Musings on Design Inspiration

[JVN]

So... I've seen a few discussions "boil over" recently concerning teams building robots inspired by other robots. I wanted to start a thread dedicated to this topic.

I put my musings on design inspiration here:
http://blog.iamjvn.com/2011/03/design-inspiration.html

I hope others will take the time to digest their feelings on this sort of thing, reflect, and share them with the community.

More food for thought...
-John

[xSAWxBLADEx]

I agree 120%, yes i can honestly say we took some of their minibot ideas example wheel placement, and materials of them but I hate when teams copies the whole minibot...it's annoying

[ejSabathia]

I know this isn't church, but I feel the need to say AMEN. That is all.
~EJ

[MagiChau]

I like you distinguishing between inspiration and just copying. In minibots the same you should be done like the robots regarding inspiration, be inspired by a design another team did but make it your own as well.

[GaryVoshol]

One reason, that John touched on but can be expanded, is that it is now so much easier to find information about other robots. Sure, there was internet 10 or 15 years ago, but how long would it have taken you do download just one picture of one robot? Now we have robot matches being posted to YouTube before the robots synch up for the next match.

The same problems that teachers and professors have with plagerism - it's just so easy to find a source that can be copied - applies to robot design as well. Once a design is out there, it can circulate freely.

I think the copied teams have nothing to worry about nor complain about. But John's points about whether you learn from copying are very valid concerns. Before you use someone else's design, ask yourself if the team will be getting as much inspiration from winning with a good robot as you would with sometimes winning with your own design.

[Jared Russell]

This happens every year, mini-bot or not (think 121's roller claw in 2008, or the "pincher" in 2010). It just happens to be extraordinarily easy to copy a mini-bot, owing to a combination of rules, materials, and the fact that the emerging "best" solution is ridiculously simple.

Designs get copied. Every year. Especially this year. But unless you start an "FRC Patent Office", I'm not sure what can be done about it. And in the end, I think that the net gain (many students being inspired by implementing, thereby learning from, what others have perfected) more than offsets the net loss (that some short-sighted teams will end up short changing students from the full engineering design process).

[AllenGregoryIV]

I will say that I agree with you on most of your post. However I will say that copying or being inspired by minibots can prove to be a great learning experience for students.

At Alamo this year the DiscoBots had several students that were very dissatisfied with the MiniBot that the team had put together. It was a 3 second minibot but they wanted to be faster. We had plans to build a faster minibot and had done some of the numbers but hadn't had the time to do it. While at Alamo several students took it upon themselves to go to the teams that had faster minibots than us and ask them how they did it. (I know the list included 118, 148, and 1429; there may be more I wasn't involved in it). The kids set out on a task to have it built by the end of the regional, they worked through Thursday, Friday and Saturday not stopping for lunch to build an entirely new minibot and modify our deployment mechanism to not only launch the new one but to remain compatible with our old much larger minibot. Thanks to the machine shop at the event for helping them with a lot of the fabrication. They were franticly testing it till the last minute we had to go on the field for what would be our last elimination match (they brought our test pole to the queuing line). It was an amazing act of dedication to the cause. In the end the minibot did not win us Alamo (as you know JVN).

However, it was crucial to our Finalist performance at Lone Star. We were in a situation were we had to beat the 624 minibot up the pole to win the match and they had a about a 6 ft head start on us by the time ours started to climb. The new minibot was able to win the race by a split second and we won the 2nd semi-final by 8 points. We went on to win the 3rd match and head into the finals.

The students involved had worked through the season to build a really good minibot but they still wanted to improve, they learned from others and went about the task. I have very rarely seen teams that copy anything from another robot detail for detail. They build from pictures and adapt ideas and through that process develop a better understanding of the problem.

There have been many examples of design features have been adapted by many teams during the build season. Traction control in 2009 comes to mind. This is the first year where it has been a very crucial part of the game that directly earns a great deal of points.

[TheOtherGuy]

From what I've seen in FIRST, teams that copy generally always learn something from it, and if they don't, the copy they produce is of much lower quality than the original. In FTC in 2008, 546 made a neat mechanism that worked pretty well, for the most part. We decided to post it online, and much to our surprise, we began seeing "copies" pop up all over the place. I can't say how many of them were developed before hand, but I did notice that many teams took the design to the next level, and the majority of them worked quite well. I guess one issue with copying is that there's no way to know how much work the teams actually put into refining the design, versus just tinkering with it until it happens to work right. I'll take a stab and say it's generally the former. I know 842's minibot was influenced by 118's, but I also know how much time, effort, and engineering went into making it work. The design isn't exactly something you could copy in 30 minutes and expect to work consistently.

In general, I believe copying in FIRST is, at heart, reverse-engineering.

EDIT: I forgot to ask, has anyone ever noticed a team copy a design explicitly without learning something? I guess my major gripe with copying is when teams reap the rewards for the design without any consideration of the team that "inspired" them.

[jvriezen]

Some things to think about...

I'm curious about the relationship between the coopertition based encouragement to loan a mini-bot (the ultimate in 'copying') vs. copying a design. In the former, the 'copy' is (perhaps) initiated by the designer, whereas in the latter the copy is initiated by the recipient.

If a team came to you and asked to borrow a minibot (and you had extras to share) would you turn them down? If they asked to borrow the design, would you turn them down?

I'd agree however that copying a design without attributing the design to the originating team (who may have stolen it from a third team?) is not very GP.

Would you loan a minibot to a team on an alliance competing against you? During Qualifications? During Eliminations?

If you loaned a very good minibot during qualifications for indefinite use during the weekend, and then found yourself competing against that same team Saturday afternoon, would you 'take it back' ?

John Vriezen
Team 2530, "Inconceivable"
Mentor, Drive Coach, Inspector

[JewishDan18]

When team 20 arrived at nationals in 2009, we found a team that had copied our robot entirely. We thought it was funny, but I doubt the students on that team learned much from it.

[Tom Ore]

I have a slightly different take on the "learning" aspect of this. Let's say a group of students worked on their minibot for the entire build season, tried lots of different design options, did their best, but only produced a rather average minibot.

Somewhere in their thought process something was missing - they just couldn't see the better solution for whatever reason. They understand the problem and the parameters, but the key to breaking through with a truly top notch design has escaped them.

As soon as they see the exceptional minibot made by another team they can study it, understand why it works, they can go back through their build season and figure out what in their thought process caused them to miss the obvious, elegant solution. If they build it, they can see firsthand how it works and begin to understand the tradeoffs within that design. In this respect, there is learning benefit to "copying" a great design.

The "competion" side of things is another matter - the minibot is so easy to copy that any team that wants a fast minibot can make one easily. This may make the outcome of many matches rather random - assuming teams can make a successful deploy mechanism. The deploy may still be the larger challenge this year.

[JesseK]

John, imagine a scenario:

You're President of a business you started years ago. You've worked months on a proposal in competition with other companies for a complex job. You show up to ask questions about submissions in a public forum where (through logical interpretations of even the cleverest wordsmithing) you've discerned some details of your competitors' submissions. However, all it revealed was that your approach to the problem was completely inadequate. What do you do?

Given the chance for improvements, I doubt you'd stay the course. After all, the morale of your employees and reputation of your company is at stake. Sure, I would be quite offended if I had a novel [insert FRC design here] mechanism that was cloned in the same season by a team I may later face; yet being a mentor you too can learn a significant lesson from this.

The rest of this is an anecdote that's merely here to say that I think there's a fine line between believing a team 'simply created a clone' versus the team 'changing course in some core aspects of the design'. When I say "you" I mean the generalized team who has a 1.5s minibot using DD shafts that weighs next to nothing.

Do you expect credit for the entire minibot, certain aspects of the minibot, what? I wasn't 'inspired' by any of the minibots seen here on CD or in web casts; I was simply proven to have made a wrong decision along the way. We visited the idea of direct-drive minibots briefly early in the season, yet shied away from it because we had to make a decision (due to scheduling and snow) before we were able to test all options. The test would have been to see whether or not the drive shaft could have made it through constant shocks while being only cantilevered. So instead we went with a more conservative, modified-gearbox design. It's heavier overall (4.0 lbs) yet has many aspects that are similar to minibots seen in videos and even the "one-day minibot" posted on CD.

If anything, these latest designs will have us revisit the minibot design after DC should we make it to Championships in order to iterate through the direct-drive options. But to say we "cloned the powerhouse teams" would be to completely discredit us, and everything we figured out on our own other than direct-drive shafts and their placement. These things include magnet placement, center of gravity placement, secure attachment to the deployment mechanisms for match play, quick releasing for deployment, etc. This specific anecdote is almost as bad as someone saying a powerhouse team is "mostly mentor built" -- it's simply not factual to paint such a broad stroke.

[Basel A]

The question of how Coopertition plays into all this is an interesting one. Copying is one thing, but in borrowing a minibot you don't even get the experience of building it, and an even lesser admiration. However, FRC is encouraging this behaviour. Perhaps when one team had no minibot at all, this is the slightest bit understandable, but in cases I've seen, a borrowed minibot has simply replaced another, and a team has undermined its own work to seek greater success.

[Andy Grady]

Originally Posted by Tom Ore

As soon as they see the exceptional minibot made by another team they can study it, understand why it works, they can go back through their build season and figure out what in their thought process caused them to miss the obvious, elegant solution. If they build it, they can see firsthand how it works and begin to understand the tradeoffs within that design. In this respect, there is learning benefit to "copying" a great design.

I think this is a fantastic point. Midseason redesign is nothing new to this competition. Since 1995 (the first time there was a regional and championship competition in a season) teams have made adjustments, often based off of what they see at a regional, in order to climb back into the race for the second chance at winning. Minibots make it much easier to copy a design and incorporate its strategy at a future competition, but I would hazard a guess that just about every team who does it has went through the painstaking learning process the first time around. They will continue to learn the lesson by building the clone and seeing what they did wrong the first time around. To be honest, I completely expect to see tons of 1.5 minibot clones at championships, and I say thank god! This game was too heavily weighted towards the minibot race from the get-go. From the standpoint of a spectator, I like the prospect of seeing more speedy minibots to make the races more interesting. The more minibots that go up the towers, the more the competition is geared towards skill in autonomous and basic scoring. With so many clones, I would imagine the focus on the minibot race will shift more towards deployment...an aspect which I think most people will say the true challenge lies in that portion of the game.

[PAR_WIG1350]

I'd be careful here, don't be quick to judge. I walked into a meeting one day and saw my team was testing a component that looked almost exactly like one I had seen in a youtube video. When I mentioned it, I got confused looks from everybody. They hadn't seen it and had developed the same component independently without knowing it had already been done. Just because it looks copied, doesn't mean that it is.