U++ forum: Welcome to the forum

Re: Value with type float [message #58227 is a reply to message #58196]

Thu, 31 March 2022 11:51

abductee23
Messages: 5
Registered: October 2018
Location: San Francisco

Promising Member

Hi Mirek,

you have no idea how much it pains me to read that.
i was very excited when i found upp back in the day 2005 ish - and realized that i have to never worry about maintaining my own ux library... and can focus on what i do best: graphics programming.

maybe i am seeing this too much from my own viewpoint - graphics?

before upp i viewed c++ as a "necessary evil" - you have with your code, and esp your forum posts have quiet an impact.
and you seem to be so many times on what i consider "the side of an argument" when it comes to technical things.

and - to be blunt - only in recent times the c++ world has caught up to where upp was a decade ago (somehow they celebrate it by doing to their standart whatever that is they are doing...)

ok, just to make sure i don't get misinterpreted as a "hey i can't start using your stuff until you give me X Y and Z"-person(you probably have had a lot of these)

anyways - over the last 15-ish years there was not a project where not on some level i either quickly hacked together something in upp - or made it a upp thing to begin with - that goes for work and private stuff.( see pictures below - only private projects stuff obviously)

beeing able to crank out a tool in a day or less has been an invaluable asset to my work and private coding life.
(the "just drag all the pngs to this tool and if will fix the alpha channel"-incident comes to mind ^^)

and the only actual problem that i am facing over and over again: floats - love'em or hate'em - the reality is they are essential to computergraphics.

dealing with float issiues/precission is a time honered "tradition" for cg and audio(esp snyth) programmers Smile

having a flot in memory - writing it out through value/json , then reading it again incurs a conversion and its not the same number - ( i guess on a personal level i am probably puzzled how this is not bothering you - but thats beside the point)

of course - i do not disagree that that more precision is better and preferred in a lot of cases - but gpus have a significant performance penalty for using doubles and there's nothing i can do about that.(opengl/d3d are also very creative when it comes to IEEE754)

so i guess my questions are: do you see any wiggleroom when it comes to Value/Jsonize?

I want to make sure its understood this is coming from a a good place and is a genuine request - and i guess say thank you for the last 15 yrs of being there and improving upp.

Attachment: amboss.png
(Size: 829.58KB, Downloaded 168 times)
Attachment: nvscene.png
(Size: 941.70KB, Downloaded 155 times)
Attachment: the_dude_abides.png
(Size: 1.28MB, Downloaded 151 times)

Report message to a moderator