ArrayArrayArrayArrayArrayArrayArrayArrayArrayArray BrainModular BrainModular Users Forum 2013-07-23T00:14:58+02:00 https://www.brainmodular.com/forums/app.php/feed/topic/3971 2013-07-23T00:14:58+02:00 2013-07-23T00:14:58+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27318#p27318 <![CDATA[Round Vs. Trunc]]> Statistics: Posted by sephult — 23 Jul 2013, 00:14

]]> 2013-07-22T16:06:15+02:00 2013-07-22T16:06:15+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27315#p27315 <![CDATA[Round Vs. Trunc]]> And a reminder that what we nowadays think of as the ultimate in numeric precision is in fact the opposite:
Always slightly "wrong", even if negligible ;-)

Cheers,

Tom

Statistics: Posted by Thomas Helzle — 22 Jul 2013, 16:06

]]> 2013-07-22T13:16:42+02:00 2013-07-22T13:16:42+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27314#p27314 <![CDATA[Round Vs. Trunc]]> With this patch
You would choose between trunc or round and which decimal you want to keep (how many numbers after float).

Statistics: Posted by joffo78 — 22 Jul 2013, 13:16

]]> 2013-07-22T11:29:54+02:00 2013-07-22T11:29:54+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27313#p27313 <![CDATA[Round Vs. Trunc]]> I am guilty for not always going to the wiki, still getting used to that. I did however this time and saw....
Constant
Constant value.
out
Value of the constant
So I became very confused :lol:
Maybe we should update this to say it is stored as a Float value, and will display as such?

Thanks Thomas/Marignasse!

Statistics: Posted by sephult — 22 Jul 2013, 11:29

]]> 2013-07-22T10:20:51+02:00 2013-07-22T10:20:51+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27311#p27311 <![CDATA[Round Vs. Trunc]]>
Okay I tested and saw the expected results using a constant.
Left me curious about constants though...I set constant to 12.85 and in the parameters it automatically sets = 12.8500003814697
This leaves me curious.
Yeah, that is something rather irritating in Usine - float values change, even if very clear values are used that should normally be represented without precision problems like your example.
it's because it's stored internally in a float variable.
in usine, for design and performance choice, all dataflow event are stored as float at some point in the engine.

it's hard to explain without deep math (search 'Single-precision floating-point' in google), but a float can represent exactly a number only if it's a power of 2.
anyway, it's an universal limitation in computer science and despite the fact it seem strange to have this king of vagueness, it work pretty well except in very specific case.
I don't know if there are internal processed doing something to the values that causes this.
Although it's more a cosmetic problem most of the time, it feels wrong ;-)
it's not an internal process but the way values are stored, maybe in other soft they do cosmetics or store values differently...
but i can assure you that the cost to resolve this is hardly compatible with efficiency in a 'real time' point of view

Statistics: Posted by martignasse — 22 Jul 2013, 10:20

]]> 2013-07-21T19:31:47+02:00 2013-07-21T19:31:47+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27300#p27300 <![CDATA[Round Vs. Trunc]]> I don't know if there are internal processed doing something to the values that causes this.
Although it's more a cosmetic problem most of the time, it feels wrong ;-)

Cheers,

Tom

Statistics: Posted by Thomas Helzle — 21 Jul 2013, 19:31

]]> 2013-07-21T17:28:48+02:00 2013-07-21T17:28:48+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27297#p27297 <![CDATA[Round Vs. Trunc]]> Left me curious about constants though...I set constant to 12.85 and in the parameters it automatically sets = 12.8500003814697

This leaves me curious.

Statistics: Posted by sephult — 21 Jul 2013, 17:28

]]> 2013-07-21T17:25:58+02:00 2013-07-21T17:25:58+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27296#p27296 <![CDATA[Round Vs. Trunc]]> I havent dug to deep in experimenting with Round/Trunc.

Always appreciated!

-S

Statistics: Posted by sephult — 21 Jul 2013, 17:25

]]> 2013-07-21T16:58:48+02:00 2013-07-21T16:58:48+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27292#p27292 <![CDATA[Round Vs. Trunc]]>

Statistics: Posted by nay-seven — 21 Jul 2013, 16:58

]]> 2013-07-21T16:32:56+02:00 2013-07-21T16:32:56+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27284#p27284 <![CDATA[Round Vs. Trunc]]>
So far I see the same results, I was expecting Trunc to allow me specify the precision at which I wanted to drop....

I think this is a feature request, or possibly if someone can fill me in on an easy method to drop precision to the place.


-S

Statistics: Posted by sephult — 21 Jul 2013, 16:32

]]> BrainModular BrainModular Users Forum 2013-07-23T00:14:58+02:00 https://www.brainmodular.com/forums/app.php/feed/topic/3971 2013-07-23T00:14:58+02:00 2013-07-23T00:14:58+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27318#p27318 <![CDATA[Round Vs. Trunc]]> Statistics: Posted by sephult — 23 Jul 2013, 00:14

]]> 2013-07-22T16:06:15+02:00 2013-07-22T16:06:15+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27315#p27315 <![CDATA[Round Vs. Trunc]]> And a reminder that what we nowadays think of as the ultimate in numeric precision is in fact the opposite:
Always slightly "wrong", even if negligible ;-)

Cheers,

Tom

Statistics: Posted by Thomas Helzle — 22 Jul 2013, 16:06

]]> 2013-07-22T13:16:42+02:00 2013-07-22T13:16:42+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27314#p27314 <![CDATA[Round Vs. Trunc]]> With this patch
You would choose between trunc or round and which decimal you want to keep (how many numbers after float).

Statistics: Posted by joffo78 — 22 Jul 2013, 13:16

]]> 2013-07-22T11:29:54+02:00 2013-07-22T11:29:54+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27313#p27313 <![CDATA[Round Vs. Trunc]]> I am guilty for not always going to the wiki, still getting used to that. I did however this time and saw....
Constant
Constant value.
out
Value of the constant
So I became very confused :lol:
Maybe we should update this to say it is stored as a Float value, and will display as such?

Thanks Thomas/Marignasse!

Statistics: Posted by sephult — 22 Jul 2013, 11:29

]]> 2013-07-22T10:20:51+02:00 2013-07-22T10:20:51+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27311#p27311 <![CDATA[Round Vs. Trunc]]>
Okay I tested and saw the expected results using a constant.
Left me curious about constants though...I set constant to 12.85 and in the parameters it automatically sets = 12.8500003814697
This leaves me curious.
Yeah, that is something rather irritating in Usine - float values change, even if very clear values are used that should normally be represented without precision problems like your example.
it's because it's stored internally in a float variable.
in usine, for design and performance choice, all dataflow event are stored as float at some point in the engine.

it's hard to explain without deep math (search 'Single-precision floating-point' in google), but a float can represent exactly a number only if it's a power of 2.
anyway, it's an universal limitation in computer science and despite the fact it seem strange to have this king of vagueness, it work pretty well except in very specific case.
I don't know if there are internal processed doing something to the values that causes this.
Although it's more a cosmetic problem most of the time, it feels wrong ;-)
it's not an internal process but the way values are stored, maybe in other soft they do cosmetics or store values differently...
but i can assure you that the cost to resolve this is hardly compatible with efficiency in a 'real time' point of view

Statistics: Posted by martignasse — 22 Jul 2013, 10:20

]]> 2013-07-21T19:31:47+02:00 2013-07-21T19:31:47+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27300#p27300 <![CDATA[Round Vs. Trunc]]> I don't know if there are internal processed doing something to the values that causes this.
Although it's more a cosmetic problem most of the time, it feels wrong ;-)

Cheers,

Tom

Statistics: Posted by Thomas Helzle — 21 Jul 2013, 19:31

]]> 2013-07-21T17:28:48+02:00 2013-07-21T17:28:48+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27297#p27297 <![CDATA[Round Vs. Trunc]]> Left me curious about constants though...I set constant to 12.85 and in the parameters it automatically sets = 12.8500003814697

This leaves me curious.

Statistics: Posted by sephult — 21 Jul 2013, 17:28

]]> 2013-07-21T17:25:58+02:00 2013-07-21T17:25:58+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27296#p27296 <![CDATA[Round Vs. Trunc]]> I havent dug to deep in experimenting with Round/Trunc.

Always appreciated!

-S

Statistics: Posted by sephult — 21 Jul 2013, 17:25

]]> 2013-07-21T16:58:48+02:00 2013-07-21T16:58:48+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27292#p27292 <![CDATA[Round Vs. Trunc]]>

Statistics: Posted by nay-seven — 21 Jul 2013, 16:58

]]> 2013-07-21T16:32:56+02:00 2013-07-21T16:32:56+02:00 https://www.brainmodular.com/forums/viewtopic.php?t=3971&p=27284#p27284 <![CDATA[Round Vs. Trunc]]>
So far I see the same results, I was expecting Trunc to allow me specify the precision at which I wanted to drop....

I think this is a feature request, or possibly if someone can fill me in on an easy method to drop precision to the place.


-S

Statistics: Posted by sephult — 21 Jul 2013, 16:32

]]>