Rectify method is not saved

[lklepner]

Hi Tim,

I've noticed that rectifying methods are not saved on the NYPL instance of the Mapwarper and was hoping you might be able to shed some light on why this might be.

I've detailed my findings here -
nypl-spacetime/nypl-warper#27

Was this an intentional design decision for some reason? Or would be make sense to begin preserving/saving the rectify method so future rectification makes use of it?

Best,
Lou

[timwaters]

Hi Lou, thanks for the issue - I think it should be saved - its a good idea. Your example of the river is a great example!

By design if no method is chosen it will automatically choose a polynominal one according to the number of control points, but it wont choose thin plate spline.

We've been discussing how to share the results of a warp between warpers and desktop GIS to make them reproducible - having the rectify method, and resampling method would be needed for that.

[lklepner]

Awesome - thanks Tim. If this is something you're able to adjust in the core codebase I'll try to carry it forward to the NYPL codebase.

[lklepner]

Even for the core mapwarper website itself I worry that not saving the rectify method is causing users who try refining a crop or adjust rectify points to inadvertently deteriorate the quality of an existing rectification by triggering a warp using the default method.. and unless they have indepth experience with the tool, they won't likely know to try switching the method.

Its a little off topic for the issue of saving the rectify method but I'm finding in my experience so far that thin plate spline seems to often times produce better results even with a small number of points.. avoiding the swirl outcome polynominal seems to often times produce.