Op woensdag 16-01-2008 om 09:54 uur [tijdzone +0100], schreef Joost van
der Sluis:
> Don't get me wrong, but some discussions are useless and do more harm
> than good for the Lazarus. project imho. This is one of hem. Normally
> I
> would simply not answer. But this takes too long imho, and too much
> time
> which can be spend much better.
Some more information.
When people try to use Lazarus for the first time. They write 'hello
world', compile, run. Add a breakpoint, compile, run. If debugging then
doesn't work, they'll toss Lazarus away. If they see lines with
hexadecimal numbers, they'll start googling or complaining.
Next step is that they compile one of their 'old, rusty' delphi/pascal
programs. To see if that still works.
That are the things we need to think about while developing Lazarus.
The binary size - no one is interested in that. Only people who are
coming from Delphi. And that's because the Delphi compiler is written in
an time in which executable size was important! You only had 2400 baud
modems, and your harddisk had 20MB of storage space! That's the reason
why Delphi has all those tricks to keep files as small as possible, even
while you're developing. Nowadays no-one really cares. Unless you are
obsessed with comparing every detail of FPC/Lazarus with Delphi.
About the compilation time, some people argue that that would be faster
if the binary size (or the size of the .o files) is smaller. Well,
hurray! Compilation time drops from 1.2 seconds to 1.198 seconds That's
0.02 seconds faster! Fast, submit it to slashdot! That will learn those
C-users to bash pascal!
Same holds for executable-loading times. (According to the benchmarks
fpc is one of the fastes compilers in this aspect)
A thing which does matter, if the amount of memory the application uses
while running. I'm still annoyed by the fact that the 'clock' in the
corner of my screen consumes 5MB of internal memory, just to show the
time. And then I didn't count the python-runtime which is loaded to use
it. GNOME (Fedora) can take up to hundred or even more MB of memmory,
just to show some widgets. That annoys me. So I'm happy with Lazarus
which only needs twice the amount of internal memory then a stupid,
simple, clock.
I doubt that any non-Delphi user will ever make an issue about the
file-size. Until they want to distribute their applications, in which
case they'll have no problems with the strip option.
Offcourse, the core-developers could spend some time on an option to
strip the debug info from the LCL, but not of the whole application. But
it's importance is the same a typo on the about screen.
While the effort needed to implement it, is comparable to implementing
the basics of a out-of-the-box CGI-web system. (Or some steps further in
the development of the internal debugger)
I'm really happy with the Lazarus developers, who would spend their time
on the CGI-functionality, instead of useless functionality like typo's
and lcl-debuginfo-strip-options.
That's also the reason why Lazarus has become what it is, and the amount
of users is growing that fast: The project is driven by a team of
excellent developers, who know where they should spend their time on.
The same holds, for example, for the iconset. I'm pretty pleased that
the core people didn't spend much time on it. And I'm pretty pleased
with Laurent and others, who have picked this up and made many patches
with new and improved Icons. They didn't start long threads on the
mailinglist to ask others to code a solution for their - from a
core-developer view - unimportant options. No they started long threads
with all sort of self-made Icons, compared them, and chooses the best of
them. Resulting in a far better looking Lazarus in just a couple of
months.
>From time to time this threads pops up. Just like the 'icon' threads did
in the past. Now take your responsibility: if you are annoyed by the
fact of the big-binaries, you have to supply patches for that. If you
think that it's a big issue which we all have been waiting for.... you
would start spending time on it, to make that patch. Even if it would
take you months, you would do it. If not, it's not a big enough issue to
spend time on. And especially not to spend others time on, writing long
mails to the mailing list. ;)
Joost.
ps: Main idea of the message above isn't even debatable. So in case
you're already writing a long answer: stop that, I won't even read. And
certainly not answer.
_________________________________________________________________
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject
archives at http://www.lazarus.freepascal.org/mailarchives
