We've been talking about community projects and standardizing an open source robot platform for awhile now. I really like/support these ideas, but something has always prevented the ideas from going further . . .
So this thread is about discussing not what these standards are, but would could prevent these standards from catching on. If this is taken into account, perhaps the standards we decide on in the future will stick.
Here are some ideas of what we need to overcome . . .
For example . . . lets say I was a company producing X robot product. And lets say that modifying the product to be standards compatible will cost an extra $1/board. Doesn't sound like much, unless you make a thousand units. Will spending an extra $1k be worth it for the company? The point I'm trying to make is that a standard has to account for mass manufacture costs, and industry opinion, if we want industry to join in.
However, lets say we don't care about industry and just make it open source standardization. That $1 more is nothing to a hobbyist if it allows all these units to fit together without much effort.
Stackable units are a great idea. But . . . a microcontroller will inevitably be much bigger than say a simple accelerometer sensor. Also, unit positioning is sometimes more important than stacking. For example some sensors are sensitive around other electronics (compass next to a motor driver, for example), and a gyro must be located at the very center of a robot. Stackable doesn't really work if a robot needs to be flat, such as for robot planes. I think stacking should be a standard for only units that don't care if they are stacked, and are about the same size (like a motor driver and a microcontroller, for example). I think its much more important to have wire connector standards (think of the servo wire standard).
Screw standards I think are really important. I dislike having different screws, of different lengths, different diameters, different head types, etc. But you also need to account for robot size with this standard. We all seem to have agreed on 4-40 (and the metric equivalent) in terms of diameter. But we also need an agreed upon screw length(s) and head-type, otherwise you'd still need 8+ sets of different 4-40 screws.
In terms of industry, any intelligent manufacturer will try to use as few screw types as possible. Only one screw type, ideally. You don't want a pick and place machine for 20 different screws! It wouldn't be hard to get industry to follow a screw standard, so sooner we pick one sooner it'll become established.
Influence. If you have 20 robot part manufacturers, and only 2 are following standards at the $1 extra/unit cost, what will happen? Those 2 companies lose out as the standard just isn't influential. Point being, we need a way to encourage uptake of a standard. For example, we could make standards-compatible $50 Robot modules to get beginners used to it from the beginning. We'd also need a critical mass of people contributing to the standard - say getting people interested by having several modules built that show how easy they connect.
Anyway, if anyone has ideas on what would stop a standard from catching on, and/or solutions to it, feel free to share! I strongly feel the solutions will help the community projects go to the next level.