Quality Control Check Rules for SMS Connections

Subject: 	Quality Control Check of SMS connections
Date: 	Tue, 09 Nov 2010 18:36:07 -0500
From: 	Ted Willard <twillard@aaas.org>
To: 	Brian Sweeney <bsweeney@aaas.org>, Francis Molina
<fmolina@aaas.org>, John Weatherley <jweather@ucar.edu>, Mike Wright
<mwright@ucar.edu>



Hi all-
As promised, here are some "things to look for" in trying to rectify
inconsistences between Atlas 1 and Atlas 2 as represented in the Strand
Map Service.
Issues that need to be examined and/or rectified:
1. Benchmark A and Benchmark B appear on Map X and Map Y. They are
connected on Map X but not connected on Map Y
2. Benchmark A and Benchmark B appear on Map X and Map Y, A --> B on Map
X but B --> A on Map Y
3. Across all maps Instances where "A --> B --> C --> A" or "A --> B -->
C --> D --> A" or "A --> B --> C --> D --> E --> A" and so on.
4. A case of A --> B where A is in a higher grade level than B is.
A couple notes:
Any time a conflict is found, I will need to know on which maps each of
the connections is found in order to work through a solution.
There will be instances for number 1 above that are ok, and require no
correction, but it would be a good to examine as part of the quality
control process.
A program that finds the recursive loop
"A --> B --> C --> D --> A",
there would also be a loops for
"B --> C --> D --> A --> B"
"C --> D --> A --> B --> C"
"D --> A --> B --> C --> D".
It would be nice to have the search consider this one problem rather
than four separate ones.
Finally, there are cases where we have decided when doing Atlas 2 and
Benchmarks Online to split a benchmark that appeared as whole in Atlas
1, and also where we took what was two Benchmarks in Atlas 1 and
combined them to form one benchmark in Atlas 2 and benchmarks online. I
am not sure how the strand map service is currently dealing with these
issues.
Let me know if you have any questions or concerns.
-Ted