[fpc-pascal]size/speed/compiler - Was:another fpc RAD: MSEide

Florian Klaempfl florian at freepascal.org
Wed Apr 19 21:49:12 CEST 2006


L505 wrote:
> Okay so we have to now consider these points:
>  1. interpreted languages can take up less memory if engineered right (says Florian)
>  2. compilation and linking is a hassle - compared to shipping or uploading an interpreted
> file
>  3. speed - not a big deal. Hardware cheap enough.
>  4. size - shipping an interpreted file usually smaller than exe/elf
>  5. uploading - uploading a 30K script file on a 56k modem is easier than uploading a 5MB
> binary
> 
> What advantage of compilation? There must be something.

2. is true but: with a interpreted language your user has the hazzle to
get proper libraries for the interpreter, get the interpreter itself
etc. while with a compiler the developer who should know his tools has
hazzle.

3. isn't true for real interpreted languages, they are really slow, we
aren't talking about JIT like Java or .NET but things like Python, Perl,
PHP they are magnitudes slower than compiled languages. Don't take the
advantages of the Python (small memory footprint) and Java (JIT, i.e.
not that slow) and say a compiled program is bad because it is a) not
much faster than java and b) takes more memory than Python :) You
overlook then that a compiled program usually takes a) less memory than
a jrt and b) is magnitudes faster than Python.

4.+5. is also only partly true. What shall I do with a 30K script if I
miss the interpreter and runtime environment with 10-20 MB?

> 
> I know Microsoft use to love compiling those EXE's because you couldn't see the source
> code to the EXE's.
> Is this the advantage of FPC? That we can hide our sources?
> 
> :-)
> 
> 
> _______________________________________________
> fpc-pascal maillist  -  fpc-pascal at lists.freepascal.org
> http://lists.freepascal.org/mailman/listinfo/fpc-pascal
> 




More information about the fpc-pascal mailing list