[vcf-midatlantic] MARCHIAN invasion of Trenton
Douglas Crawford
touchetek at gmail.com
Sat Mar 25 17:12:24 EDT 2017
On 3/25/2017 2:20 PM, Herb Johnson via vcf-midatlantic wrote:
> Herb Johnson:
>
>>> Evan is correct, in his reportage that I said his and Jeff's
>>> robotics talk was more
>>> technically informative, than my discussion with a student and
>>> faculty-member of their
>>> $40K demonstration robot. I could not get those persons, to describe
>>> the programming
>>> interface to their tower of plastic and motors. Evan showed a BASIC
>>> program with peeks and
>>> pokes. The 20th century won that "battle of the 'bots". Am I a
>>> judge? I worked on robotic
>>> system interfaces to minicomputers in 1975.
>>
> Neil Cherry:
>
>> I find great humor in this. Evan who learned on his own (with help,
>> but still) knew more
>> about the technical details of his project than the students did of
>> theirs. This says
>> good things about Evan.
>
> This is amusing but don't take my remarks too far. I'm sure a 2017
> engineering student knows a lot that I don't know. I'm sure a
> professor of electrical engineering knows a lot I don't know. Read
> what I'm saying. I said "I could not get these persons to DESCRIBE the
> programming interface...."
>
> My casual understanding from discussion, is that the college's robot
> was obtained a few months ago. And the plan is to work harder on it,
> in the summer. So it simply may not have been familiar to either
> party. And the student may have been intimidated by hard questions;
> and the instructor may not have had time to look at the programming
> features.
>
> All that said, my point was that neither person could DESCRIBE in a
> technical way, what their robot could do at the program command level.
> It was sold as a development tool, not just some pick-n-place robot. I
> asked the question a number of ways, and couldn't get a straight answer.
>
> The most direct inferences are 1) they weren't prepared for technical
> questions or 2) the robot did not come with a clearly defined "API" or
> applications programming interface or 3) the robot wasn't well
> documented.
>
> The instructor said something about supporting the "Robotics Operating
> System" which is "open source" but could not (or would not) inform me
> further. I've heard of it:
>
> http://www.ros.org/
>
> and so I go to the Web page. The home page points to "features" which
> are listed as:
>
> Standard Message Definitions for Robots
> Robot Geometry Library
> Robot Description Language
> Preemptable Remote Procedure Calls
> Diagnostics
> Pose Estimation
> Localization
> Mapping
> Navigation
>
> Creee-IPES! What a bunch of gobble-gook! I'm an engineer, dammit, not
> a linguist, or a choreographer! It's nice to know I can pass messages,
> that will help me in homeroom to send notes to my buddies and that new
> girl. But - what can I move, what can I sense, what can I *see*
> through this ROS, for this robot?
>
> The page talks about "poses" and "graphical interfaces" and
> "diagnostics", "data logging"....(sigh) I'm aware of these issues and
> methods, I understand you can write a MS-degree thesis by playing with
> this stuff. And I see their "testimonials" from PhD's and grad
> students, and one "hobbyist".
>
> This goes to my "thesis" - anyone can pose a problem as a thesis, you
> know. My thesis is this: it's worth preserving 20th century (early)
> vintage computing, because the technology of the era is simple enough
> to grasp at a fundamental level. Therefore, it's accessible to those
> of modest skills, or those looking back at forgotten technology.
>
> More important, this technology solved problems and provided access to
> a class of computing and control technology. If another generation of
> devices, or another class of technology, comes around - why not see if
> THOSE solutions will work for those NEW devices and technology? Or see
> if the same problems were already solved, or not solved - learn from
> our mistakes and successes?
>
> I don't expect personal computer collectors to have these goals and
> these considerations; they have personal reasons and interests. But
> some of us have this perspective.
>
> So - if you aren't a grad student or scholar, which robot is going to
> teach you something? the robot with the operating system and
> message-passing protocols and pre-established methods? Or the robot
> that you say "move this joint motor" with "peek(12379)" in a FOR loop?
>
> You can argue you are peeing away your time with primitive coding. But
> I'd argue you are pissing away what you might learn about control and
> sensor problems, if you only use a graphics interface and script
> messages about tasks. Not much difference between some game-design
> avatar and a robot, at that level.
>
> 20th century versus 21st century. Engineering design versus product
> design. These are old tensions. And as they will occur again, this old
> stuff is worth preserving, to show how we dealt with these conflicts,
> the last few times they occurred.
>
> Herb Johnson
> retrotechnology.com
>
Great post, Herb, on all accounts.
I had to look at ROS recently and I had the same reaction to their web
site in trying to glean
what was all about. I'm convinced its not us, their presentation is flawed.
More information about the vcf-midatlantic
mailing list