[msg=298703]In this thread?[/msg]
I for one, perused it. I did not understand some of the Greek Symbols. I did not understand most of what Kent said. But, my personal take on it was something like:
"When you shoot point 10, and it has a NS error ellipse, of 0.06', and an EW Elise of 0.03', then you shoot point 11, and it has a NS error ellipse, of 0.07', and an EW ellipse of 0.03', and you draw a plat, showing a line between 10 and 11, you have to ADD the total Northing error, and the total Easting error, to get 0.13' NS error potential, and and EW error potential, of 0.06. However, you also have to factor in the possibility that the pole bubble was off, and the user may have faced E when he shot point 10, and west when he shot point 11. Then, since you have no redundancy, in this method of surveying, you don't have a strong basis for the above total error potential, between 10 and 11. So, maybe you can add in an estimated error potential in the bubble, of maybe 0.04' (0.02'+0.02') so, we have a total error potential, of 0.17', so now can we do an ALTA survey, and RELY on these methods, and certify with full confidence in our accuracy numbers?"
That seemed to me to be what Kent was saying. Or trying to say. Now, granted, error ellipses, are never fully NS, or EW, but that was close enough, for discussion purposes.
I am an "In the dirt surveyor". I go and do stuff, in the field, and check it in methods, I am comfortable with. However, I have never been able to get away from the feeling, that I cannot get my hands on the actual accuracy numbers, And all the coordinate geometry, of our modern RTK, or post processed Static work. I don't know how to do the math, to work LAT LON, into a plane, on a sphere, and then modify it on to a geoid, and set an elevation, and work and think with it. I am trusting my computer. And, the software developer. Heck, I don't even know the least bit about machine language.
So, How many of you actually understood what Kent was saying? How can we agree, or disagree, if he is talking French, and we don't use French?
I suspect that there are more like me. And, that there are some who know alot less than I do, who are RTK practitioners. I can remember, taking a new guy, and putting him on the gun, and teaching him to wind up angles, and how to chain. We could ALWAYS get our fingers on the numbers. We have gotten quite away from that. I can see a 1 wk trainee, with RTK, and a pickup, and a shonstedt, and a shovel. And, maybe even without a shonstedt! Give him a chain saw, and a dull machete!
I am not all that comfortable with where we are headed, but the train is moving faster and faster, gaining momentum, and we are all trusting the programmers, who know nothing about what we do.
Life is like that. We are on a journey. Where will it all end?
I need to fish more.
Nate
> [msg=298703]In this thread?[/msg]
>
> "When you shoot point 10, and it has a NS error ellipse, of 0.06', and an EW Elise of 0.03', then you shoot point 11, and it has a NS error ellipse, of 0.07', and an EW ellipse of 0.03', and you draw a plat, showing a line between 10 and 11, you have to ADD the total Northing error, and the total Easting error, to get 0.13' NS error potential, and and EW error potential, of 0.06.
No, it's a bit different than that. If the uncertainties of Point 10 and Point 11 with respect to the base aren't correlated, that means that the errors in the relative positional uncertainty between them add according the root sum of squares law, not simple addition.
So in your example, the relative error ellipse between 10 and 11 at the same confidence level as the ellipses you describe for 10 and 11 would be computed as follows:
Semi-major axis = SQRT[ 0.06^2 + 0.07^2] = 0.09 ft.
Semi-minor axis = SQRT[ 0.03^2 + 0.03^2] = 0.04 ft.
As a practical matter, good least squares survey adjustment software will make this computation for all case, including those where the error ellipses are not truly independent and where the axes aren't parallel.
> So, How many of you actually understood what Kent was saying? How can we agree, or disagree, if he is talking French, and we don't use French?
>
It's mostly over my head right now. The recent posts on this subject have motivated me to start reading up on it to try to get a better understanding.
It seems like something that is important to understand for all types of measurements not just GPS.
Button Pushers
> I can see a 1 wk trainee, with RTK, and a pickup, and a shonstedt, and a shovel. And, maybe even without a shonstedt! Give him a chain saw, and a dull machete!
>
I can teach a guy what buttons to push on a robotic total station just as quick as I can teach a guy what buttons to push on a GPS. An idiot with a total station can do just as much damage as an idiot with a GPS.
Least squares is a great tool. I understand it, but only because the OIT geomatics curriculum required it. I don't know that I would have gravitated to it on my own.
We were taught the foundational statistics, vector algebra, geodesy, trig, and geometry that powers the black boxes. That is one strong argument for the 4 year requirement. If you don't understand how the box works it is pretty tough to test/rely upon your results.
Dang it, Nate
All your jibber jabber reminded me of my college days. It was summer school and the class was one no one but an engineer would be enrolled in. There were eight of us in there and I was a minority in that particular class. Only two of us were natives of the U.S. of A. Three were from Nigeria, one from Bangladesh and two from India. The professor was born in Taiwan (China, according to him). The professor had a great accent that made for fun interpretation. The Greek letter rho came out much more like zho and theta was saytuh. We used many Greek letters in that class. So we had a Chinese professor using Greek letters in an attempt to educate this mixture of nationalities while speaking English.
We had to straighten out the professor when, on Friday, he announced we would have a major examination on Monday. He didn't quite get what we meant initially when we said, "But, that is the Fourth of July." No classes were to be held on the Fourth.
I understand it, but if I hadn't had the Geomatics courses at UF, I wouldn't....and I doubt that I would be disciplined enough to follow the ins and outs of it if I didn't have a degree. I'm not knocking least squares adjustments, but adjustments like compass rule have suited most Surveyors sufficiently for years on end. For the vast majority of surveys most Surveyors perform, the differences wouldn't be overly significant anyway. For me, it's just another tool in the toolbox. If I close out a traverse within a hundredth or two and a few seconds, I'm not adjusting it. Perhaps that makes me a bad Surveyor to some, but the commissioner of common sense tells me otherwise.
I understand some of it, I get the gist of some more, and I hope to someday understand the rest.
LS adjustment is vastly superior to compass rule for several reasons, among them:
1. Ability to simultaneously adjust GPS vectors, TS data, and levelling.
2. Data doesn't need to be added in any particular order.
3. Detailed reporting.
And more. A lot more. Claiming the compass rule is good enough is just whistling past the graveyard. Compass rule is better than nothing, but not even close to LS. If you are not adjusting at all because you don't want to "add error" to your measurements then you don't understand the difference between data adjustment and blunder detection.
And LS is SIMPLE to do! It's a computer program, people! If you can manage a CR adjustment you can do an LS adjustment. If you have Carlson Survey, you have the software. Get with it!
You hit on a point I've made numerous times over the years.
Statistics have not changed appreciably since I began my career. Our tools and the nuances of expression have. My lack of early education had me in the same boat you describe. I did not like struggling to understand what I perceived as higher math. My solution was to head on down to Boise State and push myself further.
Looking back I had hit that same wall as a chief in the early 80s. I borrowed TS Madsen and bugged the crap out of everyone around me. I learned what I needed to know that what I was doing was correct.
I would like to say I am in a maintenance phase now, but I've discovered that math actually leaks out of my head over time..
Ramble off.. Tom
Well, thanks Kent.
You apparently have dedicated yourself to this, in a mathematical way, that I have not. I have been too busy surveying, to figure it out in those terms.
That is, I have been busy with the Great Tripod of Surveying, as alluded to in [msg=298907]This Thread.[/msg]
Kent seems knowledgeable of one ASPECT of ONE of the legs of that tripod. It can affect the other legs some, as it would mean we manage our uncertainties, in a more systematic, and straightforward manner. But, I am still bucking around in the woods, trying to retrace a surveyor, who sets rebar that are 10' off, from where his plat says they are.
I guess that's why it's called "Theory, and Practice" of surveying. Theory is what the plat said. Practice, is what happened in the field. Which means that not-so-professional surveying is still happening.
I want to take a minute, to thank Kent for his dedication to:
1.) Learn it all.
2.) Try to hammer it into Surveyors. (That's synonymous with trying to take a Catholic, and baptize him in a Baptist Church!, or Vice Versa, if you will!)
I for one, will take all the info I can, and practice it. But, I cannot neglect the other legs on the Great Tripod.
OK, gotta go make some money. The Great Tripod is fun, if there is enough money at the tribrach!
Thanks Everybody.
Nate
That's the thing, LSA is sooooooo easy!!!
In the work flow it's almost non-existent compared to the old days of hours and hours of adjustments, now it's a click, 20 page report generated in .6 seconds.
:good: :good: :good:
And you can use your various types of measuring equipment to close your survey the most efficient way and adjust in LS. Plus the ability to add redundancy and use that in the adjustment.
LS is also a great way to process data.
> That's the thing, LSA is sooooooo easy!!!
>
> In the work flow it's almost non-existent compared to the old days of hours and hours of adjustments, now it's a click, 20 page report generated in .6 seconds.
The threads of the last week have been focused on the minutiae. That may leave the uninitiated with the impression that LS is all about minutiae.
To my mind all those conversations fall within the limits of one golf ball.
How many dimples are on a golf ball?
> How many dimples are on a golf ball?
For a professional golfer it's important to know. The duffer takes it as it comes.
:good: :good:
I'm with Brad and Chi Chi Rodriguez.
I'll bet Chi Chi didn't count the dimples, but he knew how to find the green.
Don
Edit: and let's not forget Lee Trevino who once famously said, "It ain't the arrow, it's the Indian."
I admit that may not be relevant, after thinking about it.
Sorry;-)
> I'll bet Chi Chi didn't count the dimples, but he knew how to find the green.
Now would that be pronounced "Chee Chee", or Ki Ki? And can you shorten his name and just say Chi-squared Rodriguez?
(And of course would a Chi-squared test be a simple question like "how many dimples are on a golf ball?").
I better log off before I make more bad jokes.
That's what Chi Chi would like to have you think. But the reality is different.
Remember Les Nessman from WKRP in Cincinatti? His attempt was something along the lines of Chigh Chigh Rawdruhkweez.
