Please accept my apologies. I'm terribly embarrassed:-] After so many years reading Slashdot, it's easy to be a cynic.
On topic, I prefer to post-process my photos with programs I wrote or I can study the source of. It's relatively easy to duplicate some of the algorithms involved (beware, there are some bugs in ColorSync in Lab color space processing that may explain some of the results in the Ars Technica review). But I put out many less photos and I am very selective. I definitely see the need for the exhaustive organizing capabilities of Aperture.
Keep on with your great posts and keep us informed if Aperture continues to deliver for you.
I see I've been labeled a troll. If the distinguished moderators had bothered to follow the link, once again: http://discussions.apple.com/thread.jspa?threadID= 258656&tstart=0 they would have seen that Mr. analogueblue pasted Devon Hillard's post into this esteemed forum. I doubt they are the same person.
That's my experience as well for my programs in plain C (again numerical stuff). I have this theory that ICC outperforms GCC only for less than ideal code written by naive scientists.
just to finish this thread: i beleived this is a bug in gcc. being a perseverative person:-), i just downloaded lcc and give it a shot. now i'm _convinced_:-). yep, this is a bug.
ok, thank you for the explanations. i'm an electrical engineer, working into satellite data processing:-)
but i suppose you will find mildly amusing the results of this:
int main(void){
int a = 10;
printf("%d\n", (int) (a *.6 ));
printf("%d\n", (int) (a *.6f));
give it a try. if you hold the result of multiplication in an intermediary variable, the result is correct. what i find interesting is that it gives incorrect results only for 0.6 and 0.7.
=)
look, if you don't cast to int the second expression you get 10.000000. i also tried it on my alpha with both compilers, gcc & ccc (compaq) and i always get 10 = 10 irrespective of ieee compliancy for floating point and direction of rounding. because casting to an int is _NOT_ equivalent to a floor, it's implementation dependent. trust me, i checked for it last week. it's in k&r 2nd ed., somewhere in appendix c.
actually they don't include very much proprietary software, just a huge amount of free software (think redhat + powertools - or whatever it's called). i think there is also a running image on a cd. you get a lot of value for the money and it's very high quality. the iso it's just the bare distribution - later (if it's not already on the ftp) you can get all of it, something like 6 cd's, ouch!
hi there. i was born and lived most of my life in such a country (i.e. Romania). i can tell you that now it is kind of worse. why? because of stupid leaders that destroyed the hope of people. yeah, sure, there are lots of lights blinking over spiffy ads and the like, but inside the soul of people is cracking, they don't see the exit from misery (even the moral one).
-bogdan p.s. if you ever have a chance, please help Romania (even by learning about it from reliable sources)
Slashdot Mirror
Why not read up a bit:
i ndex.html
http://www.nasa.gov/mission_pages/stereo/mission/
http://secchi.nrl.navy.mil/
For example, this mission could be important for understanding how to protect humans out in the solar system.
Please accept my apologies. I'm terribly embarrassed :-]
After so many years reading Slashdot, it's easy to be a cynic.
On topic, I prefer to post-process my photos with programs I wrote or I can study the source of. It's relatively easy to duplicate some of the algorithms involved (beware, there are some bugs in ColorSync in Lab color space processing that may explain some of the results in the Ars Technica review).
But I put out many less photos and I am very selective. I definitely see the need for the exhaustive organizing capabilities of Aperture.
Keep on with your great posts and keep us informed if Aperture continues to deliver for you.
Best regards,
Bogdan
I see I've been labeled a troll.
If the distinguished moderators had bothered to follow the link, once again:
http://discussions.apple.com/thread.jspa?threadID
they would have seen that Mr. analogueblue pasted Devon Hillard's post into this esteemed forum. I doubt they are the same person.
You don't say, you little critter...= 258656&tstart=0
http://discussions.apple.com/thread.jspa?threadID
Go back to your little wormhole!
That's my experience as well for my programs in plain C (again numerical stuff).
I have this theory that ICC outperforms GCC only for less than ideal code written by naive scientists.
sorry, Jean-Marc. guess i badly needed this coffee 8-). i woke up now (it's 3:30am here). you are right and i always use rint(3) in fact.
:o)
i got it before & i get it again, so help me god
just to finish this thread: i beleived this is a bug in gcc. being a perseverative person :-), i just downloaded lcc and give it a shot. now i'm _convinced_ :-). yep, this is a bug.
ok, thank you for the explanations. i'm an electrical engineer, working into satellite data processing :-)
.6 ));
.6f));
but i suppose you will find mildly amusing the results of this:
int main(void){
int a = 10;
printf("%d\n", (int) (a *
printf("%d\n", (int) (a *
return 0;
}
i don't think so, it promotes to higher precision: d = (double)a * .7, so it gets 7.0 as expected :-)
sorry to beat the dead horse, but how you explain that this gives the correct result (with -O0)?
.7;
int main(void) {
int a = 10;
double d;
d = a *
printf("%f = %d\n", d, (int) d);
return 0;
}
having a further look at this, i simplified it:
.7));
#include
int main(void) {
int a = 10;
printf("%f = %d\n", a * 0.7, (int) (a *
return 0;
}
give it a try. if you hold the result of multiplication in an intermediary variable, the result is correct. what i find interesting is that it gives incorrect results only for 0.6 and 0.7.
=)
look, if you don't cast to int the second expression you get 10.000000. i also tried it on my alpha with both compilers, gcc & ccc (compaq) and i always get 10 = 10 irrespective of ieee compliancy for floating point and direction of rounding. because casting to an int is _NOT_ equivalent to a floor, it's implementation dependent. trust me, i checked for it last week. it's in k&r 2nd ed., somewhere in appendix c.
i suggest reading this http://www.oreillynet.com/pub/a/network/2000/05/19 /magazine/mozilla_vm.html for a start and hold off your sarcasm.
actually they don't include very much proprietary software, just a huge amount of free software (think redhat + powertools - or whatever it's called). i think there is also a running image on a cd. you get a lot of value for the money and it's very high quality. the iso it's just the bare distribution - later (if it's not already on the ftp) you can get all of it, something like 6 cd's, ouch!
... they also changed their webserver: Apache/1.3.9 (Unix) (Red Hat/Linux)
.be
hi there.
i was born and lived most of my life in such a country (i.e. Romania). i can tell you that now it is kind of worse. why? because of stupid leaders that destroyed the hope of people. yeah, sure, there are lots of lights blinking over spiffy ads and the like, but inside the soul of people is cracking, they don't see the exit from misery (even the moral one).
-bogdan
p.s. if you ever have a chance, please help Romania (even by learning about it from reliable sources)