You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@commons.apache.org by Gilles Sadowski <gi...@gmail.com> on 2020/04/06 17:24:51 UTC
[Geometry][Math] Equivalent of CM's "RotationConvention"
Hello.
Is there a way to replace an argument of CM's type
"RotationConvention"[1] with a type from [Geometry]?
Thanks,
Gilles
[1] https://gitbox.apache.org/repos/asf?p=commons-math.git;a=blob;f=src/main/java/org/apache/commons/math4/geometry/euclidean/threed/RotationConvention.java;h=a80fe9b553d023d4054fd17f18f02b64ca7cdd34;hb=HEAD
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [Geometry][Math] Equivalent of CM's "RotationConvention"
Posted by Matt Juntunen <ma...@hotmail.com>.
I agree. Let's remove it.
-Matt
________________________________
From: Gilles Sadowski <gi...@gmail.com>
Sent: Thursday, April 9, 2020 5:30 AM
To: Commons Developers List <de...@commons.apache.org>
Subject: Re: [Geometry][Math] Equivalent of CM's "RotationConvention"
Hi.
2020-04-09 3:30 UTC+02:00, Matt Juntunen <ma...@hotmail.com>:
> It should just be a matter of removing all RotationConvention arguments and
> replacing any branches on the rotation convention with the path that uses
> the non-negated angle. Then, documenting the convention followed by the
> class.
Rereading the discussion[1] that led to the introduction of
"RotationConvention", I now find that the argument was not
very compelling ("some people do not like this convention").
And it seems inadequate and confusing to embed the choice
in one type of transforms, rather than having a separate
concept of "coordinate transform".
However, if the "feature" is just removed, it make it even
more unlikely that people will adapt their code just to use
the version of that class that would ship with CM v4.0.
IMO, they will stick with CM 3.6.1; hence adapting the few
classes that remain in the "o.a.c.m.geometry" package
seems like a pointless exercise. I'd just remove it all (cf.
other thread).
WDYT?
Gilles
[1] https://issues.apache.org/jira/browse/MATH-1302
>>> [...]
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [Geometry][Math] Equivalent of CM's "RotationConvention"
Posted by Gilles Sadowski <gi...@gmail.com>.
Hi.
2020-04-09 3:30 UTC+02:00, Matt Juntunen <ma...@hotmail.com>:
> It should just be a matter of removing all RotationConvention arguments and
> replacing any branches on the rotation convention with the path that uses
> the non-negated angle. Then, documenting the convention followed by the
> class.
Rereading the discussion[1] that led to the introduction of
"RotationConvention", I now find that the argument was not
very compelling ("some people do not like this convention").
And it seems inadequate and confusing to embed the choice
in one type of transforms, rather than having a separate
concept of "coordinate transform".
However, if the "feature" is just removed, it make it even
more unlikely that people will adapt their code just to use
the version of that class that would ship with CM v4.0.
IMO, they will stick with CM 3.6.1; hence adapting the few
classes that remain in the "o.a.c.m.geometry" package
seems like a pointless exercise. I'd just remove it all (cf.
other thread).
WDYT?
Gilles
[1] https://issues.apache.org/jira/browse/MATH-1302
>>> [...]
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [Geometry][Math] Equivalent of CM's "RotationConvention"
Posted by Matt Juntunen <ma...@hotmail.com>.
It should just be a matter of removing all RotationConvention arguments and replacing any branches on the rotation convention with the path that uses the non-negated angle. Then, documenting the convention followed by the class.
-Matt
________________________________
From: Gilles Sadowski <gi...@gmail.com>
Sent: Wednesday, April 8, 2020 9:20 AM
To: Commons Developers List <de...@commons.apache.org>
Subject: Re: [Geometry][Math] Equivalent of CM's "RotationConvention"
Hi.
>
> I would encourage removing RotationConvention without replacing it. But, I still don't think I'm quite getting what you're saying. Do you have an example?
I wonder how to replace "RotationConvention" usage in class "FieldRotation".[1]
Regards,
Gilles
[1] https://gitbox.apache.org/repos/asf?p=commons-math.git;a=blob;f=src/main/java/org/apache/commons/math4/geometry/euclidean/threed/FieldRotation.java;h=e671acf555e31564cc2049467014ab4f873a48e7;hb=HEAD
>
> -Matt
> ________________________________
> From: Gilles Sadowski <gi...@gmail.com>
> Sent: Tuesday, April 7, 2020 12:32 PM
> To: Commons Developers List <de...@commons.apache.org>
> Subject: Re: [Geometry][Math] Equivalent of CM's "RotationConvention"
>
> Hello.
>
> Le mar. 7 avr. 2020 à 03:18, Matt Juntunen <ma...@hotmail.com> a écrit :
> >
> > Hi Gilles,
> >
> > I removed the RotationConvention enum from commons-geometry because it was more confusing than useful. Instead, I specified [1] that commons-geometry uses "active rotations" by convention (equivalent to RotationConvention.VECTOR_OPERATOR). If users want to use the equivalent of RotationConvention.FRAME_TRANSFORM, they simply need to negate the angle used as input when creating the rotation quaternions. This is actually all that the previous commons-math Rotation class did with this enum.
>
> My question was ambiguous: What I'm after is removing
> "RotationConvention" from Commons Math and replace
> it (by <something>?) where it is used within CM.
> Possible?
>
> Thanks,
> Gilles
>
> >
> > Regards,
> > Matt J
> >
> > [1] https://github.com/apache/commons-geometry/blob/master/commons-geometry-euclidean/src/main/java/org/apache/commons/geometry/euclidean/threed/rotation/package-info.java
> >
> > ________________________________
> > From: Gilles Sadowski <gi...@gmail.com>
> > Sent: Monday, April 6, 2020 1:24 PM
> > To: Commons Developers List <de...@commons.apache.org>
> > Subject: [Geometry][Math] Equivalent of CM's "RotationConvention"
> >
> > Hello.
> >
> > Is there a way to replace an argument of CM's type
> > "RotationConvention"[1] with a type from [Geometry]?
> >
> > Thanks,
> > Gilles
> >
> > [1] https://gitbox.apache.org/repos/asf?p=commons-math.git;a=blob;f=src/main/java/org/apache/commons/math4/geometry/euclidean/threed/RotationConvention.java;h=a80fe9b553d023d4054fd17f18f02b64ca7cdd34;hb=HEAD
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [Geometry][Math] Equivalent of CM's "RotationConvention"
Posted by Gilles Sadowski <gi...@gmail.com>.
Hi.
>
> I would encourage removing RotationConvention without replacing it. But, I still don't think I'm quite getting what you're saying. Do you have an example?
I wonder how to replace "RotationConvention" usage in class "FieldRotation".[1]
Regards,
Gilles
[1] https://gitbox.apache.org/repos/asf?p=commons-math.git;a=blob;f=src/main/java/org/apache/commons/math4/geometry/euclidean/threed/FieldRotation.java;h=e671acf555e31564cc2049467014ab4f873a48e7;hb=HEAD
>
> -Matt
> ________________________________
> From: Gilles Sadowski <gi...@gmail.com>
> Sent: Tuesday, April 7, 2020 12:32 PM
> To: Commons Developers List <de...@commons.apache.org>
> Subject: Re: [Geometry][Math] Equivalent of CM's "RotationConvention"
>
> Hello.
>
> Le mar. 7 avr. 2020 à 03:18, Matt Juntunen <ma...@hotmail.com> a écrit :
> >
> > Hi Gilles,
> >
> > I removed the RotationConvention enum from commons-geometry because it was more confusing than useful. Instead, I specified [1] that commons-geometry uses "active rotations" by convention (equivalent to RotationConvention.VECTOR_OPERATOR). If users want to use the equivalent of RotationConvention.FRAME_TRANSFORM, they simply need to negate the angle used as input when creating the rotation quaternions. This is actually all that the previous commons-math Rotation class did with this enum.
>
> My question was ambiguous: What I'm after is removing
> "RotationConvention" from Commons Math and replace
> it (by <something>?) where it is used within CM.
> Possible?
>
> Thanks,
> Gilles
>
> >
> > Regards,
> > Matt J
> >
> > [1] https://github.com/apache/commons-geometry/blob/master/commons-geometry-euclidean/src/main/java/org/apache/commons/geometry/euclidean/threed/rotation/package-info.java
> >
> > ________________________________
> > From: Gilles Sadowski <gi...@gmail.com>
> > Sent: Monday, April 6, 2020 1:24 PM
> > To: Commons Developers List <de...@commons.apache.org>
> > Subject: [Geometry][Math] Equivalent of CM's "RotationConvention"
> >
> > Hello.
> >
> > Is there a way to replace an argument of CM's type
> > "RotationConvention"[1] with a type from [Geometry]?
> >
> > Thanks,
> > Gilles
> >
> > [1] https://gitbox.apache.org/repos/asf?p=commons-math.git;a=blob;f=src/main/java/org/apache/commons/math4/geometry/euclidean/threed/RotationConvention.java;h=a80fe9b553d023d4054fd17f18f02b64ca7cdd34;hb=HEAD
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [Geometry][Math] Equivalent of CM's "RotationConvention"
Posted by Matt Juntunen <ma...@hotmail.com>.
Gilles,
I would encourage removing RotationConvention without replacing it. But, I still don't think I'm quite getting what you're saying. Do you have an example?
-Matt
________________________________
From: Gilles Sadowski <gi...@gmail.com>
Sent: Tuesday, April 7, 2020 12:32 PM
To: Commons Developers List <de...@commons.apache.org>
Subject: Re: [Geometry][Math] Equivalent of CM's "RotationConvention"
Hello.
Le mar. 7 avr. 2020 à 03:18, Matt Juntunen <ma...@hotmail.com> a écrit :
>
> Hi Gilles,
>
> I removed the RotationConvention enum from commons-geometry because it was more confusing than useful. Instead, I specified [1] that commons-geometry uses "active rotations" by convention (equivalent to RotationConvention.VECTOR_OPERATOR). If users want to use the equivalent of RotationConvention.FRAME_TRANSFORM, they simply need to negate the angle used as input when creating the rotation quaternions. This is actually all that the previous commons-math Rotation class did with this enum.
My question was ambiguous: What I'm after is removing
"RotationConvention" from Commons Math and replace
it (by <something>?) where it is used within CM.
Possible?
Thanks,
Gilles
>
> Regards,
> Matt J
>
> [1] https://github.com/apache/commons-geometry/blob/master/commons-geometry-euclidean/src/main/java/org/apache/commons/geometry/euclidean/threed/rotation/package-info.java
>
> ________________________________
> From: Gilles Sadowski <gi...@gmail.com>
> Sent: Monday, April 6, 2020 1:24 PM
> To: Commons Developers List <de...@commons.apache.org>
> Subject: [Geometry][Math] Equivalent of CM's "RotationConvention"
>
> Hello.
>
> Is there a way to replace an argument of CM's type
> "RotationConvention"[1] with a type from [Geometry]?
>
> Thanks,
> Gilles
>
> [1] https://gitbox.apache.org/repos/asf?p=commons-math.git;a=blob;f=src/main/java/org/apache/commons/math4/geometry/euclidean/threed/RotationConvention.java;h=a80fe9b553d023d4054fd17f18f02b64ca7cdd34;hb=HEAD
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [Geometry][Math] Equivalent of CM's "RotationConvention"
Posted by Gilles Sadowski <gi...@gmail.com>.
Hello.
Le mar. 7 avr. 2020 à 03:18, Matt Juntunen <ma...@hotmail.com> a écrit :
>
> Hi Gilles,
>
> I removed the RotationConvention enum from commons-geometry because it was more confusing than useful. Instead, I specified [1] that commons-geometry uses "active rotations" by convention (equivalent to RotationConvention.VECTOR_OPERATOR). If users want to use the equivalent of RotationConvention.FRAME_TRANSFORM, they simply need to negate the angle used as input when creating the rotation quaternions. This is actually all that the previous commons-math Rotation class did with this enum.
My question was ambiguous: What I'm after is removing
"RotationConvention" from Commons Math and replace
it (by <something>?) where it is used within CM.
Possible?
Thanks,
Gilles
>
> Regards,
> Matt J
>
> [1] https://github.com/apache/commons-geometry/blob/master/commons-geometry-euclidean/src/main/java/org/apache/commons/geometry/euclidean/threed/rotation/package-info.java
>
> ________________________________
> From: Gilles Sadowski <gi...@gmail.com>
> Sent: Monday, April 6, 2020 1:24 PM
> To: Commons Developers List <de...@commons.apache.org>
> Subject: [Geometry][Math] Equivalent of CM's "RotationConvention"
>
> Hello.
>
> Is there a way to replace an argument of CM's type
> "RotationConvention"[1] with a type from [Geometry]?
>
> Thanks,
> Gilles
>
> [1] https://gitbox.apache.org/repos/asf?p=commons-math.git;a=blob;f=src/main/java/org/apache/commons/math4/geometry/euclidean/threed/RotationConvention.java;h=a80fe9b553d023d4054fd17f18f02b64ca7cdd34;hb=HEAD
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [Geometry][Math] Equivalent of CM's "RotationConvention"
Posted by Matt Juntunen <ma...@hotmail.com>.
Hi Gilles,
I removed the RotationConvention enum from commons-geometry because it was more confusing than useful. Instead, I specified [1] that commons-geometry uses "active rotations" by convention (equivalent to RotationConvention.VECTOR_OPERATOR). If users want to use the equivalent of RotationConvention.FRAME_TRANSFORM, they simply need to negate the angle used as input when creating the rotation quaternions. This is actually all that the previous commons-math Rotation class did with this enum.
Regards,
Matt J
[1] https://github.com/apache/commons-geometry/blob/master/commons-geometry-euclidean/src/main/java/org/apache/commons/geometry/euclidean/threed/rotation/package-info.java
________________________________
From: Gilles Sadowski <gi...@gmail.com>
Sent: Monday, April 6, 2020 1:24 PM
To: Commons Developers List <de...@commons.apache.org>
Subject: [Geometry][Math] Equivalent of CM's "RotationConvention"
Hello.
Is there a way to replace an argument of CM's type
"RotationConvention"[1] with a type from [Geometry]?
Thanks,
Gilles
[1] https://gitbox.apache.org/repos/asf?p=commons-math.git;a=blob;f=src/main/java/org/apache/commons/math4/geometry/euclidean/threed/RotationConvention.java;h=a80fe9b553d023d4054fd17f18f02b64ca7cdd34;hb=HEAD
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org