From that "trouble with hobbyist-level robot hardware" thread, I surfed over to newmicros.com to see what one of the posters (apparently the founder of New Micros) was talking about. I see that I'd been here before, but hadn't stuck around long. What follows is not a criticism of New Micros in particular; I just use them as a springboard for some general comments that may be helpful to robotics companies looking to expand their access to the casual-robotics market.
First, too much information can be as problematic as too little. A feature matrix comparing controller boards, for example, can be rather bewildering to a new user. What is a "JTAG interface," and do I need one? How will a "watchdog" help me -- I assume it's not something that barks when somebody tries to break into my shop? What is "CAN" and should I care? I'm sure this is great technical info for somebody who already knows all this stuff, but to a newbie, it just says, "Go away, this is too complex for you."
What would help here is a link -- preferably near the top of such a page
-- to the effect of "New to robotics? See our beginner's guide!" which links to a page that avoids jargon, makes some reasonable assumptions about what a newbie needs to get started (see below), and presents a small number of options, explaining the differences between them in plain English, leaving out details that a newbie doesn't really need to know.
Another example of "too much information" -- say I like the looks of the TiniAVR board, because it's fairly cheap and because the short blurb says it is "the perfect board for robotic applications." So I click through to the description, which says it is intended to be plugged into a user-designed carrier board (i.e. go away kid, this is WAY too hard for you), and then launches into a general discussion about two-layer and multilayer boards, and how the latter remain prohibitively expensive. Seems like a gross digression -- but surely there's some reason that discussion is related to the TiniAVR. I just can't see it.
Next, even with all that information available, it's still not clear to me what I can actually do with the dang things. This site has about a dozen microcontroller boards, at least several of which claim to be perfect for robotics applications. I can imagine that any of them could, in principle, be used to make a simple robot. But where are the examples? I'm looking for actual projects that made good use of these devices, so I can see what's possible. It's much easier for me to look over half a dozen robots and say, "hey, that's similar to what I want," and read about how it was made, than to wade through dozens of pages of specs and try to figure out which is the one I should use. I guess this goes back to the "easy beginner's guide" idea.
Finally, for a real newbie, starting with individual boards is just too much of a hurdle. I know, if I had enough time, I could figure out what I need, buying a controller from here, a power supply from there, motors from someplace else, making or buying a chassis elsewhere still, and figuring out how to make it all work together. But I don't have that much time; I have kids to raise and a full-time job. So until I get more experience, I'm going to stick to kits that contain everything I need, including directions. Here are the kit characteristics that I think will hook a newbie:
- complete (this is a must)
- well-written directions, with photos, ideally available for download
- programmable (ideally, with any serial-capable computer)
- obvious growth path (suggested ways to swap in a more powerful controller, or add additional sensors/servos/etc.)
- low-cost (certainly no more than $100; under $50 would be better)
Note that it doesn't have to do very much -- I know you can't expect much for $50. But if it doesn't do much, it should be possible to either extend the kit itself, or take it apart and reuse the parts for other things. (This is the main problem with the OWI robot kits, for example; they can't be easily extended or reused in other projects.)
The "programmable" requirement could be relaxed in the base kit, provided there was a documented, fairly painless path to add in a programmable controller later. It's much easier for me to spend $60 now and $60 again a few months later, than it is for me to spend $100 all at once. But it's also important that when I spend in smaller chunks, I feel like I'm building up the capability to build more advanced bots, rather than just collecting a bunch of separate, useless toys.
I really think this is the key to the newbie market -- to have a suitably low entry barrier, in terms of both cost and required skills. Once they have a successful project under their belt, they'll come back for the more advanced stuff, and keep coming back provided they keep having success. But if that initial entry is too steep, there's just no way for them to get started with your product line.
HTH,
- Joe