aarondandy / pigeoid Goto Github PK

The last step in the coordinate operation expects a geographic coordinate but receives a 3D point instead. Not sure if the calculated path is wrong or if more conversions are needed.

Would be nice to have a source of ESRI coordinate systems and "well knows IDs" that can be included in this library without legal problems. For example something may request data in ESRI:102113 and then what?

• Could respond to the request with I dunno that ESRI CRS LOL ¯\(°_o)/¯
• Could use a filthy hack to remap known ESRI CRSs to EPSG equivalent.
• Maybe only accept those as full WKT, but no JS web mapping tool will supply that much detail.

Oh and it gets better because lots of people like to pretend only the EPSG exists so when you get a CRS that is really under the ESRI, OGC, SR-ORG they will force it to EPSG.

Make sure to use this advice when handling web mercator projections:

http://alastaira.wordpress.com/2011/01/23/the-google-maps-bing-maps-spherical-mercator-projection/

Should Math.Pow(Math.E,x) be converted to Math.Exp(x)?

Details: https://portal.opengeospatial.org/files/?artifact_id=8814

It may be possible to create a new "AuthorityTag" for URNs that implements the same interface but exposes a bunch more stuff. The EPSG models should use the new URN tag so version can be included. This may be important when working with B.S. CRSs such as ESRI:102100 and ESRI:102113 and associating them to a true well known identifier. Also the OGC decided it would become an authority but expose the authorities through their "PDF web API" so.... yeah it's a mess. Using the URN format with backwards authority tag support may help though.

I don't know if this would hurt or help organization. There may be lots of tests that have weird supporting files and the tests may not map 1:1 with the projects.

Some ideas in:

http://www.phailed.me/2013/05/introduction-to-scientific-computing-error-propagation/

Use stylecop to enforce a code style with warnings.