the crypt

It's definitely quieter -- maybe more like a mortuary rather than a crypt. But I have noticed the same slowdown in message traffic in most other forums and groups, too.

Interest in robotics comes and goes. It tends to peak every 5-7 years. We had a pretty good run in 1999 to about 2001 or 2002, but things have gotten progressively slower since. I expect things to pick up again in a year or two. However, much of this depends on the introduction of new and interesting low-cost products for building bots -- microcontrollers, sensors, whatever. A $75 vision system would get a lot of people back into the game, for example.

I think there's also the matter of people adding certain posters to their kill file. Not you but others. And it being the weekend, a weekend of college bowl games, etc. (For that matter, don't even think about getting many hits to the site next Sunday.)

I can say I didn't visit the sites because A) I've already visited your sites (on numerous occasions), and B) PID math isn't much of an interest of mine. But just to skew your logs I just visited all of the ones you posted!

-- Gordon

The UserFriendly cartoon

had an interesting clip that drew the conclusion that blogs are the new usenet.

Other than that, c.r.m is sort of esoteric. While you may not have appreciated my post on PID (I think we have philosophical differences), it is one of the things that I think this group needs.

Right now, it feels more like a "support group," as people post questions and hope for answers. Because of this, people won't come here to see new stuff.

I also sense, and it is hard to articulate, it is like disrespect. (everyone is guilty on some level) Sure this is usenet, but look at the thread for my PID essay. I posted a link to something I thought would make good reading for someone new or just beginning. I was asking for critique. Now there was some useful criticism, and some off topic criticism like colored comments. Those aren't bad, then there was yours, which at the very beginning seemed to imply that I hadn't done what I wrote about or at the very least attempted to minimize it, the thread quickly got out of hand. There is clearly no incentive now to write about how to hack up a simple ball mouse and make it a dual motor encoder or the other aspects of the $500 robot.

Building and maintaining a community takes, if not work, at least activity. I love robotics, I built my first robot in (I think I remember) 1978, based on an old BigTrac and a COSMAC 1802, maybe it is like Heathkit, no longer generally interesting.

People need to post, to share, to create and to feel encouraged to do so.

dpa,

I wish I had useful analysis or a theory to explain what's going on. I've felt the same thing for a long time. I do have some general statistics that might spark some discussion.

First, I'd like to remind everyone that back on October 20, Randy Dumse did some Google analysis and posted some statistics that show an decline in the number of postings on comp.robotics.misc.

1999 14,900 2000 18,400 2001 11,900 2002 11,700 2003 8,200 2004 7,550 2005 5880 (in october, projected to about 7000 by end of year)

Recently, "mlw" suggested that this might be due to the emergence of blogs. I don't have a good measure, myself, but I'm not aware of any particularly lively blogs that might be drawing off participation in general robotics discussions. Sure there are subject-matter specific groups, but these have always been around in one form or another.

Also, looking at the Seattle Robotics Society and the number of "Encoders" they published

2000 11 2001 8 2002 3 2003 2 2001 1

Publications like the Encoder "run their course" and it isn't unreasonable for them to taper off as the creative personnel move on to other challenges. Still, the Encoder was a wonderful resource and I don't see anything else that's moved in to take over its role.

I also miss the NASA "Cool Robot of the Week". I believe that a couple of your bots got featured there.

So if anyone wonders if there is actually a trend, it looks like there is evidence in a lot of different areas to support that hypothesis.

Finally, you reported that you saw relatively few hits when your pages were referenced in mlw's recent PID article. I don't know what to make of that, but I do have one alternate observation to offer. At the end of November I announced that I'd put up a pair of articles titled "Calculations Useful for Robotics" at the Rossum Project web site. Now, I'm not entirely happy with the way the articles turned out, and they were rather specialized pages with a limited audience. Even so, I saw my web visit volume more than double for the two days surrounding the post. So maybe the reason that the article didn't steer more hits toward your page (which, by the way, I have visited and enjoyed many times) was just that it just wasn't set up to do so.

Well, that's all I've got. I hope someone can make sense of all this.

Gary

P.S. For anyone who's interested, I believe that you can see DPA's robots at

I got him confused with someone else).

dpa wrote:

Interesting numbers. Both Randy and David are members of DPRG, which has/had a very active mailing list, so I'm sure they're already aware of the phenomenon there, too. The DPRG archives make it easy to see they've had a similar slowdown. Go back year-to-year for the same month, and you'll see maybe a three or four-fold decrease from say, 2000, to today

-- a proportion similar to the figures above. In fact, Dave started many of the DPRG threads for this month!

Amateur robotics has had its doldrums before. There was a big one in the early 1990s. Then came the BASIC Stamp, and love it or hate it, this one product turned robotics around. I don't imagine they were the first (or maybe they were; I don't know the exact history), all I know is that it sparked a resurgence of interest.

In the late-1990s we had a big boost. There was Mindstorms, the Sharp sensors, the ADXL line from Analog, and several other enabling technologies. This is a segment largely driven by cheap, innovative commercial products. What's out there yet to be developed...the $5 encoder, the $75 vision system? When we have a couple more of these you'll see a return to the "old days."

-- Gordon

I wonder if there is some correlation between interest in robotics and the release of StarWars movies? Seems like about the same time frame.

That makes me curious if there is perhaps some public disappointment that we're not closer to C3PO and R2D2?

I know I'm happy when I can get a PID routine to behave, or make a robot balance on two wheels, but that may seem pretty lame to folks who's inspiration and imagination have been fired by an exciting sci-fi movie.

On the other hand, I got to give a presentation to about 250

4th and 5th graders last week, and afterward they just mobbed me! Their enthusiasm was really infectious, and all the robots really did was just dart around and avoid them. That seemed to be enough.

I got a fascinating thankyou from their science teacher, who thanked me for "letting the children get so close to your robots."

Might be something to ponder.

best regards, dpa

Gord> It's definitely quieter -- maybe more like a

I think robots are of little interest because essentially most of them don't do much considering their high cost. They are just a fun application for newbies in electronics and for engineers to show off their know how.

-- JC

That might be true if there wasn't any recent interest, but that's not the case. We're talking about how groups like this one aren't as busy as they used to be.

Besides, the group was busier when robots were *more* expensive than they are today. Under a certain threshhold cost is not a major consideration, as long as there's something new. What's missing from your theory is that there hasn't been anything radically new for a while.

-- Gordon

Gord> That might be true if there wasn't any recent

Yes I was a bit off topic. I had something else in mind. Sorry.

-- JC

I read this and my first thought was "grrrrrr... no! wrong! robots good! robots good!"

Then I started wondering if maybe you're right. So that made me think, what exactly _can_ our robots do? Hmmmm.

I started to make a list of what hobbiests' experimental autonomous mobile robots can do, now. It's not so easy. It has to be what they can really do, now. That's what people see, that's what they are attracted to.

This is what I've come up with so far:

1. Maneuver.

Move around without getting stuck or hitting things.

1. Navigate.

Go somewhere.

This would have to include everything like line following, odometery, wall following, GPS following, maze running, etc.

1. Acquire a target.

candle flames, ir beacons, bright lights, orange traffic cones. maybe this is also navigation?

1. Pick and place something.

Grip or push things from one place to another.

1. Sumo.

Seems like it needs a class to itself. Or is it just pick and place?

1. Soccer.

likewise. Or just team pick and place?

There are surely more but it seems like I can fit all the things I've seen various hobby autonomous mobile robots do in one of these categories. At least all I can think of at 2am.

But maybe this is really just 2 basic capabilities:

1. Navigate (includes maneuver and acquire)

1. Pick and Place. (includes sumo and soccer)

Is that right? We've built a bunch of mobile pick and place machines? I need to sleep on this.

best, dpa

What do we do that is different? Are we a bunch of mobile pick and place machines? The common riposte to the question what is the use of xxx is, "What use is a newborn baby?"

One use they have that you left off your list is,

1. Entertainment.

They probably also act to inspire young engineers to the future possibilities for automation.

-- JC

There is certainly some interest - there are about 1000 teams entering the FIRST robotics competition this year - and yes, many of them will be little more than RC boxes, but they will be 50kg boxes.

The challenge this year includes aiming and firing balls at a target using CMUCam for aiming. Making all this sort of stuff work in the 6 weeks that they have is really demanding - many teams do not have 'proper' workshops - my local lot use someones garage.

So there is still plenty of interest - I suspect that the problem is more specific to this group - not enough interesting posts means fewer people reading and posting and so on.

Look at

for action !

Dave

Hi,

I have two people in my kill file. Each goes by three initials, neither is DPA...

In my opinion, usenet was necessary when there was no other online resource. Back in the 90's I had my "site without a purpose" and posted piccies of robots for people without web hosting. It was unique then, but now it would be just that, a site without a purpose, in that it offers nothing unique, but that is the curse of the internet. That *.wmv file that a friend sends, that you have already seen a dozen times, that electronics forum made by two brothers in some hideously cold climate, that they think is ultra cool, whatever it is, it is tough to be unique. The internet has enabled the dissemination of information. Not good information, not useful information, just information in what ever form.

Now, once month, someone comes through with an announcement for a great robotics site. Now there are a variety of different forums. (google offers up 25 pages of "robotics forum" links) Now there are user groups for a variery of processors. ( google offers 80 pages of "basic stamp", 78 pages of "pic microcontroller robot",)

The cohesive force is "google". Search CRM for "google", I get 6270 hits ( until google threads this post...), many of which are something to the effect of "don't post your assinine question here, google it". Google in essence is a binding force.

Robotics has bifurcated several times over into a variery of different types of robotics domains. Walkers 3+ legs, bipeds, rollers, combat robotics large, medium, small, sumo, mini sumo, micro sumo, lego, solar, and so forth. Each with a unique following, some with crossover, but all clustering interests.

Interest and commonality gather people together. Everyone on C.R.M. has one thing in common, robotics. Everyone in a robotics club has at least two things in common, location and robotics. Additionally, clubs often have common useful tools, common processors and boards amongst the members, and websites with forums for their members and whomever else joins. For these people, C.R.M. holds little value if they have an expert. Robotics is alive and moderately well-ish, more clubs are forming than dying, as far as I can tell, but diffusion is killing usenet.

So, in short, I believe diffusion of information, and clustering of like minded people and information is killing usenet.

Well, it's not just robotics. For many years I was an avid reader of comp.graphics and its subsequent subgroups. Early on, there was a plethora of good activity, spawned in part by the lack of adequate textbooks and information. Back then, comp.graphics formed the center of a community of computer graphics people who were interested to learn and share information.

Now, it too, is basically a wasteland.

It's not that there are fewer graphics wizards out there, or even that they are any less motivated to share their knowledge with others: it's simply that there are different, better forums for this kind of exchange than newsgroups. Let's face it, most of the questions asked on groups like this have been asked and answered many times before. But the netnews mechanisms simply do not provide the kinds of search capabilities that would enable people to weed out these redundant threads. The web, for all its many problems, does. Yes, it's less conversational, but it can be more informative, and combined with the ability to display graphics and video media, it simply becomes the preferred way to talk about things.

Mark

But isn't this what Google and similar usenet archives do? I regularly check usenet posts through Google when I'm researching. Some people use Google (or another Web-usenet gateway) to post usenet messages.

In Google's case, the messages can appear threaded, and the search mechanism is far better than what you'd find with phpBB, UBB, or their ink. So I'm not sure this is the reason for the decline in usenet posts. I rather think there is a decline *everywhere*, probably because there are more resources to choose from, and only a finite set of eyes. In another post Mike hits it on the head: there is a bifurcation of robotics and people are gravitating toward the resources that best fit their needs. C.r.m. is a general purpose group.

-- Gordon

If we could get a robot to do something truly useful then the large manufacturing companies, always looking for new markets, certainly could. The more robotics becomes a commercial venture the less interest there will be in hobbyists building them.

Why do people build and fly model airplanes? I don't imagine it's because one day they want to build a larger airplane and fly in it; they simply like to build things, they like the learning process, and they like the challenge. They may also like the social aspect of meeting other likeminded individuals, sharing ideas, and enterting contests.

Yet another valid purpose for robot building: Robotics is a unique blend of disciplines that covers three major vocational areas: electronics, programming, and mechanics. That makes it an ideal teaching tool whose usefulness reaches far beyond what the machine actually does. As you've found, kids are thrilled just watching these things move. It's the adult that asks, "What does it do?" And if some little robot gets a kid to take engineering or programming or electronics courses, who's to say the little robot didn't have a useful purpose.

-- Gordon

--The killer app that would be good for this group and for many other suddenly-low-traffic usenet groups, would be something like tribe.net, but with usenet "style"; i.e. something that shows threading like shell, etc. Trouble is, I don't see anything like this on the horizon. If it existed it would be possible to "migrate" this group's traffic over there, then we'd pick up all (well, most) of the people who no longer have, or no longer want to bother with shell access. Just a thought..

--IMO the trouble is that *none* of these new forums (fora?) are as easy to use or to read as usenet. I've still got a jones for shell (using PuTTy and tin), because it shows things you don't get to see in web-based communities, like degree of divergence of thread. And of course there's the wonderful killfile, heh. On the web if it ain't moderated you're subjected to every jerk's whims which really messes with the signal to noise ratio. Bottom line: I really think that the robotics world would find, and flock to a web-based forum that had usenet-like utility, but so far...

What would you want in a vision system? Do you have a feature wishlist in mind?

My wife's cellphone has a 1.3 megapixel camera that is about the size of the eraser on the end of a pencil. It can take still photos, and 30fps movies.

I would love to have one of those for a robot, but

1.3 Mpix x 30fps is a lot of data. I think the only way to deliver that much data would be to use something like high-speed USB (480 Mbps). Not many microcontroller based robots are going to be able to handle that. The minimum system would be maybe a mini-itx.

Intel has announced a new ARM Xscale processor, the PXA271, that can interface directly to a CMOS camera. It runs at 400MHz, has 32MB of RAM and 32MB of flash on chip. It should be able to run Linux with no additional components other than a crystal. It has the integer XMM SIMD instructions, which should give it enough power to do image processing tasks like edge-detection. It is clearly designed for PDAs and cellphones, but it looks like it could be a dandy robot controller with integrated vision. I don't think it is actually shipping yet, and I don't know what the price will be.

Just delivering the data is not enough, to make a vision system, you gotta do something with it as well. My robot is ITX based, and I have been pondering a low-cost vision system for it.

1.3Mpix is pretty high, 1280x1024 pixels is a lot of processing, hell 640x480 is a lot of processing, I would be satisfied with 324x240 at 30fps *and* be able to do something.