Well, that was not the most fun I've had recently, but it had to be done. The task was to plot up a tract of land that was surveyed in 1974 and conveyed by a metes and bounds description traversing 340 (yes, 340) sides. The deed of conveyance as recorded was not a masterpiece of legibility, consisting of pages of single-space run-on typing, several courses to a line, that hadn't evidently been all that great when presented for record since the County Clerk had felt obliged to hit it with the "Original Instrument Not Legible" stamp before recordation.
If there was an easier way to have done iit, I'd like to know for future reference, but the best method that I could come up with was to convert the description to ascii text by OCR software and then edit the text to turn it into a Star*Net input file. It was still a considerably tedious task, but was easier to debug the data entry and rerun the whole works in Star*Net than by any other method I've used to date.
A sample of the description:
The client is certainly getting his monies worth on this one, eh?
And I thought we had some wild legals....
did it close?
I probably would have slugged through it course by course in CAD marking each one out as I went along. Then I would have thrown my computer out the window when I got to the end and it didn't close by 500 feet. The copy you posted was pretty clear. Lucky for you it was OCRable.
Same here. Would help to have one person read it aloud and a second recite it as it's being entered as a double check. Slow but effective. Would also want to number each course and so indicate on the drawing to simplify any edits that might be needed.
Were there any errors in the original? Something like 68 seconds or 372 degrees?
A large number of very short distances pretty much negates the meaningfulness of the seconds listed. Plus or minus several seconds and you're still closer than the bubble allowed.
Man...that's a LOT of pebble piles to look for!
o.O
Cameron Watson PLS, post: 361490, member: 11407 wrote: Lucky for you it was OCRable.
I did Photoshop the images of the pages to get the lines of typing parallel with the frame of the page and erase some clutter. The OCR worked a heck of a lot better on the parts of the text where the characters weren't faint or burned through. The simple word "feet" got translated into what would pass for at least nine different languages, including "foet" and "feot", which I assume are Flemish and Old Icelandic.
The single most misread character was the "8". That tended to come out as "0". It would have worked much better with text that was at least of quality sufficient for recordation. As it was, parts of the description were missing and had to be recovered from an earlier Notice of Lis Pendens that incorporated the same description, but one that was too faint to scan well.
Loyal, post: 361494, member: 228 wrote: Man...that's a LOT of pebble piles to look for!
Actually, nearly every one of those corners was marked with a 1/2-inch iron rod. Considering that was 1974, I'd say that they probably ran a traverse through nearly all of them on true line, with hardly any sideshots.
astrodanco, post: 361481, member: 7558 wrote: The client is certainly getting his monies worth on this one, eh?
I like to think so. :>
You didn't say if the written legal actually closed very well or not. At least 1:10,000?
Cameron Watson PLS, post: 361500, member: 11407 wrote: You didn't say if the written legal actually closed very well or not. At least 1:10,000?
I'm still working on debugging the description. There was at least one obvious quadrant error in it. I'll mention what the final closure turns out to be when I get there.
Kent McMillan, post: 361479, member: 3 wrote:
If there was an easier way to have done iit, I'd like to know for future reference, but the best method that I could come up with was to convert the description to ascii text by OCR software and then edit the text to turn it into a Star*Net input file. It was still a considerably tedious task, but was easier to debug the data entry and rerun the whole works in Star*Net than by any other method I've used to date.
Could you share a sample of what the Star*net input file looked like? And what was the output? I thought Star*net didn't do you much good without redundant data?
rfc, post: 361502, member: 8882 wrote: Could you share a sample of what the Star*net input file looked like?
The entry was in MAP MODE. A generic entry format for a closed loop beginning and ending on 100 would be:
.MAP ON
C 100 [Northing] [Easting] ! !
TB
T 100 [Bearing] [Distance]
T 101 [Bearing] [Distance]
[...]
T 430 [Bearing] [Distance]
TE 100
Cameron Watson PLS, post: 361490, member: 11407 wrote: I probably would have slugged through it course by course in CAD marking each one out as I went along. Then I would have thrown my computer out the window when I got to the end and it didn't close by 500 feet. The copy you posted was pretty clear. Lucky for you it was OCRable.
My thoughts exactly. Imagine. Then having go back through each course to make sure you didn't type something in wrong.... just thinking about it is making my hair jump out of my scalp to save itself from me pulling it out
rfc, post: 361502, member: 8882 wrote: I thought Star*net didn't do you much good without redundant data?
Least squares is no better than compass rule without redundant data. But the Star*Net program can double as a deed checker by calculating your closure and letting you explore possible blunders.
Rich., post: 361507, member: 10450 wrote: My thoughts exactly. Imagine. Then having go back through each course to make sure you didn't type something in wrong.... just thinking about it is making my hair jump out of my scalp to save itself from me pulling it out
That is one of the advantages to using Star*Net to make the calculation. The input file is simple ascii that is run as a batch, so fixing entry errors is just a matter of editing the entry file and rerunning it.
StarNet is great as a sort of Cogo program on something like this.
Kent McMillan, post: 361509, member: 3 wrote: That is one of the advantages to using Star*Net to make the calculation. The input file is simple ascii that is run as a batch, so fixing entry errors is just a matter of editing the entry file and rerunning it.
Hmm. I keep hearing a few talk about this. I'll have to check it out
Rich., post: 361511, member: 10450 wrote: Hmm. I keep hearing a few talk about this. I'll have to check it out
I just did. Very cool indeed. And really transparent; easy to edit.
Is there a way to do an inverse within Starnet? I mean other than doing the math manually on the Coordinate Output file?
