slow resource
hogs. Even worse they find their customers (the end
users) complaining of difficulties with the products.
It might be a good idea to require developers to do a
pass on a normal machine before releasing their work
to the next stage.
. Ed
Quote
Hendrik Schober wrote in message
news: XXXX@XXXXX.COM ...

(I had been in a shop where the boss said
that developers should have slow machines,
so that they make faster software. I quickly
switched and never looked back -- except in
disgust.)
...
 

{smallsort}

Re:Re: BDNtv: C++Builder in Delphi

Hendrik Schober wrote:
Quote
(I had been in a shop where the boss said
that developers should have slow machines,
so that they make faster software. I quickly
switched and never looked back -- except in
disgust.)
Actually, he was almost right. Testers should have slower (or
"average") machines.
--Steve
 

Re:Re: BDNtv: C++Builder in Delphi

This is exactly what we do.
We need to use top of the range (and as clean as possible) PC's as anything
else would impinge on our productivity.
But, of course we use "real world" spec PC's to benchmark our applications
from time to time.
Stewart
"Ed Mulroy [TeamB]" < XXXX@XXXXX.COM >wrote in message
Quote
It might be a good idea to require developers to do a
pass on a normal machine before releasing their work
to the next stage.

. Ed
 

Re:Re: BDNtv: C++Builder in Delphi

Stephen Waits < XXXX@XXXXX.COM >wrote:
Quote
Hendrik Schober wrote:
>(I had been in a shop where the boss said
>that developers should have slow machines,
>so that they make faster software. I quickly
>switched and never looked back -- except in
>disgust.)

Actually, he was almost right. Testers should have slower (or
"average") machines.
I wasn't a tester. I was a developer.
Quote
--Steve
Schobi
--
XXXX@XXXXX.COM 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
 

Re:Re: BDNtv: C++Builder in Delphi

Ed Mulroy [TeamB] wrote:
Quote
It might be a good idea to require developers to do a
pass on a normal machine before releasing their work
to the next stage.
That is why our test machines are of the same spec as our users.
Test machines are not only a great idea for catching performance
problems, but issues that might arise in deployment because you forgot
you were using an ActiveX control, or had created certain config files,
or numerous other reasons that files installed on the developer machine
or not in the release environment. Releasing CodeGuarded builds was a
classic mistake this picked up! (although we have better QA to stop
CodeGuard leaking out in the first place now ;?)
AlisdairM(TeamB)
 

Re:Re: BDNtv: C++Builder in Delphi

"Ed Mulroy [TeamB]" < XXXX@XXXXX.COM >wrote:
Quote
While I do not aggree with that boss, I understand what
prompted that. [...]
So do I. What he, however, didn't accept
was that to do that, developers would
also be a lot slower. Mind you, this was
a software meant to manipulate and print
huge bitmaps -- huge meaning anything
above 100MB. In the end, the minimum
requirements for running the app were
above what we had for debugging it.
I was doing threads for this using BCB3
on Win95. (NT4 would crawl on this box.
There was an nad MT bug in the std lib's
'std::string' class of BCB3 and I was
booting the box five times per hour. That
was taking almost five minutes each. Go
figure how much time I had left to do
some work...
Heck, it took the de{*word*81} 10 seconds
just to jump to the next line! Instead of
debugging the damn thing, we put in
gazillions of logging statements -- some
of which certainly got forgotten and in
the end caused decreased performance.
Developers were putting off badly needed
changes, because they required changing
some headers -- which meant we all would
be staring at the compiler counting lines
for four hours. In the end, some changes
got lost, some were forgotten, and some
were put off infinitely. -- All this badly
hurting the app's stability.
After a few months of this I asked the boss
to sit beside me as I would like to show
him something in the app we were developing.
I then "accidently" touched a header, which
meant I first needed to re-compile some of
the project. We use the time to talk about
a few issues of the app. But then we hit a
forgotten breakpoint and had to wait until
the system swapped out the app to load the
de{*word*81}. Then -- again just "accidently" --
I stepped over the code exposing what I
wanted to show him and had to restart all
this again.
After 45mins it happened. He almost went
through the roof yelling at me I would waste
his precious time. Well, of course that's
exaclty where I wanted to have him and I had
some telling and yelling to do, too. Three
days later I had a new machine.
Nevertheless, another few months later I left
the shop. I had to learn that, where they get
this wrong, they get other things wrong, too.
I still regret I put up with this for so long.
That's why I'm touchy when someone mentions
he is working on an inferiour machine.
Quote
It might be a good idea to require developers to do a
pass on a normal machine before releasing their work
to the next stage.
This is something very different to which
I certainly agree.
Quote
. Ed
[...]
Schobi
--
XXXX@XXXXX.COM 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
 

Re:Re: BDNtv: C++Builder in Delphi

"Hendrik Schober" < XXXX@XXXXX.COM >writes:
[snip]
Quote
That's why I'm touchy when someone mentions
he is working on an inferiour machine.
I'm working on an inferior machine, for two reasons:
1: I don't want to switch to another machine to do testing.
2: Switching to a 3.x GHz machine from my 550 MHz one would save me
almost no time at the end of the day, except when I do profiling or
memory checking, which is a rare event and would be slow on a fast
machine as well.
[snip]
--
Oscar
 

Re:Re: BDNtv: C++Builder in Delphi

Oscar Fuentes < XXXX@XXXXX.COM >wrote:
Quote
[...]
2: Switching to a 3.x GHz machine from my 550 MHz one would save me
almost no time at the end of the day, except when I do profiling or
memory checking, which is a rare event and would be slow on a fast
machine as well.
Good for you.
Here, compile and link times are a
significant issue.
Schobi
--
XXXX@XXXXX.COM 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
 

Re:Re: BDNtv: C++Builder in Delphi

Hendrik Schober wrote:
Quote

Good for you.
Here, compile and link times are a
significant issue.

Same here. The systems we shipped with our instruments at my last
company were dual P4s but very expensive (just when P4s came out) but
compilation speed got a huge increase compared to our single processor
P3 dev machines.
Cheers
Russell
 

Re:Re: BDNtv: C++Builder in Delphi

"John Kaster (Borland)" < XXXX@XXXXX.COM >wrote in <41c389ea$1
@newsgroups.borland.com>:
Quote
bdn.borland.com/article/0,1410,32862,00.html

Let us know what you think. Thanks for your time and attention.

That's a wonderfull christmas gift :) Keep up the good work :)
--
Regards from Quebec.
David Charron
 

Re:Re: BDNtv: C++Builder in Delphi

Quote
Does that mean that C++ projects will finally be able to use precompiled
Delphi packages? As it currently stands, BCB has never been able to do
that. To use Delphi components in BCB, the Pascal source files have to be
re-compiled using BCB's compilers.
And the worse part: the Delphi-To-C++ translator has some serious problems
sometimes, that render the headers not really usable, or the delphi sources
must have a lot of tricks and stuff like that so the hpp works from BCB.
If it's going to be as easy as use the Delphi packages, then it will be just
GREAT!
Regards,
--
Rodrigo Gómez
rgomez.msa.com.mx/gallery/
"Remy Lebeau (TeamB)" < XXXX@XXXXX.COM >escribi?en el mensaje
Quote

"John Kaster (Borland)" < XXXX@XXXXX.COM >wrote in message
news: XXXX@XXXXX.COM ...

>No package incompatibilities any more, or at least nothing like they
>used to be (gotta leave myself some wiggle room) ... THIS is what's
>going to "Rock!" for C++Builder customers. You'll no longer be feeling
>like you're always one step behind.

Does that mean that C++ projects will finally be able to use precompiled
Delphi packages? As it currently stands, BCB has never been able to do
that. To use Delphi components in BCB, the Pascal source files have to be
re-compiled using BCB's compilers.


Gambit


 

Re:Re: BDNtv: C++Builder in Delphi

Hendrik Schober wrote:
Quote
Here, compile and link times are a significant issue.
In C++, yes, compile and link time is awful. Delphi is fast... an order of
magnitude faster. Did I mention it's fast? :-)
 

Re:Re: BDNtv: C++Builder in Delphi

Lee Grissom < XXXX@XXXXX.COM >wrote:
Quote
Hendrik Schober wrote:
>Here, compile and link times are a significant issue.

In C++, yes, compile and link time is awful. Delphi is fast... an order of
magnitude faster. Did I mention it's fast? :-)
1. C++ compilers spent a considerable amount
of that extra time parsing templates and
doing other stuff that Delphi doesn'tdo. I
especially like templates as they allow me
to do more stuff at compile time instead
of at run-time. If this fails, fails on my
desk, not on the customer's.
2. In our case, most of the time wasted is due
to some <insert swear word here>including
everything but the kitchen sink.
Schobi
--
XXXX@XXXXX.COM 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