Posted on the FRC Blog, 12/5/19 by Frank Merrick.
One More Team Update and Q&A Closing Extension
Written by Frank Merrick.
Check out the latest Team Update here.
Because of this update, we are extending the closing of the official Q&A system to Noon Eastern Time on Tuesday, December 10. Please submit any additional questions you have by then.
Frank
Ahahahah. Youāre welcome.
With the Q&A extension, does anyone want to ask about the recent technicalities brought up by the Venom upgrade program?
Edit: Also Q84 is extremely important
Hmm. So, hypothetically, do we have to consider our teamsā software to have value? Based, perhaps, on the number of student-hours taken to write it times minimum wage or something? (Remembering that āfair market valueā is supposed to be used for cost accounting, rather than the money you actually had to spend to get the thing.) If I, as a mentor, contribute directly or indirectly to the code, should I charge my hourly rate to the CAW?
I feel this falls under the same precedent as mechanical design, so any time is assumed to be donated
I think the question you are asking is:
Can software development labor provided by team members (including sponsor employees who are members of the team) be excluded from the BOM?
Unfortunately,
This question is about a rule which is not part of the pre-kickoff rules release. If you still have a question after reading the 2020 Game Manual released at Kickoff, please re-submit.
Software development is labor. Assuming thereās a 2020 rule equivalent to 2019ās R14:
R14. The BOM cost of each non-KOP item must be calculated based on the unit Fair Market Value for the material and/or labor, except for labor provided by team members (including sponsor employees who are members of the team), members of other teams, event provided Machine Shops and shipping.
If done by a team member, you may exclude the value of that labor from the BOM.
Thatās gonna need to be a post kickoff question if you want an answer.
So, WPILib does not come in the Kit of Parts, or the Virtual Kit of Parts, and with rare exception, is not made by team members. How do you account for the thousands of commits on this crucial piece of robot code infrastructure, given that thereās no easily calculable value, but you canāt deny that Thad, Peter, and the others time has value?
Or maybe we need to consider that this rule change was simple in theory and has ramifications beyond what they intended, as I doubt they want almost every single team to be too expensive before they even have a single component made.
Thank you for that! 
I think that with respect to costing, however, the fair market value for anything that the codeās license allows you to use for free is $0. I think itās a pretty unreasonable reading, for example, for this rule to require you to estimate the development cost of Linux if you stick a Raspberry Pi on the robot. Same goes for vendor librariesāwhile they certainly have a development cost, and I think at least some of them have licensing that only allow you to use them with their products, they are free for any team to download, install, and use, so they should have a FMV of $0.
I think this pokes at an interesting question, which is, āwhat is āfair market valueāā? Is it the cost to reproduce something? The cost to purchase an identical item? The cost to buy this specific item?
Let me give an example. This is a RAD750 computer:
(Source)
That board costs several hundred thousand dollars at least, because itās space qualified and radiation hardened. The purple chip in the bottom right is the RAD750 itself, and the five gold chips across the top are its RAM.
Here is one of those RAM chips in my hand:
Now, hereās my question. Suppose my team decided to use one of my rad-hard RAM chips on our robot. What would be the āfair market valueā? Would it be the $10,000 that BAE would charge for a new one? The price that I could get for one on ebay ($40-$100 or so, depending on gold content)? Or the $7 price of a functionally equivalent, but non-rad-hard, 32MB SDRAM chip on Digikey?
Fair Market Value was defined in the Blue Box under R14 in the 2019 manual. Iām hoping that section will get some updates prior to the 2020 release.
The phrase āfair market valueā was used in R14. It wasnāt defined anywhere, that I can find.
Granted, Iām unlikely to actually use a space-qualified SDRAM module on an FRC robot. For a slightly more relatable example, what is the fair market value of a Spike relay (ignoring any KOP exemptions)? There isnāt really a manufacturerās price, because no one makes them anymore. Is the FMV what IFI originally charged, or is it the price you can buy a Spike for on ebay?
āThe Fair Market Value of a COTS part is its price defined by a VENDOR.ā
And then the blue box continues to give multiple examples. That seems like a definition to me but maybe we have different definitions of definition.
Ah, yes, sorry about that. My local PDF of the manual must be outdated.
But I think my question about the FMV of a Spike still stands. Obviously individual ebay sellers donāt count as VENDORS, unless they have really large quantities of Spikes. But thenā¦ what is the FMV of a Spike? Is it $0, since there are no VENDORS (that I can find)? Is it the most recent VENDOR-advertised price? And, getting back to my original example, what would be the FMV of a BAE rad-hard SDRAM chip? Bear in mind that BAE wonāt even give you a quote for one unless youāre a national space agency or major contractor; my $10,000 figure was an estimate.
Am I nitpicking? Sure. But Iām opposed to the idea that you can disregard corner cases in your specifications (rules) just because you donāt think anyoneās likely to exercise them.
Blue box, Example E, is the closest thing in this particular case. That said, for a discontinued item, I would ask the question in Q&A of āIf no Vendor sells a particular item (as it is discontinued), but it can be obtained used from other Teams, online auctions, etc., would the FMV be the price obtained for it or the most recent Vendor price?ā
For the rad-hard SDRAM chip, I would follow the ābetter safe than sorryā approach and either do a scattershot Google search for prices of similar items, contact the company indicating that you have somehow acquired these and for cost accounting purposes you need to know an approximate FMV, or just not use them. I wouldnāt use the Ebay or functional-equivalent cost.
Iām opposed to the idea of covering all the corner cases, because that leads to really thick rulebooks. If it can be done REASONABLY (e.g. a line or two for several of these), thatās awesome. But thatās partly what the Q&Aās for, and partly what the LRI/Head Ref are there for.
To use the Spike example: If Iām an LRI, and someone comes up to me and says, āHey, I think a teamās skipping the FMV of a Spikeā, then Iām going to ask two questions. The first is āHow close is the team to the maximum BOM limit, and how many do they have?ā (Given my time in FRC, I know the Spikes are under the individual item limit even if I donāt know the exact price. Offhand, theyāre under $200 from manufacturer, new, IF they were still being made.) If the team isnāt close, Iāll ask the team to do some research before their next event, point them in the right direction, and pass them anyways. If the team IS close, then the goal is to work with the team to figure this stuff outāincluding ācan we open up the BOM a little moreā, a message to Big Al and his co-chief, and possibly even a message to VEX/IFI. Youāre right, Iām dodging the question. But it gets the team legal faster. Quite possibly thereās a followup in the next rules update as well.
I agree that thatās a good Q&A question, for posterityās sake if nothing else. Iāll submit it in the 2020 Q&A, assuming nobody beats me to it.
Good luck with that 
I get your point. Youāre saying that (L)RIs should apply reasonable judgement, and, for example, not demand proof of FMV for Spikes from a team thatās otherwise at $2000 total. I donāt disagree with that at all, and Iām in favor of overlooking purely procedural violations in order to let a team get on the field.
I guess my sticking point is that āreasonableā is subjective, and that thereās no real concept of stare decisis in FRC. If (L)RIs have previously declared your grey-area design to be legal, that doesnāt keep (L)RIs at your next event from declaring the same thing illegal. I believe Marshall refers to this as āLRI-a-rooā. My position is that itās better to have more restrictive but consistent rules, rather than leave a lot of latitude to LRIs at individual events. Or, alternatively, to have less restrictive but consistent rules, and accept that some teams may profit from it.
Ultimately, I think that itās better for a team to fail inspection at their first event, than to pass at their first event and fail at their second event without changing their robot.
I recognize that thatās a philosophical judgement, and eventually the decision on that point needs to flow down from FIRST HQ.
So, there is a question that we deliberately did not ask that speaks to FMV and the subject of āwhenā. Iāll post it here and someone else is more than capable of running with it if they want to:
Can you clarify when (in time) to evaluate the cost of materials. The text preceeding R11 seems to imply that costs from the time of Kickoff 2019 would be used as the cost of the robot is being reverted to that time. Is the time to evaluate costs at purchase time, at use time, at inspection time, or at the time of Kickoff 2019? In addition, can you clarify how prices should be evaluated to adjust for inflation? Should the standard Consumer Price Index (CPI) valuation be used or would it be acceptable to use the Retail Price Index (RPI)?
It needs some editing before being asked. Also, shout out to Planet Money and The Indicator podcasts that myself and other Zebracorns are fans of.
So is this RIP Robotpy? or as someone else mentioned WPIlib?
How about Chameleonvision?
OpenSight?
Pathfinder?
What about anything built in Onshape since itās not on the KOP page?
