CanOpen and Orocos

Hi all,

Does someone has already interfaced Orocos with some Can drivers ? CanOpen
stacks ? Are these components publics ?

If the anwer is no, is there anything planned in Orocos to share user
components ? If I develop this and whant to share do I have a mean or should
I do it on my own ?
I know these kind of components (I mean Hardware) shall not be include in
the official Orocos releases, but isn't a good thing for Orocos to create a
place to hold it so that Apllication Builder may look in it before choosing
or not Orocos as their framework ?

This question is quite related the the recent Comedi discussions on the
mailing list.

CanOpen and orocos

> Hi all,
>
> Does someone has already interfaced Orocos with some Can drivers ? CanOpen
> stacks ? Are these components publics ?

Do you mean CAN or CanOpen?
WRT CAN, this is supported (see OCL api documentation)
WRT CanOPEN, we've done this a long time ago using the CANFestival.
It worked, but it was a hacky solution. AFAIK the code never got
public.

Klaas

CanOpen and orocos

Hi, I just had a look in the ocl 1.12 directory. All this stuff has
effectively disappear in 2.X.

I am very interested in having the CanFestival integration. Do you know
where I can found it ?

2011/2/17 Klaas Gadeyne <klaas [dot] gadeyne [..] ...>

> > Hi all,
> >
> > Does someone has already interfaced Orocos with some Can drivers ?
> CanOpen
> > stacks ? Are these components publics ?
>
> Do you mean CAN or CanOpen?
> WRT CAN, this is supported (see OCL api documentation)
> WRT CanOPEN, we've done this a long time ago using the CANFestival.
> It worked, but it was a hacky solution. AFAIK the code never got
> public.
>
> Klaas
> --
> Orocos-Users mailing list
> Orocos-Users [..] ...
> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
>

CanOpen and orocos

[...]
> Hi, I just had a look in the ocl 1.12 directory. All this stuff has
> effectively disappear in 2.X.
>
> I am very interested in having the CanFestival integration. Do you know
> where I can found it ?

[cfr earlier remarks]
AFAIR, it is in a private repository somewhere at this company.

Regards,

Klaas

CanOpen and orocos

On Fri, Feb 18, 2011 at 6:27 PM, Willy Lambert <lambert [dot] willy [..] ...>wrote:

> Hi, I just had a look in the ocl 1.12 directory. All this stuff has
> effectively disappear in 2.X.
>
> I am very interested in having the CanFestival integration. Do you know
> where I can found it ?
>

Willy,

There was an interesting thread on ros-users recently related to CanOpen,
and CanFestival in particular. You might be interested in giving it a read.

https://code.ros.org/lurker/message/20101201.093150.e833cb68.en.html

Cheers,

Adolfo.

> 2011/2/17 Klaas Gadeyne <klaas [dot] gadeyne [..] ...>
>
>> > Hi all,
>>
>> >
>> > Does someone has already interfaced Orocos with some Can drivers ?
>> CanOpen
>> > stacks ? Are these components publics ?
>>
>> Do you mean CAN or CanOpen?
>> WRT CAN, this is supported (see OCL api documentation)
>> WRT CanOPEN, we've done this a long time ago using the CANFestival.
>> It worked, but it was a hacky solution. AFAIK the code never got
>> public.
>>
>> Klaas
>> --
>> Orocos-Users mailing list
>> Orocos-Users [..] ...
>> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
>>
>
>
> --
> Orocos-Users mailing list
> Orocos-Users [..] ...
> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
>
>

CanOpen and orocos

Sure, thanks for it.

But I am still interesed in the CanFestival Orocos component, even if it is
not in a working state, my aim is to have an Orocos component

2011/2/21 Adolfo Rodríguez Tsouroukdissian <
adolfo [dot] rodriguez [..] ...>

>
>
> On Fri, Feb 18, 2011 at 6:27 PM, Willy Lambert <lambert [dot] willy [..] ...>wrote:
>
>> Hi, I just had a look in the ocl 1.12 directory. All this stuff has
>> effectively disappear in 2.X.
>>
>> I am very interested in having the CanFestival integration. Do you know
>> where I can found it ?
>>
>
> Willy,
>
> There was an interesting thread on ros-users recently related to CanOpen,
> and CanFestival in particular. You might be interested in giving it a read.
>
> https://code.ros.org/lurker/message/20101201.093150.e833cb68.en.html
>
> Cheers,
>
> Adolfo.
>
>
>> 2011/2/17 Klaas Gadeyne <klaas [dot] gadeyne [..] ...>
>>
>>> > Hi all,
>>>
>>> >
>>> > Does someone has already interfaced Orocos with some Can drivers ?
>>> CanOpen
>>> > stacks ? Are these components publics ?
>>>
>>> Do you mean CAN or CanOpen?
>>> WRT CAN, this is supported (see OCL api documentation)
>>> WRT CanOPEN, we've done this a long time ago using the CANFestival.
>>> It worked, but it was a hacky solution. AFAIK the code never got
>>> public.
>>>
>>> Klaas
>>> --
>>> Orocos-Users mailing list
>>> Orocos-Users [..] ...
>>> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
>>>
>>
>>
>> --
>> Orocos-Users mailing list
>> Orocos-Users [..] ...
>> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
>>
>>
>
>
> --
> Adolfo Rodríguez Tsouroukdissian
>
> Robotics engineer
> PAL ROBOTICS S.L
> http://www.pal-robotics.com
> Tel. +34.93.414.53.47
> Fax.+34.93.209.11.09
>
> CONFIDENTIALITY NOTICE: This e-mail and the accompanying document(s) may
> contain confidential information which is privileged and intended only for
> the individual or entity to whom they are addressed. If you are not the
> intended recipient, you are hereby notified that any disclosure, copying,
> distribution or use of this e-mail and/or accompanying document(s) is
> strictly prohibited. If you have received this e-mail in error, please
> immediately notify the sender at the above e-mail address.
>

CanOpen and orocos

2011/2/17 Klaas Gadeyne <klaas [dot] gadeyne [..] ...>

> > Hi all,
> >
> > Does someone has already interfaced Orocos with some Can drivers ?
> CanOpen
> > stacks ? Are these components publics ?
>
> Do you mean CAN or CanOpen?
>
If the choice exists, of course CanOpen

> WRT CAN, this is supported (see OCL api documentation)
>

Is true in 2.x series ? Is it still the aim of OCL to propose such
libraries ? I thougth OCL was kind of integrated to Orocos toolchain and
that it is not the aim of OCL to propose such "user libraries" (to be
honest, I do not really know what are the mainline for ocl)
in a more general manner, I have difficulties to have a clear view about how
users may share their code to the Orocos community

> WRT CanOPEN, we've done this a long time ago using the CANFestival.
> It worked, but it was a hacky solution.

Perfect, I am trying to interface CanFestival on my robot, was it a question
of time or due to technical problem ?

> AFAIK the code never got
> public.
>

Is it a lack of maintainer or time, or a will to keep it not used too much ?

>
> Klaas
> --
> Orocos-Users mailing list
> Orocos-Users [..] ...
> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
>

CanOpen and orocos

> 2011/2/17 Klaas Gadeyne <klaas [dot] gadeyne [..] ...>
[...]
>> WRT CAN, this is supported (see OCL api documentation)
>>
>
> Is true in 2.x series ? Is it still the aim of  OCL to propose such
> libraries ? I thougth OCL was kind of integrated to Orocos toolchain and
> that it is not the aim of OCL to propose such "user libraries" (to be
> honest, I do not really know what are the mainline for ocl)
> in a more general manner, I have difficulties to have a clear view about how
> users may share their code to the Orocos community

Honestly, I wouldn't know. my statements concerned the 1.x series. I
have not followed the 2.x developments closely, and I have no clue
what the status in 2.x is (I cannot recall any update message saying
that x and y were no longer supported)

>> WRT CanOPEN, we've done this a long time ago using the CANFestival.
>> It worked, but it was a hacky solution.
>
> Perfect, I am trying to interface CanFestival on my robot, was it a question
> of time or due to technical problem ?

I didn' t develop the code myself, but it was a kind of demonstrator
prototype. It also was a bit a special case, in which we tried to
emulate the behaviour of multiple CANOpen slaves as part of a single
program, and IIRC this complicated stuff. If you would just be
interested in implementing a CANOpen master, I guess there will be no
problems at all (cfr. EtherCAT master implementations which have been
done in the past)

>> AFAIK the code never got
>> public.
>
> Is it a lack of maintainer or time, or a will to keep it not used too much ?

As described above, it was a kind of "special case" and therefor not
generally applicable (without major effort), but the project was at
some point also discontinued (for other reasons), and hence a lack of
(maintainer) time is certainly another reason why it was never
published.

HTH,

Klaas
--
Orocos-Users mailing list
Orocos-Users [..] ...
http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users

CanOpen and orocos

2011/2/17 Klaas Gadeyne <klaas [dot] gadeyne [..] ...>

> > 2011/2/17 Klaas Gadeyne <klaas [dot] gadeyne [..] ...>
> [...]
> >> WRT CAN, this is supported (see OCL api documentation)
> >>
> >
> > Is true in 2.x series ? Is it still the aim of OCL to propose such
> > libraries ? I thougth OCL was kind of integrated to Orocos toolchain and
> > that it is not the aim of OCL to propose such "user libraries" (to be
> > honest, I do not really know what are the mainline for ocl)
> > in a more general manner, I have difficulties to have a clear view about
> how
> > users may share their code to the Orocos community
>
> Honestly, I wouldn't know. my statements concerned the 1.x series. I
> have not followed the 2.x developments closely, and I have no clue
> what the status in 2.x is (I cannot recall any update message saying
> that x and y were no longer supported)
>
> >> WRT CanOPEN, we've done this a long time ago using the CANFestival.
> >> It worked, but it was a hacky solution.
> >
> > Perfect, I am trying to interface CanFestival on my robot, was it a
> question
> > of time or due to technical problem ?
>
> I didn' t develop the code myself, but it was a kind of demonstrator
> prototype. It also was a bit a special case, in which we tried to
> emulate the behaviour of multiple CANOpen slaves as part of a single
> program, and IIRC this complicated stuff. If you would just be
> interested in implementing a CANOpen master, I guess there will be no
> problems at all (cfr. EtherCAT master implementations which have been
> done in the past)
>
> >> AFAIK the code never got
> >> public.
> >
> > Is it a lack of maintainer or time, or a will to keep it not used too
> much ?
>
> As described above, it was a kind of "special case" and therefor not
> generally applicable (without major effort), but the project was at
> some point also discontinued (for other reasons), and hence a lack of
> (maintainer) time is certainly another reason why it was never
> published.
>

thanks for all theses precisions.

May any RTT developper tell me about where is the place for such a component
? is it the job of OCL v2 ?

>
> HTH,
>
> Klaas
> --
> Orocos-Users mailing list
> Orocos-Users [..] ...
> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
>

CanOpen and orocos

On Thursday 17 February 2011 17:29:28 Willy Lambert wrote:
> 2011/2/17 Klaas Gadeyne <klaas [dot] gadeyne [..] ...>
>
> > > 2011/2/17 Klaas Gadeyne <klaas [dot] gadeyne [..] ...>
> >
> > [...]
> >
> > >> WRT CAN, this is supported (see OCL api documentation)
> > >
> > > Is true in 2.x series ? Is it still the aim of OCL to propose such
> > > libraries ? I thougth OCL was kind of integrated to Orocos toolchain
> > > and that it is not the aim of OCL to propose such "user libraries" (to
> > > be honest, I do not really know what are the mainline for ocl)
> > > in a more general manner, I have difficulties to have a clear view
> > > about
> >
> > how
> >
> > > users may share their code to the Orocos community
> >
> > Honestly, I wouldn't know. my statements concerned the 1.x series. I
> > have not followed the 2.x developments closely, and I have no clue
> > what the status in 2.x is (I cannot recall any update message saying
> > that x and y were no longer supported)
> >
> > >> WRT CanOPEN, we've done this a long time ago using the CANFestival.
> > >> It worked, but it was a hacky solution.
> > >
> > > Perfect, I am trying to interface CanFestival on my robot, was it a
> >
> > question
> >
> > > of time or due to technical problem ?
> >
> > I didn' t develop the code myself, but it was a kind of demonstrator
> > prototype. It also was a bit a special case, in which we tried to
> > emulate the behaviour of multiple CANOpen slaves as part of a single
> > program, and IIRC this complicated stuff. If you would just be
> > interested in implementing a CANOpen master, I guess there will be no
> > problems at all (cfr. EtherCAT master implementations which have been
> > done in the past)
> >
> > >> AFAIK the code never got
> > >> public.
> > >
> > > Is it a lack of maintainer or time, or a will to keep it not used too
> >
> > much ?
> >
> > As described above, it was a kind of "special case" and therefor not
> > generally applicable (without major effort), but the project was at
> > some point also discontinued (for other reasons), and hence a lack of
> > (maintainer) time is certainly another reason why it was never
> > published.
>
> thanks for all theses precisions.
>
> May any RTT developper tell me about where is the place for such a
> component ? is it the job of OCL v2 ?

No. Create a package repository similar to how a ROS package tree works (ie
identical :-) , and point people to that repository. You can use git or svn,
or whatever you feel most comfortable with.

Peter

CanOpen and orocos

2011/2/22 Peter Soetens <peter [..] ...>

> On Thursday 17 February 2011 17:29:28 Willy Lambert wrote:
> > 2011/2/17 Klaas Gadeyne <klaas [dot] gadeyne [..] ...>
> >
> > > > 2011/2/17 Klaas Gadeyne <klaas [dot] gadeyne [..] ...>
> > >
> > > [...]
> > >
> > > >> WRT CAN, this is supported (see OCL api documentation)
> > > >
> > > > Is true in 2.x series ? Is it still the aim of OCL to propose such
> > > > libraries ? I thougth OCL was kind of integrated to Orocos toolchain
> > > > and that it is not the aim of OCL to propose such "user libraries"
> (to
> > > > be honest, I do not really know what are the mainline for ocl)
> > > > in a more general manner, I have difficulties to have a clear view
> > > > about
> > >
> > > how
> > >
> > > > users may share their code to the Orocos community
> > >
> > > Honestly, I wouldn't know. my statements concerned the 1.x series. I
> > > have not followed the 2.x developments closely, and I have no clue
> > > what the status in 2.x is (I cannot recall any update message saying
> > > that x and y were no longer supported)
> > >
> > > >> WRT CanOPEN, we've done this a long time ago using the CANFestival.
> > > >> It worked, but it was a hacky solution.
> > > >
> > > > Perfect, I am trying to interface CanFestival on my robot, was it a
> > >
> > > question
> > >
> > > > of time or due to technical problem ?
> > >
> > > I didn' t develop the code myself, but it was a kind of demonstrator
> > > prototype. It also was a bit a special case, in which we tried to
> > > emulate the behaviour of multiple CANOpen slaves as part of a single
> > > program, and IIRC this complicated stuff. If you would just be
> > > interested in implementing a CANOpen master, I guess there will be no
> > > problems at all (cfr. EtherCAT master implementations which have been
> > > done in the past)
> > >
> > > >> AFAIK the code never got
> > > >> public.
> > > >
> > > > Is it a lack of maintainer or time, or a will to keep it not used too
> > >
> > > much ?
> > >
> > > As described above, it was a kind of "special case" and therefor not
> > > generally applicable (without major effort), but the project was at
> > > some point also discontinued (for other reasons), and hence a lack of
> > > (maintainer) time is certainly another reason why it was never
> > > published.
> >
> > thanks for all theses precisions.
> >
> > May any RTT developper tell me about where is the place for such a
> > component ? is it the job of OCL v2 ?
>
> No. Create a package repository similar to how a ROS package tree works (ie
> identical :-) , and point people to that repository. You can use git or
> svn,
> or whatever you feel most comfortable with.
>

Peter
>

I found this in Rock :
http://rock-robotics.org/package_directory/packages/drivers_orogen_canbu...

is it a usable Orocos component (I mean out of rock) ?

Orocos and canfestival

Hi all, I need to use a opencan DS 402 profile with some elmo boards.

I like canfestival, is there something already done in Orocos?

Regards, Fabrizio

Orocos and canfestival

2013/1/23 <fabrizio [dot] boriero [..] ...>:
> Hi all,
> I need to use a opencan DS 402 profile with some elmo boards.
>
> I like canfestival, is there something already done in Orocos?
>

Yes and no.
Yes, because I think we are several people to play with CanFestival and Orocos.

No because none of this stuff is public AFAIK. On my own, it is not
public because it is not clean and genric enougth to be public. But
basically what I have done is to create standalone ports (I mean out
of a component) that are fitted in Canfestival Callbacks. I have A
CanManager Component that is doing the "server" work and one component
by Can device.

I'm quite sure we are several here interested in a nice CanFestival
integration. I'm personnaly very interested in this, but I have other
plates boiling. Nevertheless, even if CanFestival do the work, I don't
find that this framework is very clean, living, ergonomic, up to date,
evolutive. I'm split into puting effort to rework all this in C++ or
waiting that Can is replaced by other bus with higher bandwidth (in
short everything that is playing with Ethernet wiring).

That said, my robot is perfectly working with it for now.

> Regards,
> Fabrizio
> --
> Orocos-Users mailing list
> Orocos-Users [..] ...
> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users

Orocos and canfestival

FYI.

Some months ago I needed a CANOpen master for a project and frustrated with
the API of CanFestival and its (apparent?) limitations in terms of
flexibility, I decided to write mine from scratch.

I know, I know... reinventing the f***ing wheel. But I couldn't find on
CanFestival a way to have different ObjectDictionaries, i.e. to control
different brands of motors (for example Elmo + Maxon).

The good news is that my current version work and it is way more
understandable and flexible.

I am currently working hard to make a commercial software out of it, maybe
with a double licensing model (GPL for the community + commercial )... I
don't know.
Integration with Orocos would be straight forward through a thin wrapper.
Actually I designed an extra layer that I called Motor Abstract Layer, that
make possible to control the behaviour of the motor without the need to
learn the CANopen standard, i.e. you talk about position, velocities,
torques, etc., and you don't have to deal directly with NTM, SDO, PDO, SYNC
and all that stuff, unless you want to (you can still do it through an
additional API).

If there is anyone that want to be a use/test it, let me know. I should be
able to release it in a couple of weeks.

Regards

Davide

On Wed, Jan 23, 2013 at 7:11 PM, Willy Lambert <lambert [dot] willy [..] ...>wrote:

> 2013/1/23 <fabrizio [dot] boriero [..] ...>:
> > Hi all,
> > I need to use a opencan DS 402 profile with some elmo boards.
> >
> > I like canfestival, is there something already done in Orocos?
> >
>
> Yes and no.
> Yes, because I think we are several people to play with CanFestival and
> Orocos.
>
> No because none of this stuff is public AFAIK. On my own, it is not
> public because it is not clean and genric enougth to be public. But
> basically what I have done is to create standalone ports (I mean out
> of a component) that are fitted in Canfestival Callbacks. I have A
> CanManager Component that is doing the "server" work and one component
> by Can device.
>
> I'm quite sure we are several here interested in a nice CanFestival
> integration. I'm personnaly very interested in this, but I have other
> plates boiling. Nevertheless, even if CanFestival do the work, I don't
> find that this framework is very clean, living, ergonomic, up to date,
> evolutive. I'm split into puting effort to rework all this in C++ or
> waiting that Can is replaced by other bus with higher bandwidth (in
> short everything that is playing with Ethernet wiring).
>
> That said, my robot is perfectly working with it for now.
>
>
>
>
> > Regards,
> > Fabrizio
> > --
> > Orocos-Users mailing list
> > Orocos-Users [..] ...
> > http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
> --
> Orocos-Users mailing list
> Orocos-Users [..] ...
> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
>

Orocos and canfestival

On Thu, 24 Jan 2013, Davide Faconti wrote:

> FYI.
> Some months ago I needed a CANOpen master for a project and frustrated with the API of
> CanFestival and its (apparent?) limitations in terms of flexibility, I decided to write
> mine from scratch.
>
> I know, I know... reinventing the f***ing wheel. But I couldn't find on CanFestival a
> way to have different ObjectDictionaries, i.e. to control different brands of motors
> (for example Elmo + Maxon).
>
> The good news is that my current version work and it is way more understandable and
> flexible.
>
> I am currently working hard to make a commercial software out of it, maybe with a
> double licensing model (GPL for the community + commercial )... I don't know.
> Integration with Orocos would be straight forward through a thin wrapper. 
> Actually I designed an extra layer that I called Motor Abstract Layer, that make
> possible to control the behaviour of the motor without the need to learn the CANopen
> standard, i.e. you talk about position, velocities, torques, etc., and you don't have
> to deal directly with NTM, SDO, PDO, SYNC and all that stuff, unless you want to (you
> can still do it through an additional API).

This "layer" is worth being "standardized", I think, because it is useful
in all motion contrl applications, and not just in the context of CAN :-)

> If there is anyone that want to be a use/test it, let me know. I should be able to
> release it in a couple of weeks.

I am definitely interested in discussing the Motor Abstract Layer!

> Regards
>
> Davide

Herman

Orocos and canfestival

2013/1/25 Herman Bruyninckx <Herman [dot] Bruyninckx [..] ...>:
> On Thu, 24 Jan 2013, Davide Faconti wrote:
>
>> FYI.
>> Some months ago I needed a CANOpen master for a project and frustrated
>> with the API of
>> CanFestival and its (apparent?) limitations in terms of flexibility, I
>> decided to write
>> mine from scratch.
>>
>> I know, I know... reinventing the f***ing wheel. But I couldn't find on
>> CanFestival a
>> way to have different ObjectDictionaries, i.e. to control different brands
>> of motors
>> (for example Elmo + Maxon).
>>
>> The good news is that my current version work and it is way more
>> understandable and
>> flexible.
>>
>> I am currently working hard to make a commercial software out of it, maybe
>> with a
>> double licensing model (GPL for the community + commercial )... I don't
>> know.
>> Integration with Orocos would be straight forward through a thin wrapper.
>> Actually I designed an extra layer that I called Motor Abstract Layer,
>> that make
>> possible to control the behaviour of the motor without the need to learn
>> the CANopen
>> standard, i.e. you talk about position, velocities, torques, etc., and you
>> don't have
>> to deal directly with NTM, SDO, PDO, SYNC and all that stuff, unless you
>> want to (you
>> can still do it through an additional API).
>
>
> This "layer" is worth being "standardized", I think, because it is useful
> in all motion contrl applications, and not just in the context of CAN :-)
>

agree

>
>> If there is anyone that want to be a use/test it, let me know. I should be
>> able to
>> release it in a couple of weeks.
>
>
> I am definitely interested in discussing the Motor Abstract Layer!
>

I will be a pleasure to do that, please come
http://sourceforge.net/projects/libds402 here to chat avout this. Even
if the project is related to CAN now, we will do cleaning later when
things will be clear.

>> Regards
>>
>> Davide
>
>
> Herman

Orocos and canfestival

On Fri, 25 Jan 2013, Willy Lambert wrote:

[...]
>>> Actually I designed an extra layer that I called Motor Abstract Layer,
>>> that make possible to control the behaviour of the motor without the
>>> need to learn the CANopen standard, i.e. you talk about position,
>>> velocities, torques, etc., and you don't have to deal directly with
>>> NTM, SDO, PDO, SYNC and all that stuff, unless you want to (you can
>>> still do it through an additional API).
>>
>> This "layer" is worth being "standardized", I think, because it is useful
>> in all motion contrl applications, and not just in the context of CAN :-)
>
> agree
>
>>> If there is anyone that want to be a use/test it, let me know. I should be
>>> able to release it in a couple of weeks.
>>
>> I am definitely interested in discussing the Motor Abstract Layer!
>
> I will be a pleasure to do that, please come
> http://sourceforge.net/projects/libds402 here to chat avout this.

Thanks for this invitation, but web forums do not fit well with my
communication work flow :-) They are too much "pull" oriented, in the sense
that I have to go to the web forum to see whether someone has made a
comment; and replying via web forms is also not so efficient.

But maybe you can configure the forum to also allow mailing list-like
access, as orocos.org is using?

> Even if the project is related to CAN now, we will do cleaning later when
> things will be clear.

Herman

Orocos and canfestival

On Fri, Jan 25, 2013 at 12:44 AM, Herman Bruyninckx <
Herman [dot] Bruyninckx [..] ...> wrote:

> On Fri, 25 Jan 2013, Willy Lambert wrote:
>
> [...]
>
> I will be a pleasure to do that, please come
>> http://sourceforge.net/**projects/libds402<http://sourceforge.net/projec... to chat avout this.
>>
>
> Thanks for this invitation, but web forums do not fit well with my
> communication work flow :-) They are too much "pull" oriented, in the sense
> that I have to go to the web forum to see whether someone has made a
> comment; and replying via web forms is also not so efficient.
>
> But maybe you can configure the forum to also allow mailing list-like
> access, as orocos.org is using?
>

I made a mailing list, when is ready i will share the link...

Orocos and canfestival

Hi All,
the mailing list of libDS402 project is ready:
https://lists.sourceforge.net/lists/listinfo/libds402-developers.

Join if you like :)

Thanks all,
Bye!

Fabrizio

On Fri, Jan 25, 2013 at 11:23 AM, Fabrizio Boriero <
fabrizio [dot] boriero [..] ...> wrote:

> On Fri, Jan 25, 2013 at 12:44 AM, Herman Bruyninckx <
> Herman [dot] Bruyninckx [..] ...> wrote:
>
>> On Fri, 25 Jan 2013, Willy Lambert wrote:
>>
>> [...]
>>
>> I will be a pleasure to do that, please come
>>> http://sourceforge.net/**projects/libds402<http://sourceforge.net/projec... to chat avout this.
>>>
>>
>> Thanks for this invitation, but web forums do not fit well with my
>> communication work flow :-) They are too much "pull" oriented, in the
>> sense
>> that I have to go to the web forum to see whether someone has made a
>> comment; and replying via web forms is also not so efficient.
>>
>> But maybe you can configure the forum to also allow mailing list-like
>> access, as orocos.org is using?
>>
>
> I made a mailing list, when is ready i will share the link...
>

Orocos and canfestival

2013/1/24 Davide Faconti <faconti [..] ...>:
> FYI.
>
> Some months ago I needed a CANOpen master for a project and frustrated with
> the API of CanFestival and its (apparent?) limitations in terms of
> flexibility, I decided to write mine from scratch.
>
> I know, I know... reinventing the f***ing wheel. But I couldn't find on
> CanFestival a way to have different ObjectDictionaries, i.e. to control
> different brands of motors (for example Elmo + Maxon).
>
> The good news is that my current version work and it is way more
> understandable and flexible.
>
> I am currently working hard to make a commercial software out of it, maybe
> with a double licensing model (GPL for the community + commercial )... I
> don't know.
> Integration with Orocos would be straight forward through a thin wrapper.
> Actually I designed an extra layer that I called Motor Abstract Layer, that
> make possible to control the behaviour of the motor without the need to
> learn the CANopen standard, i.e. you talk about position, velocities,
> torques, etc., and you don't have to deal directly with NTM, SDO, PDO, SYNC
> and all that stuff, unless you want to (you can still do it through an
> additional API).
>
> If there is anyone that want to be a use/test it, let me know. I should be
> able to release it in a couple of weeks.

All depend on your licensing. I'm very interested to try it if it's
released on open license. But if it's commercial, it will be your
effort to do the testing ^^.

I personnaly don't think the closed licensing is the way to go on such
a subject, but it's just an opinion.

>
> Regards
>
> Davide
>
>
>
> On Wed, Jan 23, 2013 at 7:11 PM, Willy Lambert <lambert [dot] willy [..] ...>
> wrote:
>>
>> 2013/1/23 <fabrizio [dot] boriero [..] ...>:
>> > Hi all,
>> > I need to use a opencan DS 402 profile with some elmo boards.
>> >
>> > I like canfestival, is there something already done in Orocos?
>> >
>>
>> Yes and no.
>> Yes, because I think we are several people to play with CanFestival and
>> Orocos.
>>
>> No because none of this stuff is public AFAIK. On my own, it is not
>> public because it is not clean and genric enougth to be public. But
>> basically what I have done is to create standalone ports (I mean out
>> of a component) that are fitted in Canfestival Callbacks. I have A
>> CanManager Component that is doing the "server" work and one component
>> by Can device.
>>
>> I'm quite sure we are several here interested in a nice CanFestival
>> integration. I'm personnaly very interested in this, but I have other
>> plates boiling. Nevertheless, even if CanFestival do the work, I don't
>> find that this framework is very clean, living, ergonomic, up to date,
>> evolutive. I'm split into puting effort to rework all this in C++ or
>> waiting that Can is replaced by other bus with higher bandwidth (in
>> short everything that is playing with Ethernet wiring).
>>
>> That said, my robot is perfectly working with it for now.
>>
>>
>>
>>
>> > Regards,
>> > Fabrizio
>> > --
>> > Orocos-Users mailing list
>> > Orocos-Users [..] ...
>> > http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
>> --
>> Orocos-Users mailing list
>> Orocos-Users [..] ...
>> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
>
>
>
>
> --
> Davide Faconti
> ----------------------------------------------
> Icarus Technology SLU
> ----------------------------------------------
> NIF: B64096571
> tel. +34666722156
> ----------------------------------------------
> c/ paris 45-47, entrlo 3º
> 08029, Barcelona, Spain
> ----------------------------------------------
> www.icarustechnology.com
>

Orocos and canfestival

I am a straight supporter of the free software.
Moreover I am working under an European project so, public money, public
license...

So, Willy and others, what you said if we start to write a clean and
open-source DS-CIA 402 library?
I think we can start from here: http://sourceforge.net/projects/libepos/

I really need this library as soon as possible so I want to start pretty
soon, who's with me? :)

Fabrizio

On Thu, Jan 24, 2013 at 9:22 AM, Willy Lambert <lambert [dot] willy [..] ...>wrote:

> 2013/1/24 Davide Faconti <faconti [..] ...>:
> > FYI.
> >
> > Some months ago I needed a CANOpen master for a project and frustrated
> with
> > the API of CanFestival and its (apparent?) limitations in terms of
> > flexibility, I decided to write mine from scratch.
> >
> > I know, I know... reinventing the f***ing wheel. But I couldn't find on
> > CanFestival a way to have different ObjectDictionaries, i.e. to control
> > different brands of motors (for example Elmo + Maxon).
> >
> > The good news is that my current version work and it is way more
> > understandable and flexible.
> >
> > I am currently working hard to make a commercial software out of it,
> maybe
> > with a double licensing model (GPL for the community + commercial )... I
> > don't know.
> > Integration with Orocos would be straight forward through a thin wrapper.
> > Actually I designed an extra layer that I called Motor Abstract Layer,
> that
> > make possible to control the behaviour of the motor without the need to
> > learn the CANopen standard, i.e. you talk about position, velocities,
> > torques, etc., and you don't have to deal directly with NTM, SDO, PDO,
> SYNC
> > and all that stuff, unless you want to (you can still do it through an
> > additional API).
> >
> > If there is anyone that want to be a use/test it, let me know. I should
> be
> > able to release it in a couple of weeks.
>
> All depend on your licensing. I'm very interested to try it if it's
> released on open license. But if it's commercial, it will be your
> effort to do the testing ^^.
>
> I personnaly don't think the closed licensing is the way to go on such
> a subject, but it's just an opinion.
>
> >
> > Regards
> >
> > Davide
> >
> >
> >
> > On Wed, Jan 23, 2013 at 7:11 PM, Willy Lambert <lambert [dot] willy [..] ...>
> > wrote:
> >>
> >> 2013/1/23 <fabrizio [dot] boriero [..] ...>:
> >> > Hi all,
> >> > I need to use a opencan DS 402 profile with some elmo boards.
> >> >
> >> > I like canfestival, is there something already done in Orocos?
> >> >
> >>
> >> Yes and no.
> >> Yes, because I think we are several people to play with CanFestival and
> >> Orocos.
> >>
> >> No because none of this stuff is public AFAIK. On my own, it is not
> >> public because it is not clean and genric enougth to be public. But
> >> basically what I have done is to create standalone ports (I mean out
> >> of a component) that are fitted in Canfestival Callbacks. I have A
> >> CanManager Component that is doing the "server" work and one component
> >> by Can device.
> >>
> >> I'm quite sure we are several here interested in a nice CanFestival
> >> integration. I'm personnaly very interested in this, but I have other
> >> plates boiling. Nevertheless, even if CanFestival do the work, I don't
> >> find that this framework is very clean, living, ergonomic, up to date,
> >> evolutive. I'm split into puting effort to rework all this in C++ or
> >> waiting that Can is replaced by other bus with higher bandwidth (in
> >> short everything that is playing with Ethernet wiring).
> >>
> >> That said, my robot is perfectly working with it for now.
> >>
> >>
> >>
> >>
> >> > Regards,
> >> > Fabrizio
> >> > --
> >> > Orocos-Users mailing list
> >> > Orocos-Users [..] ...
> >> > http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
> >> --
> >> Orocos-Users mailing list
> >> Orocos-Users [..] ...
> >> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
> >
> >
> >
> >
> > --
> > Davide Faconti
> > ----------------------------------------------
> > Icarus Technology SLU
> > ----------------------------------------------
> > NIF: B64096571
> > tel. +34666722156
> > ----------------------------------------------
> > c/ paris 45-47, entrlo 3º
> > 08029, Barcelona, Spain
> > ----------------------------------------------
> > www.icarustechnology.com
> >
>

Orocos and canfestival

2013/1/24 Fabrizio Boriero <fabrizio [dot] boriero [..] ...>:
> I am a straight supporter of the free software.
> Moreover I am working under an European project so, public money, public
> license...
>
> So, Willy and others, what you said if we start to write a clean and
> open-source DS-CIA 402 library?
> I think we can start from here: http://sourceforge.net/projects/libepos/
>
> I really need this library as soon as possible so I want to start pretty
> soon, who's with me? :)
>
> Fabrizio

I do with a strong will to integrate this (after) with Orocos, but
depends on how, and what. My aim is to have a complete CanOpen
library (in C++ at choice but arguable). DS402 is just a profile (ie a
dictionnary structure) for motors.

But we should switch to other place to discuss details. CanFestival is
not related to Orocos ^^. You could even ask it directly on
CanFestival mailing list which is quite dead anyway.

>
>
> On Thu, Jan 24, 2013 at 9:22 AM, Willy Lambert <lambert [dot] willy [..] ...>
> wrote:
>>
>> 2013/1/24 Davide Faconti <faconti [..] ...>:
>> > FYI.
>> >
>> > Some months ago I needed a CANOpen master for a project and frustrated
>> > with
>> > the API of CanFestival and its (apparent?) limitations in terms of
>> > flexibility, I decided to write mine from scratch.
>> >
>> > I know, I know... reinventing the f***ing wheel. But I couldn't find on
>> > CanFestival a way to have different ObjectDictionaries, i.e. to control
>> > different brands of motors (for example Elmo + Maxon).
>> >
>> > The good news is that my current version work and it is way more
>> > understandable and flexible.
>> >
>> > I am currently working hard to make a commercial software out of it,
>> > maybe
>> > with a double licensing model (GPL for the community + commercial )... I
>> > don't know.
>> > Integration with Orocos would be straight forward through a thin
>> > wrapper.
>> > Actually I designed an extra layer that I called Motor Abstract Layer,
>> > that
>> > make possible to control the behaviour of the motor without the need to
>> > learn the CANopen standard, i.e. you talk about position, velocities,
>> > torques, etc., and you don't have to deal directly with NTM, SDO, PDO,
>> > SYNC
>> > and all that stuff, unless you want to (you can still do it through an
>> > additional API).
>> >
>> > If there is anyone that want to be a use/test it, let me know. I should
>> > be
>> > able to release it in a couple of weeks.
>>
>> All depend on your licensing. I'm very interested to try it if it's
>> released on open license. But if it's commercial, it will be your
>> effort to do the testing ^^.
>>
>> I personnaly don't think the closed licensing is the way to go on such
>> a subject, but it's just an opinion.
>>
>> >
>> > Regards
>> >
>> > Davide
>> >
>> >
>> >
>> > On Wed, Jan 23, 2013 at 7:11 PM, Willy Lambert <lambert [dot] willy [..] ...>
>> > wrote:
>> >>
>> >> 2013/1/23 <fabrizio [dot] boriero [..] ...>:
>> >> > Hi all,
>> >> > I need to use a opencan DS 402 profile with some elmo boards.
>> >> >
>> >> > I like canfestival, is there something already done in Orocos?
>> >> >
>> >>
>> >> Yes and no.
>> >> Yes, because I think we are several people to play with CanFestival and
>> >> Orocos.
>> >>
>> >> No because none of this stuff is public AFAIK. On my own, it is not
>> >> public because it is not clean and genric enougth to be public. But
>> >> basically what I have done is to create standalone ports (I mean out
>> >> of a component) that are fitted in Canfestival Callbacks. I have A
>> >> CanManager Component that is doing the "server" work and one component
>> >> by Can device.
>> >>
>> >> I'm quite sure we are several here interested in a nice CanFestival
>> >> integration. I'm personnaly very interested in this, but I have other
>> >> plates boiling. Nevertheless, even if CanFestival do the work, I don't
>> >> find that this framework is very clean, living, ergonomic, up to date,
>> >> evolutive. I'm split into puting effort to rework all this in C++ or
>> >> waiting that Can is replaced by other bus with higher bandwidth (in
>> >> short everything that is playing with Ethernet wiring).
>> >>
>> >> That said, my robot is perfectly working with it for now.
>> >>
>> >>
>> >>
>> >>
>> >> > Regards,
>> >> > Fabrizio
>> >> > --
>> >> > Orocos-Users mailing list
>> >> > Orocos-Users [..] ...
>> >> > http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
>> >> --
>> >> Orocos-Users mailing list
>> >> Orocos-Users [..] ...
>> >> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
>> >
>> >
>> >
>> >
>> > --
>> > Davide Faconti
>> > ----------------------------------------------
>> > Icarus Technology SLU
>> > ----------------------------------------------
>> > NIF: B64096571
>> > tel. +34666722156
>> > ----------------------------------------------
>> > c/ paris 45-47, entrlo 3º
>> > 08029, Barcelona, Spain
>> > ----------------------------------------------
>> > www.icarustechnology.com
>> >
>
>

Orocos and canfestival

I made this project: http://sourceforge.net/projects/libds402 we can
discuss there...

Fabrizio

On Thu, Jan 24, 2013 at 10:31 AM, Willy Lambert <lambert [dot] willy [..] ...>wrote:

> 2013/1/24 Fabrizio Boriero <fabrizio [dot] boriero [..] ...>:
> > I am a straight supporter of the free software.
> > Moreover I am working under an European project so, public money, public
> > license...
> >
> > So, Willy and others, what you said if we start to write a clean and
> > open-source DS-CIA 402 library?
> > I think we can start from here: http://sourceforge.net/projects/libepos/
> >
> > I really need this library as soon as possible so I want to start pretty
> > soon, who's with me? :)
> >
> > Fabrizio
>
> I do with a strong will to integrate this (after) with Orocos, but
> depends on how, and what. My aim is to have a complete CanOpen
> library (in C++ at choice but arguable). DS402 is just a profile (ie a
> dictionnary structure) for motors.
>
> But we should switch to other place to discuss details. CanFestival is
> not related to Orocos ^^. You could even ask it directly on
> CanFestival mailing list which is quite dead anyway.
>
>
> >
> >
> > On Thu, Jan 24, 2013 at 9:22 AM, Willy Lambert <lambert [dot] willy [..] ...>
> > wrote:
> >>
> >> 2013/1/24 Davide Faconti <faconti [..] ...>:
> >> > FYI.
> >> >
> >> > Some months ago I needed a CANOpen master for a project and frustrated
> >> > with
> >> > the API of CanFestival and its (apparent?) limitations in terms of
> >> > flexibility, I decided to write mine from scratch.
> >> >
> >> > I know, I know... reinventing the f***ing wheel. But I couldn't find
> on
> >> > CanFestival a way to have different ObjectDictionaries, i.e. to
> control
> >> > different brands of motors (for example Elmo + Maxon).
> >> >
> >> > The good news is that my current version work and it is way more
> >> > understandable and flexible.
> >> >
> >> > I am currently working hard to make a commercial software out of it,
> >> > maybe
> >> > with a double licensing model (GPL for the community + commercial
> )... I
> >> > don't know.
> >> > Integration with Orocos would be straight forward through a thin
> >> > wrapper.
> >> > Actually I designed an extra layer that I called Motor Abstract Layer,
> >> > that
> >> > make possible to control the behaviour of the motor without the need
> to
> >> > learn the CANopen standard, i.e. you talk about position, velocities,
> >> > torques, etc., and you don't have to deal directly with NTM, SDO, PDO,
> >> > SYNC
> >> > and all that stuff, unless you want to (you can still do it through an
> >> > additional API).
> >> >
> >> > If there is anyone that want to be a use/test it, let me know. I
> should
> >> > be
> >> > able to release it in a couple of weeks.
> >>
> >> All depend on your licensing. I'm very interested to try it if it's
> >> released on open license. But if it's commercial, it will be your
> >> effort to do the testing ^^.
> >>
> >> I personnaly don't think the closed licensing is the way to go on such
> >> a subject, but it's just an opinion.
> >>
> >> >
> >> > Regards
> >> >
> >> > Davide
> >> >
> >> >
> >> >
> >> > On Wed, Jan 23, 2013 at 7:11 PM, Willy Lambert <
> lambert [dot] willy [..] ...>
> >> > wrote:
> >> >>
> >> >> 2013/1/23 <fabrizio [dot] boriero [..] ...>:
> >> >> > Hi all,
> >> >> > I need to use a opencan DS 402 profile with some elmo boards.
> >> >> >
> >> >> > I like canfestival, is there something already done in Orocos?
> >> >> >
> >> >>
> >> >> Yes and no.
> >> >> Yes, because I think we are several people to play with CanFestival
> and
> >> >> Orocos.
> >> >>
> >> >> No because none of this stuff is public AFAIK. On my own, it is not
> >> >> public because it is not clean and genric enougth to be public. But
> >> >> basically what I have done is to create standalone ports (I mean out
> >> >> of a component) that are fitted in Canfestival Callbacks. I have A
> >> >> CanManager Component that is doing the "server" work and one
> component
> >> >> by Can device.
> >> >>
> >> >> I'm quite sure we are several here interested in a nice CanFestival
> >> >> integration. I'm personnaly very interested in this, but I have other
> >> >> plates boiling. Nevertheless, even if CanFestival do the work, I
> don't
> >> >> find that this framework is very clean, living, ergonomic, up to
> date,
> >> >> evolutive. I'm split into puting effort to rework all this in C++ or
> >> >> waiting that Can is replaced by other bus with higher bandwidth (in
> >> >> short everything that is playing with Ethernet wiring).
> >> >>
> >> >> That said, my robot is perfectly working with it for now.
> >> >>
> >> >>
> >> >>
> >> >>
> >> >> > Regards,
> >> >> > Fabrizio
> >> >> > --
> >> >> > Orocos-Users mailing list
> >> >> > Orocos-Users [..] ...
> >> >> > http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
> >> >> --
> >> >> Orocos-Users mailing list
> >> >> Orocos-Users [..] ...
> >> >> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
> >> >
> >> >
> >> >
> >> >
> >> > --
> >> > Davide Faconti
> >> > ----------------------------------------------
> >> > Icarus Technology SLU
> >> > ----------------------------------------------
> >> > NIF: B64096571
> >> > tel. +34666722156
> >> > ----------------------------------------------
> >> > c/ paris 45-47, entrlo 3º
> >> > 08029, Barcelona, Spain
> >> > ----------------------------------------------
> >> > www.icarustechnology.com
> >> >
> >
> >
>

Orocos and canfestival

On Thu, Jan 24, 2013 at 6:59 PM, Fabrizio Boriero <
fabrizio [dot] boriero [..] ...> wrote:

> I am a straight supporter of the free software.
> Moreover I am working under an European project so, public money, public
> license...
>

This is not the place to start any philosophical arguing about open source
(even it happened before in a post about the RTI "open" license).

Isn't GPL the most public license ? ;)
So, what is the problem with a double license model?
As far as I know there is no problem using GPL code in European project, am
I right?

I support open source too, but I need to pay my bills :D
Is it evil to pay my rent selling software I spent weeks writing and
testing?

If you need the library as soon as possible, i will be happy to give you
mine and in a couple of days at most you have done, otherwise you take your
money back... that is zero :D

Davide

Orocos and canfestival

On Thu, Jan 24, 2013 at 07:22:24PM +0100, Davide Faconti wrote:
>
>
>
> On Thu, Jan 24, 2013 at 6:59 PM, Fabrizio Boriero <fabrizio [dot] boriero [..] ...>
> wrote:
>
> I am a straight supporter of the free software.
> Moreover I am working under an European project so, public money, public
> license...
>
>
> This is not the place to start any philosophical arguing about open source
> (even it happened before in a post about the RTI "open" license).
>
> Isn't GPL the most public license ? ;)
> So, what is the problem with a double license model?
> As far as I know there is no problem using GPL code in European project, am I
> right?

The biggest problem with such dual license models is how to deal with
contributions. Without additional legal fuss (copyright assigments,
etc) you can not merge GPL'ed contributions without makeing your
project GPL only. Needless to say that many free software people
(incl. myself) will be reluctant to give up their copyright to a
commercial entity.

> I support open source too, but I need to pay my bills :D
> Is it evil to pay my rent selling software I spent weeks writing and testing?
>
> If you need the library as soon as possible, i will be happy to give you mine
> and in a couple of days at most you have done, otherwise you take your money
> back... that is zero :D

Best regards
Markus

Orocos and canfestival

On Thu, 24 Jan 2013, Davide Faconti wrote:

> On Thu, Jan 24, 2013 at 6:59 PM, Fabrizio Boriero <fabrizio [dot] boriero [..] ...> wrote:
> I am a straight supporter of the free software.
> Moreover I am working under an European project so, public money, public
> license...
>
>
> This is not the place to start any philosophical arguing about open source (even it
> happened before in a post about the RTI "open" license).
>
> Isn't GPL the most public license ? ;) 
> So, what is the problem with a double license model?
> As far as I know there is no problem using GPL code in European project, am I right?

Partially: all partners in the consortium have to agree on that. And since
GPL is "viral",
<http://en.wikipedia.org/wiki/Viral_license>
integrating it into system with software from various partners _might_ put
some of these partners in an unfortunate situation with reusing the
project's outcome later on.

I have been in many such projects that worked with/for open source, but GPL
licenses never made it into the Consortium Agreements; LGPL, BSD, MPL, EPL
did, because of their "non-viral" nature.

> I support open source too, but I need to pay my bills :D
> Is it evil to pay my rent selling software I spent weeks writing and testing?
>
> If you need the library as soon as possible, i will be happy to give you mine and in a
> couple of days at most you have done, otherwise you take your money back... that is
> zero :D
>
> Davide

Herman

Orocos and canfestival

2013/1/25 Herman Bruyninckx <Herman [dot] Bruyninckx [..] ...>:
> On Thu, 24 Jan 2013, Davide Faconti wrote:
>
>> On Thu, Jan 24, 2013 at 6:59 PM, Fabrizio Boriero
>> <fabrizio [dot] boriero [..] ...> wrote:
>> I am a straight supporter of the free software.
>> Moreover I am working under an European project so, public money,
>> public
>> license...
>>
>>
>> This is not the place to start any philosophical arguing about open source
>> (even it
>> happened before in a post about the RTI "open" license).
>>
>> Isn't GPL the most public license ? ;)
>> So, what is the problem with a double license model?
>> As far as I know there is no problem using GPL code in European project,
>> am I right?
>
>
> Partially: all partners in the consortium have to agree on that. And since
> GPL is "viral",
> <http://en.wikipedia.org/wiki/Viral_license>
> integrating it into system with software from various partners _might_ put
> some of these partners in an unfortunate situation with reusing the
> project's outcome later on.
>
> I have been in many such projects that worked with/for open source, but GPL
> licenses never made it into the Consortium Agreements; LGPL, BSD, MPL, EPL
> did, because of their "non-viral" nature.
>

I personnaly prefer non viral licensing. Davide, we should open a
thread on licensing. I'm not good at licensing, but viral license must
be in part that won't contamine all client work (else no one will use
the library). If it's a money problem, I think the best way is to have
"closed" license until initial funding have been found, then release
it public.

>
>> I support open source too, but I need to pay my bills :D
>> Is it evil to pay my rent selling software I spent weeks writing and
>> testing?
>>
>> If you need the library as soon as possible, i will be happy to give you
>> mine and in a
>> couple of days at most you have done, otherwise you take your money
>> back... that is
>> zero :D
>>
>> Davide
>
>
> Herman
> --
> Orocos-Users mailing list
> Orocos-Users [..] ...
> http://lists.mech.kuleuven.be/mailman/listinfo/orocos-users
>

(Non)Viral licensing (Was: Orocos and canfestival)

On Fri, 25 Jan 2013, Willy Lambert wrote:

> 2013/1/25 Herman Bruyninckx <Herman [dot] Bruyninckx [..] ...>:
>> On Thu, 24 Jan 2013, Davide Faconti wrote:
>>
>>> On Thu, Jan 24, 2013 at 6:59 PM, Fabrizio Boriero
>>> <fabrizio [dot] boriero [..] ...> wrote:
>>> I am a straight supporter of the free software.
>>> Moreover I am working under an European project so, public money,
>>> public license...
>>>
>>> This is not the place to start any philosophical arguing about open source
>>> (even it happened before in a post about the RTI "open" license).
>>>
>>> Isn't GPL the most public license ? ;)
>>> So, what is the problem with a double license model?
>>> As far as I know there is no problem using GPL code in European project,
>>> am I right?
>>
>> Partially: all partners in the consortium have to agree on that. And since
>> GPL is "viral",
>> <http://en.wikipedia.org/wiki/Viral_license>
>> integrating it into system with software from various partners _might_ put
>> some of these partners in an unfortunate situation with reusing the
>> project's outcome later on.
>>
>> I have been in many such projects that worked with/for open source, but GPL
>> licenses never made it into the Consortium Agreements; LGPL, BSD, MPL, EPL
>> did, because of their "non-viral" nature.
>
> I personnaly prefer non viral licensing. Davide, we should open a
> thread on licensing.

Just did... :-)

> I'm not good at licensing, but viral license must
> be in part that won't contamine all client work (else no one will use
> the library). If it's a money problem, I think the best way is to have
> "closed" license until initial funding have been found, then release
> it public.

>>> I support open source too, but I need to pay my bills :D

There are a couple of business models around open source code:
- CAAS ("coding as a service"): a company charges to write code, with an
open source license. The different business models inside of CAAS are:
- That license can be viral or not: in the viral case, the company can
charge for a commercial license that allows the code to be embedded in
a client's application. This was the business model of MySql, in the
beginning.
- The company can decide to accept external contributions or not. In the
latter case, the company can make a commercial version without having
to ask anyone.
- AAAS ("application as a service"): the company charges to help a client
make an application with open source code. The license does not matter
too much.

>>> Is it evil to pay my rent selling software I spent weeks writing and
>>> testing?
>>>
>>> If you need the library as soon as possible, i will be happy to give you
>>> mine and in a
>>> couple of days at most you have done, otherwise you take your money
>>> back... that is
>>> zero :D
>>>
>>> Davide

Herman

(Non)Viral licensing (Was: Orocos and canfestival)

2013/1/25 Herman Bruyninckx <Herman [dot] Bruyninckx [..] ...>:
> On Fri, 25 Jan 2013, Willy Lambert wrote:
>
>> 2013/1/25 Herman Bruyninckx <Herman [dot] Bruyninckx [..] ...>:
>>>
>>> On Thu, 24 Jan 2013, Davide Faconti wrote:
>>>
>>>> On Thu, Jan 24, 2013 at 6:59 PM, Fabrizio Boriero
>>>> <fabrizio [dot] boriero [..] ...> wrote:
>>>> I am a straight supporter of the free software.
>>>> Moreover I am working under an European project so, public money,
>>>> public license...
>>>>
>>>> This is not the place to start any philosophical arguing about open
>>>> source
>>>> (even it happened before in a post about the RTI "open" license).
>>>>
>>>> Isn't GPL the most public license ? ;)
>>>> So, what is the problem with a double license model?
>>>> As far as I know there is no problem using GPL code in European project,
>>>> am I right?
>>>
>>>
>>> Partially: all partners in the consortium have to agree on that. And
>>> since
>>> GPL is "viral",
>>> <http://en.wikipedia.org/wiki/Viral_license>
>>> integrating it into system with software from various partners _might_
>>> put
>>> some of these partners in an unfortunate situation with reusing the
>>> project's outcome later on.
>>>
>>> I have been in many such projects that worked with/for open source, but
>>> GPL
>>> licenses never made it into the Consortium Agreements; LGPL, BSD, MPL,
>>> EPL
>>> did, because of their "non-viral" nature.
>>
>>
>> I personnaly prefer non viral licensing. Davide, we should open a
>> thread on licensing.
>
>
> Just did... :-)

Thanks, but I meant outside Orocos

>
>> I'm not good at licensing, but viral license must
>> be in part that won't contamine all client work (else no one will use
>> the library). If it's a money problem, I think the best way is to have
>> "closed" license until initial funding have been found, then release
>> it public.
>
>
>>>> I support open source too, but I need to pay my bills :D
>
>
> There are a couple of business models around open source code:
> - CAAS ("coding as a service"): a company charges to write code, with an
> open source license. The different business models inside of CAAS are:
> - That license can be viral or not: in the viral case, the company can
> charge for a commercial license that allows the code to be embedded in
> a client's application. This was the business model of MySql, in the
> beginning.
> - The company can decide to accept external contributions or not. In the
> latter case, the company can make a commercial version without having
> to ask anyone.
> - AAAS ("application as a service"): the company charges to help a client
> make an application with open source code. The license does not matter
> too much.
>
>>>> Is it evil to pay my rent selling software I spent weeks writing and
>>>> testing?
>>>>
>>>> If you need the library as soon as possible, i will be happy to give you
>>>> mine and in a
>>>> couple of days at most you have done, otherwise you take your money
>>>> back... that is
>>>> zero :D
>>>>
>>>> Davide
>
>
> Herman

(Non)Viral licensing (Was: Orocos and canfestival)

Conscious that we are annoying someone with this non-Orocos topic, I can't
resist replying.

>>>
> >>>
> >>> Partially: all partners in the consortium have to agree on that. And
> >>> since
> >>> GPL is "viral",
> >>> <http://en.wikipedia.org/wiki/Viral_license>
> >>> integrating it into system with software from various partners _might_
> >>> put
> >>> some of these partners in an unfortunate situation with reusing the
> >>> project's outcome later on.
>

That is the whole point of the irony (IMHO):

1) Fabrizio (and other people like him, I just use him as a void* to
indicate a more general class) is working with public money and wants to
share his work with others. Cool.
2) At the same time the industrial partners in the consortium don't want
viral software to have the possibility to use that software done by
Fabriziowith public money for their own commercial purpose (i.e. to
make money).

Of course this is not a bad thing, since the whole community can take
advantage of the work of Fabrizio. Cool.

On the other hand, I wouldn't demonize those individuals or companies that
don't have public money to spend and want to honestly take a profit with
their work offering at the same time the possibility for a certain kind of
users (the non-for-profit ones) to get the software and use it for their
purpose.

The double license model sounds fair: if you want a non viral license for
you commercial product, just pay for it a end of the story.

> >> I personnaly prefer non viral licensing. Davide, we should open a
> >> thread on licensing.
> >
>

I swear that this is my last email on this mailing list... but I am
crossing my fingers ;)

>
> >> I'm not good at licensing, but viral license must
> >> be in part that won't contamine all client work (else no one will use
> >> the library). If it's a money problem, I think the best way is to have
> >> "closed" license until initial funding have been found, then release
> >> it public.
>

Would a closed-source, free for academic and non-commercial use make you
happy ? ;)
I am shooting my last bullet :D
Ok, last attempt... I can include also a mug with your name on it ;)

> >
>
> > There are a couple of business models around open source code:
> > - CAAS ("coding as a service"): a company charges to write code, with an
> > open source license. The different business models inside of CAAS are:
> > - That license can be viral or not: in the viral case, the company can
> > charge for a commercial license that allows the code to be embedded
> in
> > a client's application. This was the business model of MySql, in the
> > beginning.
> > - The company can decide to accept external contributions or not. In
> the
> > latter case, the company can make a commercial version without having
> > to ask anyone.
> > - AAAS ("application as a service"): the company charges to help a client
> > make an application with open source code. The license does not matter
> > too much.
>

Of course I am open to these kinds of models, but I believe that in this
case (CANOpen 402 and 301 master), it is more convenient for both sides (by
on order of magnitude) to have the client paying a very low price for a
software that would take several man-weeks to be written, maintained and
debugged.

A model that doesn't convince me very much is the "free-software, paid
support" model (I don't know if there is a better name with initials) is
the documentation dilemma: if you don't have enough documentation, no one
use your code, if it is too good, no one will need the paid support.

Anyway, thanks all of you for your patience, I know I abused a little bit
of this mailing-list

Have a nice day.

Davide

(Non)Viral licensing (Was: Orocos and canfestival)

On Fri, 25 Jan 2013, Davide Faconti wrote:

> Conscious that we are annoying someone with this non-Orocos topic, I can't resist
> replying. 
>
> >>>
> >>>
> >>> Partially: all partners in the consortium have to agree on that. And
> >>> since
> >>> GPL is "viral",
> >>>  <http://en.wikipedia.org/wiki/Viral_license>
> >>> integrating it into system with software from various partners _might_
> >>> put
> >>> some of these partners in an unfortunate situation with reusing the
> >>> project's outcome later on.
>
>
> That is the whole point of the irony (IMHO):
>
> 1) Fabrizio (and other people like him, I just use him as a void* to indicate a more
> general class) is working with public money and wants to share his work with others.
> Cool.
> 2) At the same time the industrial partners in the consortium don't want viral software
> to have the possibility to use that software done by Fabrizio with public money for
> their own commercial purpose (i.e. to make money).
>
> Of course this is not a bad thing, since the whole community can take advantage of the
> work of Fabrizio. Cool.
>
> On the other hand, I wouldn't demonize those individuals or companies that don't have
> public money to spend and want to honestly take a profit with their work offering at
> the same time the possibility for a certain kind of users (the non-for-profit ones) to
> get the software and use it for their purpose.

There is a flaw in your reasoning, which is only revealed if you bring in
the business model _and_ the project collaboration policy in the picture,
and not only the licenses: a company that releases code under GPL but does
not acceptg user contributions (except under a copyright transfer clause)
because that would prevent the company from exploiting the dual software
license business model, should not be surprised that users are hesitant to
contribute. Because most users do not just download and use software once,
but they want to subscribe to a "life cycle" in the evolution of the
software, and a no-acceptance collaboration policy demotivates users
tremendously fast, since they will have to maintain their non-accepted
contributions themselves.

> The double license model sounds fair: if you want a non viral license for you
> commercial product, just pay for it a end of the story.

Of course it is fair. But it is not perfect for all users, especially not
the creative ones.

> >> I personnaly prefer non viral licensing.  Davide, we should open a
> >> thread on licensing.
>
> I swear that this is my last email on this mailing list... but I am crossing my fingers
> ;)

I stopped swearing :-) But this discussion is, as far as I am concerned,
on-topic, since Peter Soetens and myself _very conciously_ made the decision
for the choice of license for Orocos in 2001. And that decision was to
follow a collaboration and licensing model that would (only) allow the CAAS
and AAAS business models, and not the dual licensing. I think the current
situation proves us right in the sense that there are a lot of external
contributions because these users are not limited at all by a commercial
agenda of the "Orocos project".

> >> I'm not good at licensing, but viral license must
> >> be in part that won't contamine all client work (else no one will use
> >> the library). If it's a money problem, I think the best way is to have
> >> "closed" license until initial funding have been found, then release
> >> it public.
>
> Would a closed-source, free for academic and non-commercial use make you happy ? ;)

Only in the case that there is no decent free software alternative. Adobe
Reader was the last one that survived on my computers in this way, but that
is already many years ago.

> I am shooting my last bullet :D
> Ok, last attempt... I can include also a mug with your name on it ;)

Aha! Bribing has _always_ proven to work as a business model :-)

> > There are a couple of business models around open source code:
> > - CAAS ("coding as a service"): a company charges to write code, with an
> >   open source license. The different business models inside of CAAS are:
> >   - That license can be viral or not: in the viral case, the company can
> >     charge for a commercial license that allows the code to be embedded in
> >     a client's application. This was the business model of MySql, in the
> >     beginning.
> >   - The company can decide to accept external contributions or not. In the
> >     latter case, the company can make a commercial version without having
> >     to ask anyone.
> > - AAAS ("application as a service"): the company charges to help a client
> >   make an application with open source code. The license does not matter
> >   too much.
>  
> Of course I am open to these kinds of models, but I believe that in this
> case (CANOpen 402 and 301 master), it is more convenient for both sides
> (by on order of magnitude) to have the client paying a very low price for
> a software that would take several man-weeks to be written, maintained
> and debugged. 

This kind of "driver" code is _never_ an end-application, and must always
be integrated in a larger system; hence, the price of the module in itself
is dwarfed by the potential price its viral license could have. This is of
course something that depend more on the application then on the code you
provide.

> A model that doesn't convince me very much is the "free-software, paid support" model
> (I don't know if there is a better name with initials)

"SAAS": software as a service. Red Hat makes over $1B a year with that
model; some cloud service companies are approaching the $500M mark.
This model does definitely not fit to the driver code this thread was
started on, because that is too little code without any interdependencies
between several versions and other software. SAAS works in the "Red Hat"
case: people pay to get a working system, where the company takes care of
all interdependencies, updates, version hells, etc.

In your case "CAAS" would fit best, with a non-viral license: I have
already paid several times to let third parties develop driver code. For
the same reason as why I pay my bakkery for my bread, or my plumber for
servicing my house: they do the job better, faster and cheaper than I
could, and they do not lock me in into anything, in the sense that I can
call go to another bakkery, or call in another plumber, whenever I want.

> is the documentation dilemma: if you don't have enough documentation, no
> one use your code, if it is too good, no one will need the paid support.

> Anyway, thanks all of you for your patience, I know I abused a little bit of this
> mailing-list 

I don't think so, because software-driven businesses _are_ among the
"orocos-users" that I have had in mind since the beginning, and still have. :-)

> Have a nice day.
>
> Davide

Herman