Statistics on DARPAGC
Randy M. Dumse
GC. I wonder if there is any statistical support we might measure my
contentions with.
I'll just throw out an idea if someone wants to work on it. My
impression was this newsgroup was much more active and vital before we
got sidetracked with the DARPAGC. I hope it will again become more vital
now that it has passed.
So here's the idea. Suppose there's a way to check post per day prior to
the GC and again after it started? Can we again track it afterwards? I'm
pretty sure we could use Google statistics to get this raw information.
Then it would be nice if we could extract the GC related posts, to see
them as their own category. Particularly in the time right around the
races themselves, and the occassional controversies prior to the first
race.
I'm just guessing statistically you'll find there were more posts in the
year before GC, than the year afterward, and even less the year
following that.
--
Randy M. Dumse
Caution: Objects in mirror are more confused than they appear.
Randy M. Dumse
news:meQ5f.103$EP6...@eagle.america.net...
Okay, just a quick taste.
I searched Google for comp.robotics.misc for any post with a " " in it
(should be all) between Jan 1 and Dec 31 for the past 6 years. Here are
the results:
1999 14,900
2000 18,400
2001 11,900
2002 11,700
2003 8,200
2004 7,550
and so far ~10/12th of the way through this year
2005 5,880
Folks, we're dying as a group. Something around 2003 seems to have
injured interest in talking about robotics. Curiously, the original
DARPA press release is dated January 2, 2003.
http://www.darpa.mil/grandchallenge04/media/announcement.pdf
Doing some more research, http://robots.net/article/587.html ,
battlebots was cancelled in late September 2002.
If you have another theory why robotics interest dried up so sharply,
I'd be curious to hear it.
newtype
@eagle.america.net:
A novelty for younger people, well got to go to school now, have to get a
job, pay bills, have to give up expensive hobbies.
t...@robotcrazy.com
in occasionally and lurk around to see what the latest news is. This
used to be a very positive NG.
I'm no longer involved in hobby robotics; I'm personally concentrating
on my career as a software developer. I fully intend to return when I
have more time.... Although I'll always be searching for the "what's
the application/requirements", "where is the need" etc.. Also, I'd
need to convince myself that what I wanted to achieve couldnt be done
with pure simulation at a fraction of the cost.
My theory is that there was an idea that the hobbyist in a workshop
could come up with something revolutionary and worthwhile. The well
publicised 'system' breakthroughs are fantastically complex and beyond
the means of a single hobbyist.
You just need to look at the DARPA competition vehicles, ASIMO, QRIO
and appreciate their complexity to realize that its just hopeless
trying....for now. As the technology develops and becomes more
wide-spread and understood with industrial standards, perhaps there
will be scope for hobbyists to adapt off the shelf hardware.
IMHO of course.
Regards
Tim
Brian Dean
> 1999 14,900
> 2000 18,400
> 2001 11,900
> 2002 11,700
> 2003 8,200
> 2004 7,550
>
> and so far ~10/12th of the way through this year
>
> 2005 5,880
>
> Folks, we're dying as a group. Something around 2003 seems to have
> injured interest in talking about robotics. Curiously, the original
> DARPA press release is dated January 2, 2003.
Looks like the drop started much earlier, Randy, with a 35% drop from
2000 to 2001. There was only a 29% drop during from 2002 to 2003 and
only an 8% drop from 2003 to 2004.
How do you explain the largest drop from 2000 to 2001?
> Doing some more research, http://robots.net/article/587.html ,
> battlebots was cancelled in late September 2002.
While many robotics purists look down on battlebots, it is a great way
to generate interest in kids especially. I got my kids a couple of
Scooterbots from Gordon's BudgetRobotics store one year as a Holiday
gift and the first thing they did was tape a bunch of sharp objects to
them and crash them into each other :-) I'm not even sure they ever
watched battlebots.
> If you have another theory why robotics interest dried up so
> sharply, I'd be curious to hear it.
Interest in robots seems to go in cycles, usually coinciding with
enabling technology - low cost MCU processors, advances in light
weight high energy batteries, etc. I do think the TV show battlebots
was a huge boon. I personally don't think the DGC sapped us of all
the best minds from robotics ... you don't really think they hang out
on USENET anyway, do you? :-)
-Brian
--
Brian Dean
ATmega128 based MAVRIC controllers
http://www.bdmicro.com/
blueeyedpop
Back in the early 90's, Usenet was a great place for everyone to get
together. Now with various clubs, forums, etc. I feel that the ratio of
meeting places to interested people has become unbalenced, leaving us with
less interested people in any one place.
Imagine a town with 100,000 people, and 100 drinkers. Open up 10 bars, and
you will have 10 patrons per bar, and none fairing all that well.
If all of these robotics resources were commercial ventures, most would have
gone out of business, and the best would have survived. Instead we have the
opposite, we have banded into micro cliques of similarly interested people,
generally divided by processor preference.
Remember back when it was 8051 vs 6502? Now we have AVR, ARM, PIC,
Motorola, Intel, and so forth. While the professional software engineer
sticks with what he knows out of convenience, popping from one processor to
another or language to language is a non issue. In the personal robotics
realm, processor/language wars can take on religious fervor. This divides
us, when in actuality among professionals, it is a trivial matter to
migrate.
To adapt an old addage "A place for every processor, and every processor in
its place." It is a truism, but ignoring it divides us, in my opinion.
It really is a sort of Zen thing, Fighting for what you believe in, yet
acknowledging other views. I have been guilty of nearsightedness as well. I
truly believe that IsoPods are a remarkable product. When I found that my
new employers were big IsoPod consumers, it warmed the cockels of my little
heart. None the less, one must acknowledge the strengths and weaknesses of
all processors, languages and development enviornments.
Pondering on all this, reflecting, and re-reading my words, I think to
myself that despite my altruistic intentions, you would have to pry my
IsoMax based processors, and my new love, Python, from my cold dead hands.
Despite this, at work I am happy to drop into Java, C, C++, PICs, ARMs, AVRs
if I must. Another day, another language, another paycheck. It is a simple
matter that on my own time, I like to get things done quickly.
(am I getting silly here?)
My only solution to this is to teach programming for the sake of
programming. Maybe I am just tired, but it seems like we need to teach our
next generation to program in a simple useless language on a simple
relitively useless box. Something with some LEDs, buttons, keyboard, and
display. A language that allows one to use a stack, register calls, function
calls, and a variety of maths.
After you learn programming concepts out of the context of a real language,
then you learn real languages where you now know about program flow and
proper programming constructs, and now concentrate on syntax, proper
methodologies, and efficient programming for what ever language it is you
are learning.
It seems to me that teaching people to be non processor biased may one of
the keys to helping to promote unity. It does not cure diffusion completely,
but I think it goes a long way.
Any thoughts?
"Randy M. Dumse" <r...@newmicros.com> wrote in message
news:adS5f.112$EP6...@eagle.america.net...
Randy M. Dumse
news:2005102102...@bdmicro.com...
> Looks like the drop started much earlier, Randy, with a 35% drop
> from 2000 to 2001. There was only a 29% drop during from
> 2002 to 2003 and only an 8% drop from 2003 to 2004.
>
> How do you explain the largest drop from 2000 to 2001?
Well, that is an interesting point, however, you see the previous year
was a rise, ~25% and if we start from there, 2001 is only a ~25% fall,
so do we see 2000 as significant peak, or an anomoly.
Maybe someone could write a Kalman filter for it.
The question might be better asked why was 2000 so hot. And lest we
forget, 9/11 2001 after math seemed to put lots of robot guys on the
streets.
> Interest in robots seems to go in cycles, usually coinciding with
> enabling technology - low cost MCU processors, advances in light
> weight high energy batteries, etc. I do think the TV show
> battlebots was a huge boon.
Me too.
Yes, I think it goes in cycles, and yes I think it follows enabling
technology. I guess I'm saying I thought we'd have more enabling
technology the past couple years, but new stuff has been rather thin.
Not much in the way of new sensors specifically. Think some of my new
products might be targeted in that direction in the future.
> I personally don't think the DGC sapped us of all
> the best minds from robotics ... you don't really think they hang out
> on USENET anyway, do you? :-)
Ah... guess that was my assumption. You're here!
John Nagle participates.
Marvin Minsky has posted here. Joe Jones has dropped in. I think Ron
Arkin posted here about his summer seminars. Gordon is here. Quite a few
other authors, like Dennis and Mike, and Mike and D Jay. And a host of
other swell fellows. Who else do you think we need? But seriously, these
guys usually respond more often than not, rather than starting new
posts. So that's why I think its not their lack of participation, but
the general level of interest that is falling.
BTW, what happened to Paul Grayson, and the ... IRRF was it? Web site is
there, but I can't find any activity after late 2004.
Why doesn't Michael Simpson post any more of his neat little app note
links?
Sergio had some exciting links he posted a while ago. Intersting stuff.
Those I thought had the feel of the good old days.
Matthias Melcher
> You just need to look at the DARPA competition vehicles, ASIMO, QRIO
> and appreciate their complexity to realize that its just hopeless
> trying....for now.
Nonono, the opposite is true. Sure, the Qrio is insanely complex being
created by a bunch of highly paid scientists and engineers, but at the
same time, R/C servos have become dirt cheap, BlueTooth is a reliable
and affordbable wireless solution, and "Sugar Cube" cameras are
available under $30. Put all that together, using your PC as the main
brain, and you get a Qrio style walking talking living robot for under $500.
Four years ago, the servos alone would have cost over a $1000.
This is also the time where robots start to du useful things. You can
analyse the video picture on a PC fast enough now to recognize faces, to
find your old socks on the ground and to finally have robot *do*
something (other than following a line and bumping into chairs).
Brian Dean
> > I personally don't think the DGC sapped us of all
> > the best minds from robotics ... you don't really think they hang out
> > on USENET anyway, do you? :-)
>
> Ah... guess that was my assumption. You're here!
>
> John Nagle participates.
>
> Marvin Minsky has posted here. Joe Jones has dropped in. I think Ron
> Arkin posted here about his summer seminars. Gordon is here. Quite a few
> other authors, like Dennis and Mike, and Mike and D Jay. And a host of
> other swell fellows. Who else do you think we need? But seriously, these
> guys usually respond more often than not, rather than starting new
> posts. So that's why I think its not their lack of participation, but
> the general level of interest that is falling.
>
> BTW, what happened to Paul Grayson, and the ... IRRF was it? Web site is
> there, but I can't find any activity after late 2004.
>
> Why doesn't Michael Simpson post any more of his neat little app note
> links?
>
> Sergio had some exciting links he posted a while ago. Intersting stuff.
> Those I thought had the feel of the good old days.
Don't forget yourself, Mike Keesling, D. Jay Newman Dennis Clarke,
Michael Owings, to mention only a few more. And we can't forget MLW
(who seems notably missing lately)!
My previous post ended with a smiley so I was only half kidding on
that last comment. But also there are tremendous number of folks that
don't post here. AFAIK, only John Nagle and I are the only DGC team
members that post here - I could be wrong about that. I know I posted
here before, during, and after the DGC. I know John did during and
after, not sure about before.
Padu
> My personal take is that it is a matter of diffusion.
>
That is exactly what I think too. I am a newbie in robotics, started last
year (well, you guys can see from the types of question I ask here) and
besides this forum, I participate/lurk on at least 10 more (don't let my
employer know 8^)
So, if I have a question pertaining to programming, I will ask on the
language bulleting board, so on and so forth.
It doesn't mean that I'm not interested in robotics, it only means that my
attention is divided among many forums. Also, try to measure the level of
interest for robotics having the only reference this ng is crazyness, most
of the folks on my lab don't participate to this forum, so statistically it
is not a good indicator of serious robotic activity.
Cheers
Padu
Chris Spencer
> t...@DELETEMErobotcrazy.com wrote:
>
>> You just need to look at the DARPA competition vehicles, ASIMO, QRIO
>> and appreciate their complexity to realize that its just hopeless
>> trying....for now.
>
>
> Nonono, the opposite is true. Sure, the Qrio is insanely complex being
> created by a bunch of highly paid scientists and engineers, but at the
> same time, R/C servos have become dirt cheap, BlueTooth is a reliable
> and affordbable wireless solution, and "Sugar Cube" cameras are
> available under $30. Put all that together, using your PC as the main
> brain, and you get a Qrio style walking talking living robot for under
> $500.
I think you're exaggerating here a little. For instance, while standard
R/C servos are quite inexpensive, they're also relatively useless for
robotics, especially bipedal robots. R/C servos are typically weak,
prone to breakage, and void of any built-in sensor feedback. A few Asian
companies sell servos that address these problems, but they lose their
"dirt cheapness" in the process.
Lynxmotion, best known for its walking robot kits, uses the more
expensive digital servos for its bipedal kit. Oh, but the kit still
requires a tether because even digital servos aren't strong enough to
carry a meager load. They recommend even more expensive servos if you
want their mobile kit to be, well, mobile, but now you're easily talking
over $2000 in parts, and that's only for basic locomotion. We haven't
even touched upon gyros, tactile sensors, cameras, and the complicated
software required to get the thing doing anything remotely useful.
While I agree with your point, that a lot can be done with the
technology currently available to hobbyists, the "interesting" stuff is
still well outside our reach.
It's not too surprising Randy's numbers coincide with the burst of the
Dot-com bubble. It seems many have come to realize the field of robotics
is still far outpaced by even the most humble imaginations.
Chris
D. Jay Newman
> My personal take is that it is a matter of diffusion.
>
> Back in the early 90's, Usenet was a great place for everyone to get
> together. Now with various clubs, forums, etc. I feel that the ratio of
> meeting places to interested people has become unbalenced, leaving us with
> less interested people in any one place.
Interesting. I think that you have something here.
I don't think that it's everything, but it is something. There are a
lot of robotic forums out there as well as this one. My favorite
theory is that with the demise of robotic combat on TV young people
have less exposure to "fun" robots. Perhaps RoboSapiens and company
can reverse this.
> If all of these robotics resources were commercial ventures, most would have
> gone out of business, and the best would have survived. Instead we have the
> opposite, we have banded into micro cliques of similarly interested people,
> generally divided by processor preference.
I don't know about that. The processor people do tend to hang out around
their own forums, but they also hang out on robotics forums if they
are robotics people.
--
D. Jay Newman
http://enerd.ws/robots/
blueeyedpop
processor. It makes sense from a learning point of view. THere are other
online groups like megabitty, that are processor intensive. I recall there
were others, but can't recall what they were.
I think movies may have hurt robotics to some degree, in that one can simply
never build R2D2 functionality in the scope of what we are capable of...
That is a bit of a reach, but who knows.
Mike
"D. Jay Newman" <j...@sprucegrove.com> wrote in message
news:djceqk$15ba$1...@f04n12.cac.psu.edu...
makc.th...@gmail.com
blueeyedpop wrote:
> I think movies may have hurt robotics to some degree, in that one can simply
> never build R2D2 functionality in the scope of what we are capable of...
Yep, but if someone would com up and say, "here, now you can do it,
like this", how many guys you think would start doing it right away?
Mike Young
The big jump was 2000 to 2001. The dot-com bust was responsible for that,
one can presume. DARPA-GC doesn't strike me as a societal event of even
close to the same magnitude. That's just me speaking as a layperson. I'm not
the slightest way involved with the -GC; it had zero impact on my life. My
wife's and neighbor's also, come to think of it. And I don't recall seeing
even a passing mention on the news, not that that means anything.
Other explanations abound.
Prolifereation of webboards, and the general dimishment of usenet. All
newsgroups are down compared to 5 years ago.
Stagnation of technology. Hard to believe, isn't it? In the end, what can
the average person do that he couldn't do 5 years ago, and how exciting is
it? It's not the technology that stagnated, but certainly the level of
technology easily grasped did. DIP PICs and RC servos are accessible.
Multi-GHz dsp is not. Maybe you don't see it, but the barriers to entry
there are insurmountable to all but very few. And who has time or incentive
to use it to reinvent, what?, boring old battlebots? I think I'd summarize
this is lack of relevance and access for the common denominator.
2000'ish also coincides with widespread outsourcing of tech jobs to SE Asia.
College enrollment is down. Stateside tech careers are reputedly scarce and
dwindling. Also, the promised labor savings of tech investments are starting
to pay off. The end result is a smaller labor force. Tech jobs are going,
and menial jobs are gone. What's left? Who's your audience?
A general shift in economics. We had a good run of it, didn't we? More than
ten years of double digit growth on investments. We were rich. Now we're
poor, and getting poorer. Money is scarcer, and we channel it differently.
High tech gadgets aren't the big line item in my budget they were once. Is
yours any different? There was a time when half, or more, of some store's
shelf space pandered to PC games. Not anymore.
Last, as much as we like to diminish the role of the country's leadership,
GW has been nothing but a big blanket smothering growth. I can't quite put
my finger on it, but it's widespread. Almost like it comes into every home
the same way, and smothers hope and growth the same way. Almost as though
CNN was the new brainwash channel. Scary, isn't it?
Mike Young
> "Randy M. Dumse" <r...@newmicros.com> wrote in message
> news:adS5f.112$EP6...@eagle.america.net...
> the average person do that he couldn't do 5 years ago, and how exciting is
> it? It's not the technology that stagnated, but certainly the level of
> technology easily grasped did. DIP PICs and RC servos are accessible.
> Multi-GHz dsp is not. Maybe you don't see it, but the barriers to entry
> there are insurmountable to all but very few. And who has time or
> incentive to use it to reinvent, what?, boring old battlebots? I think I'd
> summarize this is lack of relevance and access for the common denominator.
Also, I want to add that Randy is right to focus on the sensors. Solving the
machine vision problem and making it accessible would open the floodgates
wide in ways unimaginable, not just by us tinkerers, but in every segment of
industry. In the meantime, our 'bots are deaf, dumb, and blind. Deaf for
essentially the same reasons as blind. Voice recognition software on PCs are
gadgets rather than essentials. 'Bots, with very few exceptions use sound
only as a crude replacement for the missing eyes. I heard of very few that
respond to environment sounds in a way that resembles intelligence. Dumb,
simply because it has very little to say. It means both mute, and the other
meaning of dumb.
Machine vision and environment sound recognition are the problems to solve
now. This is breaking new ground, not just for the hobbyist, but in industry
in general. It is more of a software and conceptual problem than it is of
motion or device control, which are what most people think when they hear
'robotics'. So, you see, the forefront is not here; it's in the software and
science groups. Our newsgroup traffic is down because simply spinning around
and flashing a few lights is old hat, not very challenging.
Last, I think it likely the vision part is solvable with today's processors,
and might already have been solved. It would be of such strategic
importance, though, that we don't have access to it. What if DGC was not so
much to harvest our thoughts and ideas, but really to gauge the civilian
state of the art in environment recognition? It would be just like those
sneaky bastards... ;-P