2nd Most Awaited Q and A Answer?

[Squillo]

I see that a new question has been asked, and not yet answered by the GDC (thanks 1619). Hopefully they will know of our confusion and resolve it, one way or another.

I am wondering about this: an "H" shaped construction that has two pieces extending past the perimeter, joined by a bar that crosses between them (and is outside the perimeter when the 'appendage' is extended, but crosses from inside to outside during the course of extension), and then two more pieces that continue past the cross bar. Or for that matter, two pieces that cross the perimeter and are simply joined into a "u" with a crossbar outside the perimeter.

Any difference between those two? Either legal/illegal? Thoughts?

BTW, I just have to get this out there. I know it is the GDC's language, and I should take it up with them, but as a lawyer (and mentor/'rulesmeister') I really resent the negative connotation that has been placed on the term "lawyering". It just so happens that a large part of proper legal reasoning, argument and decision-making is focused on discerning the INTENT behind the rule, statute, or contract, and explaining how/why our interpretation is in accord with that INTENT - NOT, as some believe, looking for 'technical loopholes' that are contrary to the intent of the legislature/court/contract drafter. You engineers are MUCH better at that than we lawyers. Maybe we should call it "engineering" the rules...

BTW and FYI, the January 2012 issue of the magazine GPSolo, put out by the general practice, solo and small firm division of the American Bar Association, which is focused on volunteering and community service, contains an article entitled "Mentoring a High School Robotics Team" by yours truly, which details why I do this and plugs robotics in a big way. As soon as the online link is available, I'll post it. Would anyone care to suggest a forum or sub-forum that would be appropriate for that?

[MrForbes]

Originally Posted by Squillo

BTW, I just have to get this out there. I know it is the GDC's language, and I should take it up with them, but as a lawyer (and mentor/'rulesmeister') I really resent the negative connotation that has been placed on the term "lawyering". It just so happens that a large part of proper legal reasoning, argument and decision-making is focused on discerning the INTENT behind the rule, statute, or contract, and explaining how/why our interpretation is in accord with that INTENT

I think we're aware that that is what a large part of proper legal reasoning is. The thing that gives lawyers a bad reputation is the occasional improper use of the technique to find loopholes.

Some of us have learned over the years to look right away for the most conservative interpretation of the rules, it saves a lot of redesign.

[RRLedford]

At what point does an appendage become an appendage?

Is it at its attachment point inside the bot, or at the point where it crosses outward beyond the perimeter? Along what paths are we supposed to trace out contiguity tests. If all bots are 100% contiguous then what establishes dis-contiguousness as relating to appendages?

What exact application of appendageness and contiguity determine whether we have just one, or more than one appendage protruding?

What if our whole robot just expanded in all directions (above the bumpers of course) for 14" beyond its initial size, and we consider this our "frame appendage"? If it is contiguous are we in violation? What if we deploy diagonally at a corner? Can the appendage still only be 14" diagonally from the corner, or (14") X (1.414) -- still within a rectangle going 14" further outward than the frame?

-RRledford

[Tuba4]

Originally Posted by RRLedford

What if our whole robot just expanded in all directions (above the bumpers of course) for 14" beyond its initial size, and we consider this our "frame appendage"? If it is contiguous are we in violation? What if we deploy diagonally at a corner? Can the appendage still only be 14" diagonally from the corner, or (14") X (1.414) -- still within a rectangle going 14" further outward than the frame?

-RRledford

This would clearly violate the following:

[G21] Robots may extend one appendage up to 14 in. beyond a single edge of their frame perimeter at any time.

No need to discuss continuity here. One edge at a time!!!!!

[RRLedford]

Originally Posted by Tuba4

This would clearly violate the following:

[G21] Robots may extend one appendage up to 14 in. beyond a single edge of their frame perimeter at any time.

No need to discuss continuity here. One edge at a time!!!!!

While this rule may well be intended to mean "beyond ONLY a MAXIMUM of ONE SINGLE EDGE," it is not really worded to accomplish this.
Plus, it should have added ==> "and may NOT extend ANY AMOUNT beyond ANY OTHER EDGES of the robot." -- if this was what they really meant.

Consider this example: Let's suppose I tell you to deploy an appendage diagonally at 45 degrees off the corner of our bot for 14" of extension, and the robot is a rectangle.
I then tell you to check whether this arm extends more than 14" beyond any single edge of the robot.
You then report back to me that compared to NO SINGLE EDGE does the robot arm extend more then 14". You verified this by holding up a long straight edge spaced 14" away from each side, one at a time, to confirm this.
This rule's wording can also simply mean that you are comparing the tip of the arm's position to EACH INDIVIDUAL FRAME EDGE LINE ==> ONE LINE AT A TIME. So even though the appendage clearly extends beyond two of the four edges, it does NOT extend beyond 14" for "any single edge " measured ALONE.

BTW, if a circular robot is allowed, does it only have one edge? If so could it deploy a skirt outward, all the way around the robot for up to 14" as long as the skirt formed a "contiguous" loop?
There does seem to be a distinct bias toward rectilinear design concepts with the structure of the FIRST game rules.

-RRLedford

[JamesCH95]

For the love of Andy Baker can we please stop lawyering this into oblivion and use some common sense? The GDC doesn't want to artificially limit designs through inane rule interpretation.

You may have 1 mechanical appendage at a time that may extend up to a 14" offset of your frame perimeter. If the robot is a circle, any appendage must remain within a circle of a 14" larger radius with the same center as the frame perimeter. If your robot is a rectangle then any appendage must remain within a rectangle of 28" greater width and 28" greater length with the same center as the frame perimeter.

The appendage may have forks or splits in it as long as it is mechanically connected in such a way as they must function unison. Your arm+hand+fingers is considered one appendage. Your two arms acting in unison through brain commands (i.e. robot code) are still two separate appendages.

[Tuba4]

Originally Posted by JamesCH95

For the love of Andy Baker can we please stop lawyering this into oblivion and use some common sense? The GDC doesn't want to artificially limit designs through inane rule interpretation.

You may have 1 mechanical appendage at a time that may extend up to a 14" offset of your frame perimeter. If the robot is a circle, any appendage must remain within a circle of a 14" larger radius with the same center as the frame perimeter. If your robot is a rectangle then any appendage must remain within a rectangle of 28" greater width and 28" greater length with the same center as the frame perimeter.

The appendage may have forks or splits in it as long as it is mechanically connected in such a way as they must function unison. Your arm+hand+fingers is considered one appendage. Your two arms acting in unison through brain commands (i.e. robot code) are still two separate appendages.

I concur. You said it much more elegantly and succinctly than I could. I was being too specific. You were more general. Now if only the GDC agrees as well.

[Tristan Lall]

Originally Posted by JamesCH95

You may have 1 mechanical appendage at a time that may extend up to a 14" offset of your frame perimeter. If the robot is a circle, any appendage must remain within a circle of a 14" larger radius with the same center as the frame perimeter. If your robot is a rectangle then any appendage must remain within a rectangle of 28" greater width and 28" greater length with the same center as the frame perimeter.

The appendage may have forks or splits in it as long as it is mechanically connected in such a way as they must function unison. Your arm+hand+fingers is considered one appendage. Your two arms acting in unison through brain commands (i.e. robot code) are still two separate appendages.

The GDC still needs to say so, and say so in a way that clears up the minor inconsistencies. This wouldn't be a bad interpretation for a team to employ (in the absence of clarity), because it's conservative and likely to stand up to scrutiny by competition officials.

But it's not helpful for FIRST officials to each be enforcing slightly different variations on the rules, because the GDC wouldn't clarify things a bit further. That's particularly problematic with the forked appendage rule: what's a valid mechanical connection? Is it literally anything—e.g. the rest of the robot? Is it anything that bears a load of more than a certain amount? Is it anything that looks structural? Is it an issue of degrees of freedom between appendages? Or would a thread tied around two appendages make them one? What if the thread was instead a coathanger? What if it was a tie rod with ball joints at each end?

When inspectors/referees are making rulings, hopefully they're also considering the general case—because what seems good in specific circumstances may have implications for future rulings (if they're consciously attempting to be consistent, which they usually are).

Originally Posted by RRLedford

While this rule may well be intended to mean "beyond ONLY a MAXIMUM of ONE SINGLE EDGE," it is not really worded to accomplish this.
Plus, it should have added ==> "and may NOT extend ANY AMOUNT beyond ANY OTHER EDGES of the robot." -- if this was what they really meant.

Consider this example: Let's suppose I tell you to deploy an appendage diagonally at 45 degrees off the corner of our bot for 14" of extension, and the robot is a rectangle.
I then tell you to check whether this arm extends more than 14" beyond any single edge of the robot.
You then report back to me that compared to NO SINGLE EDGE does the robot arm extend more then 14". You verified this by holding up a long straight edge spaced 14" away from each side, one at a time, to confirm this.
This rule's wording can also simply mean that you are comparing the tip of the arm's position to EACH INDIVIDUAL FRAME EDGE LINE ==> ONE LINE AT A TIME. So even though the appendage clearly extends beyond two of the four edges, it does NOT extend beyond 14" for "any single edge " measured ALONE.

BTW, if a circular robot is allowed, does it only have one edge? If so could it deploy a skirt outward, all the way around the robot for up to 14" as long as the skirt formed a "contiguous" loop?
There does seem to be a distinct bias toward rectilinear design concepts with the structure of the FIRST game rules.

I'm not sure I'm totally on board with the one-edge-at-a-time measurement scheme, because I would tend to assume the rule is to be interpreted simultaneously with respect to all edges—but I concur that it's not clear whether the appendage has to physically cross the projection of an edge, or simply be extended into the space beyond an edge. (Imagine a piece that crosses only one side, but then is actuated so it curves into the space beside another edge. Was it "extend[ed]...up to 14 in. beyond a single edge"?)

As for the rest, you beat me to posting it. FIRST has a history of issuing interpretations that don't make sense with respect to non-rectilinear robots. And they frequently omit things like maxima and minima (or any tolerancing at all).

[Jon Stratis]

In another thread on here (I can't find it now, but it was dealing with bumper rules), someone postulated that a perfectly circular robot, rather than having a single edge on their frame perimeter, would actually have an infinite number of exterior vertices. Under that interpretation, any appendage extending over the frame perimeter would be crossing multiple edges.

As far as inspecting/reffing this rule... there's nothing in inspections that should be affected by this rule. It's perfectly legal for a robot to have as many appendages as they want, and to have them on one side or multiple sides. They just can't extend them all at the same time. So it really comes down to reffing. Refs are going to follow a rather simple rule in calling penalties for this - if it looks like multiple appendages, then it gets penalized. Here in Minnesota, we have 60+ teams at each competition. It's going to be next to impossible for the refs to remember which robots have multiple appendages on the same side, versus which ones have a single appendage that just looks like multiple appendages.

So save yourself and the refs some headaches and make things obvious.

Edit: found the post referencing circular designs: http://www.chiefdelphi.com/forums/sh...2&postcount=43

[RRLedford]

Originally Posted by JamesCH95

For the love of Andy Baker can we please stop lawyering this into oblivion and use some common sense? The GDC doesn't want to artificially limit designs through inane rule interpretation.

You may have 1 mechanical appendage at a time that may extend up to a 14" offset of your frame perimeter. If the robot is a circle, any appendage must remain within a circle of a 14" larger radius with the same center as the frame perimeter. If your robot is a rectangle then any appendage must remain within a rectangle of 28" greater width and 28" greater length with the same center as the frame perimeter.

The appendage may have forks or splits in it as long as it is mechanically connected in such a way as they must function unison. Your arm+hand+fingers is considered one appendage. Your two arms acting in unison through brain commands (i.e. robot code) are still two separate appendages.

Even this "common sense" explanation still lacks clarity as to whether or not an appendage can deploy diagonally off a corner, and, if so, how far?
There may be a 90 degree angle forbidden zone at the vertex of each 90 degree frame corner?
Just going by the "rectangle of 28" greater width and 28" greater length" analysis, could allow a 40+% longer (14" X 1.414) appendage length at corners if the diagonal deploy is legal, but it would still be crossing both of the bot's adjacent frame edge lines, which may not be allowed.

-RRLedford

[Tuba4]

Originally Posted by RRLedford

Even this "common sense" explanation still lacks clarity as to whether or not an appendage can deploy diagonally off a corner, and, if so, how far?
There may be a 90 degree angle forbidden zone at the vertex of each 90 degree frame corner?
Just going by the "rectangle of 28" greater width and 28" greater length" analysis, could allow a 40+% longer (14" X 1.414) appendage length at corners if the diagonal deploy is legal, but it would still be crossing both of the bot's adjacent frame edge lines, which may not be allowed.

-RRLedford

I would think the corner of the "zone" would not be a 90 degree angle, but a radius of 14 inches. The extended appendage can not exceed 14"!!!

[JamesCH95]

Originally Posted by Tuba4

I would think the corner of the "zone" would not be a 90 degree angle, but a radius of 14 inches. The extended appendage can not exceed 14"!!!

^This.

It is exactly common sense. The appendage in this case is never more that 14in from the frame perimeter. In my opinion this is what the GDC is going for, and it makes sense (I do not speak for the GDC).

[DMetalKong]

Let me try my hand at this:

1) Two elements are contiguous if the degrees of freedom between them is zero (i.e. when power is not applied, given the orientation and position of one element it is possible to compute the exact orientation and position of the other element).

2) Any elements that are both inside and outside the frame boundary (i.e. reaching across the frame boundary) must be contiguous.

3) Any elements outside the frame boundary must not extend outside of the boundary formed by extending the frame boundary 14" perpendicularly outward and rounding any resulting vertices with radius 14"

4) Any element crossing the frame boundary must form an angle of no more than 90 degrees with any other element that crosses the frame (measured from the centroid of the frame boundary)

The language could be cleaned up (especially in points 3 and 4), but I think this covers all of the situations that have been discussed.

[Tristan Lall]

Originally Posted by DMetalKong

Let me try my hand at this:

The language could be cleaned up (especially in points 3 and 4), but I think this covers all of the situations that have been discussed.

That's a good start. Allow me to pick it apart a little.

• The degrees of freedom language is a good idea, but how do you account for component and assembly flexibility? If I join two things with a bar of aluminum, does that imply 0 DOF? What if the bar is really thin and flexible? (Basically, is there a threshold beyond which you consider something to be a DOF?)
• Maybe you want to describe "crossing" the frame boundary, and mention that "contiguous" refers to the parts on either side of that boundary? (Otherwise, it could be interpreted as meaning contiguous with respect to some other thing.)
• From what parts of the appendages is the relative angle determined?
• Using the centroid is good in principle. However, depending on whether your definition of frame boundary can vary due to robot configuration changes, you might have a uniqueness problem. (Was "frame boundary" meant to be the same as the "frame perimeter"?) Also, unfortunately the centroid is imaginary and hard to locate.
• I assume you understand that the 90° spec you outline is not equivalent to the existing constraint. Also, presumably you mean the smallest angle between them. (And incidentally, isn't 75° a lot like 105°? Why would one be illegal and the other not?)

This is actually the exercise the GDC needs to go through internally (who knows, maybe they do) to settle on verbiage that reflects their intent accurately, and exposes the potential for misinterpretation.

[DMetalKong]

Originally Posted by Tristan Lall

• The degrees of freedom language is a good idea, but how do you account for component and assembly flexibility? If I join two things with a bar of aluminum, does that imply 0 DOF? What if the bar is really thin and flexible? (Basically, is there a threshold beyond which you consider something to be a DOF?)

IMO an unconstrained flexible element fails the 0 DoF test because when power is not applied, given the position of any other element (taken pair-wise) that is part of the appendage it is not immediately apparent what position the element in question will occupy. As to whether or not an element is "flexible" (as all elements will have some degree of deflection to them), I feel that this is something to which the reasonable man test can be applied (i.e. an element is flexible if a reasonable man believes that the element was designed to take advantage of its deformation)

Originally Posted by Tristan Lall

• Using the centroid is good in principle. However, depending on whether your definition of frame boundary can vary due to robot configuration changes, you might have a uniqueness problem. (Was "frame boundary" meant to be the same as the "frame perimeter"?) Also, unfortunately the centroid is imaginary and hard to locate.

I did mean "frame perimeter" when I wrote "frame boundary"; since the frame perimeter must not articulate, I believe this covers your first point. As to the centroid being imaginary, I see no other precise solution that would offer a definite "center" to the robot; if there is a major disagreement between a inspector and a team the centroid can be (albeit with difficulty) be calculated.

Originally Posted by Tristan Lall

• Maybe you want to describe "crossing" the frame boundary, and mention that "contiguous" refers to the parts on either side of that boundary? (Otherwise, it could be interpreted as meaning contiguous with respect to some other thing.)

The idea was that any part that crosses the frame perimeter must be contiguous with any other part that crosses the frame perimeter, in a pair-wise fashion. A part that "crosses" the frame perimeter is one that is both inside and outside the frame perimeter simultaneously.

Originally Posted by Tristan Lall

• From what parts of the appendages is the relative angle determined?
• I assume you understand that the 90° spec you outline is not equivalent to the existing constraint. Also, presumably you mean the smallest angle between them. (And incidentally, isn't 75° a lot like 105°? Why would one be illegal and the other not?)

The 90° was chosen because barring requiring all robots to be rectangular in shape I see no reasonable way to define the "sides" of a robot in a way that allows for various geometric shapes, while retaining what I believe is the intent of the rule: to allow appendages to extend, but in a relatively narrow direction. Perhaps a better test would involve rotating a 90° cone around the centroid.

Given your feedback (much appreciated by the way), here is a revised list:

1) Two elements are contiguous if the degrees of freedom between them is zero (i.e. when power is not applied, given the orientation and position of one element it is possible to compute the exact orientation and position of the other element). Flexible elements will be considered to add to the degrees of freedom if a reasonable man believes that the element was designed in such a way as to take advantage of its deformation.

2) Any elements that are simultaneously both inside and outside the frame perimeter (i.e. reaching across the frame perimeter) must be contiguous in a pair-wise fashion (i.e. any element crossing the frame perimeter must be contiguous with any other element crossing the frame perimeter).

3) Any elements outside the frame perimeter must not extend outside of the boundary formed by extending the frame perimeter 14" perpendicularly outward and rounding any resulting vertices with radius 14".

4) Any elements outside of the frame perimeter must lie within the right isoceles triangular prism constructed with infinite height and infinite leg length and placed so that the vertical edge of the right angle must be coincident with a vertical axis placed through the centroid of the robot.