On 2017-09-07 14:46 +0100, Andrew Atkinson via Therion wrote: > On 07/09/17 14:23, Wookey via Therion wrote: > > I've wanted to do this too, but wasn't sure how. > > > > Alternatively you can give them different names (with variances), then > > equate them. > > Okay that probably will get round it, but is that the right solution? As > this entrance has 3 locations it gets a variance, but most of the other > 9 entrances only have one fix. This would mean they are deemed perfectly > correct, while the ones with more than one location will be moved by the > averaging and the surveys that connects all the entrances together. So > in this case I could just go through and give a the variances for all > the fixes in the data files (it is only a small set.) However it is part > of a very large data set, with something like 200 entrances, that will > take me sometime (or a script.) Now one of the answers which I might > start to do is always be explicit in the variance, however, is it really > reasonable for survex and therefore therion to assume a fix is perfect, > we know that they are not.
No. Almost all fixed points really have some sort of variance. Having standard 'GPS' assumptions for GPS points would be helpful. We really should be putting some numbers in for this until Survex/Therion does it for us. Fixed points coming off map benchmarks or laser surveys could have very small variances which might be close enough to zero that leaving them as zero is not too innacurate. > 2 gps locations fixed by survey legs all the > error is distributed to the survey legs, does not seem right. > So what I'm suggesting is that the default for fix to be perfect should > be looked at and maybe amended to almost perfect variance. Agreed, or at least some shorthand for 'type of fix'. Wookey -- Principal hats: Linaro, Debian, Wookware, ARM http://wookware.org/
signature.asc
Description: Digital signature
_______________________________________________ Therion mailing list Therion@speleo.sk https://mailman.speleo.sk/listinfo/therion
