Is ROS death?

[Manuel Rodriguez]

The ROS Project was founded as an opensource integrative robotics framework. Most universities and private research institute are using ROS today for their work. Even Boston Dynamics (Petman) and Moley Kitchen Robot have installed a ROS instance for driving their robots. The problem is, that in both cases the sourcecode is not free and only few people know, that this code exists in general. Thats the antithesis to the ROS ideology.

The ROS framework which can currently be downloaded from the offical website is not the real ROS, it is only a part of ROS which is released for educational reasons. Its like a light-version of the real program which is developed by the DARPA. So my question is, what should programmers do with this kind of software?

In my opinion ROS is a big lie. The message is, that all is opensource and free while in reality nothing usefull is released, only the guys who work for DARPA have full reading access to the ROS repository, not the mass.

[Hendrik Wiese]

I'm not quite sure what you're talking about... the full source code of
the ROS base system is available on GitHub
(http://wiki.ros.org/RecommendedRepositoryUsage/CommonGitHubOrganizations),
most packages are as well. Yes, there are some packages that I know of
which have not been made publicly available. But they aren't
essential/core packages.

I don't see how you got to your conclusion. Can you provide some
evidence for your allegation?

[v4hn]

Hey Manuel,

I'm sorry, but I fail to see how this relates to MoveIt! and its user base
(and thus I don't see why you sent this to the moveit-users mailinglist).

ROS is an OpenSource *Framework* implemented with the BSD software license.
A lot of software modules for this framework are the same.
While I believe that GPL is the better choice for Free Software in general,
I also believe that ROS would not be as popular/supported as it is today
if industry, and yes sadly also military, would be forced to release the code
they build on top. So this distribution model is *essential* for ROS and not a "big lie".

On Sat, Jul 23, 2016 at 12:30:51PM -0700, Manuel Rodriguez wrote:
> The ROS framework which can currently be downloaded from the offical
> website is not the real ROS, it is only a part of ROS which is released for
> educational reasons. Its like a light-version of the real program which is
> developed by the DARPA.

> [...]

> In my opinion ROS is a big lie. The message is, that all is opensource and
> free while in reality nothing usefull is released, only the guys who work
> for DARPA have full reading access to the ROS repository, not the mass.

This is a plain conspiracy theory. No need to reply to that.

> So my question is, what should programmers do with this kind of software?

Improve on it and improve it.
Pull Requests & Patches are always welcome!
If you come up with a useful software module / concept,
implement it in a general, use-case-agnostic way and release it under a
Free Software License.

Best,


v4hn

[Manuel Rodriguez]

Am Sonntag, 24. Juli 2016 08:47:25 UTC+2 schrieb Hendrik Wiese:

I don't see how you got to your conclusion. Can you provide some
evidence for your allegation?

Sure. The Moley Kitchen Robot was developed with ROS, see the job description here https://moley.recruiterbox.com/jobs/fk06ulo/
And other advanced Robotics like Yamaha MotoBot or Google Atlas box pushing robot are not possible without using an integrated robotframework with vision, action planning, manipulation and realtime functionality. The only software on the market is ROS. It is not possible that these robots are programmed with a different software.

But if I search into the ROS repository there is no package from Yamaha, Moley, Google or others. So these are hidden packages which are more than "minor packages with less interesting features".

[G.A. vd. Hoorn - 3ME]

On 24-7-2016 14:21, Manuel Rodriguez wrote:
> Am Sonntag, 24. Juli 2016 08:47:25 UTC+2 schrieb Hendrik Wiese:
>>
>> I don't see how you got to your conclusion. Can you provide some
>> evidence for your allegation?
>>
>
> Sure. The Moley Kitchen Robot was developed with ROS, see the job
> description here https://moley.recruiterbox.com/jobs/fk06ulo/
> And other advanced Robotics like Yamaha MotoBot or Google Atlas box pushing
> robot are not possible without using an integrated robotframework with
> vision, action planning, manipulation and realtime functionality. The only
> software on the market is ROS. It is not possible that these robots are
> programmed with a different software.

Don't want to feed the troll, but this is a non-sense statement.

What about Orocos & ROCK, OpenRTM, Player, URBI, Roboframe, ASEBA, YARP,
CORBIS, MRDS, CLARAty, Smartsoft, MOOS, ORCA, Jasmine, OpenROV,
OpenJAUS, MIRPA, Naoqi and EEROS to name a few?

And then the thousands of in-house developed frameworks and control
architectures / extensions like those developed for Walk-Man, HRP,
Asimo, Justin, etc, etc?

Please inform yourself before posting such unfounded statements, and
don't insult the robotics (software) community by implying that only the
people at OSRF are capable of creating robot control software
frameworks. However nice ROS may be for some applications, it's
certainly not the only viable choice for robotic software development.


Gijs

[Manuel Rodriguez]

Am Sonntag, 24. Juli 2016 15:43:14 UTC+2 schrieb gavanderhoorn:

What about Orocos & ROCK, OpenRTM, ...

I like them all. Not because they are so powerfull like ROS, but because its is better to reinvent the wheel again. It makes no sense to use a software which is magical if it's unclear who has developed it or what this software can't do. If somebody writes his own C++ or Java Tool for controlling a robot it is obvious that this project will fail because the task is hard and programming is complex. But to fail is better than what Willow Garage has done ...

Failing means to learn, failing means to write a paper about the issue. Failing means not to use technology for promoting fear in that form like google, apple or microsoft use technology.

[G.A. vd. Hoorn - 3ME]

On 24-7-2016 16:10, Manuel Rodriguez wrote:
> Am Sonntag, 24. Juli 2016 15:43:14 UTC+2 schrieb gavanderhoorn:
>>

>> What about Orocos& ROCK, OpenRTM, ...

>
>
> I like them all. Not because they are so powerfull like ROS, but because
> its is better to reinvent the wheel again. It makes no sense to use a
> software which is magical if it's unclear who has developed it or what this
> software can't do. If somebody writes his own C++ or Java Tool for
> controlling a robot it is obvious that this project will fail because the
> task is hard and programming is complex. But to fail is better than what
> Willow Garage has done ...
>
> Failing means to learn, failing means to write a paper about the issue.
> Failing means not to use technology for promoting fear in that form like
> google, apple or microsoft use technology.

I responded specifically to:

> It is not possible that these robots are programmed with a different
software.

Which is non-sense.

The rest I will ignore.


Gijs

[Hendrik Wiese]

Am 24.07.2016 um 16:10 schrieb Manuel Rodriguez:
> Am Sonntag, 24. Juli 2016 15:43:14 UTC+2 schrieb gavanderhoorn:
>
> What about Orocos & ROCK, OpenRTM, ...
>
>
> I like them all. Not because they are so powerfull like ROS, but because
> its is better to reinvent the wheel again. It makes no sense to use a
> software which is magical if it's unclear who has developed it or what
> this software can't do. If somebody writes his own C++ or Java Tool for
> controlling a robot it is obvious that this project will fail because
> the task is hard and programming is complex. But to fail is better than
> what Willow Garage has done ...

Jesus... What, for God's sake, has Willow Garage done??? They haven't
done anything that'd justify your ranting! ROS is still open source.
Entirely! Yes, there are packages that are not open source because they
are very specific and might contain secret code. But who cares? That's
not Willow Garage's fault. ROS is open and remains open. Seriously, get
your homework done before you come up with bullshit like this!

>
> Failing means to learn, failing means to write a paper about the issue.
> Failing means not to use technology for promoting fear in that form like
> google, apple or microsoft use technology.

Who has failed? And in what way?

Stop ranting and provide some info!

[Ugo Cupcic]

Hey,

Just a quick note from me (I don't want to start feeding the troll). Since we've developed Moley at Shadow, I can tell you that most of the work was done as open source - just check shadow-robot contribution to the ros ecosystem.... Then the work specific to moley was closed source for obvious reasons. But it doesn't mean that this kind of work does not feedback in the community. On the contrary, this kind of commercial projects move ros forward in my humble opinion.

Cheers,

Ugo

Sent from a tiny keyboard, excuse the brevity!

We're hiring Software Developers! - visit our website for more details.

---

Shadow Robot Company Ltd.
251 Liverpool Road, N1 1LX, UK
Registered Number 3308007 (England & Wales)

---

[Manuel Rodriguez]

Am Montag, 25. Juli 2016 08:46:01 UTC+2 schrieb Ugo Cupcic:

Then the work specific to moley was closed source for obvious reasons.

Are you sure, that the Dynamic Movement Primitive (DMP) which controlls the moley kitchen robot are working correct and that no teleoperated control during the presentation is necessary?