In its current form, fmi relies on rwfs for much of the functionality. It only implements one subclass and rest is setting up things. The generic architecture in rwfs is useful in principle, but it is painful to develop fmi... The class inheritance is unclear even for me and figuring what gets called and when is obscure. I've been struggling with this for years and I just don't thing the generic nature of rwfs justifies the complexity. Especially so since rgdal has improved a lot in terms of WFS access since rwfs was first created and sf is making things much easier. Much of the WFS-specific functionality that was in rwfs is just not needed anymore. Thus, I'm thinking about doing a re-write of the whole fmi package and retiring rwfs completely.
In its current form,
fmirelies onrwfsfor much of the functionality. It only implements one subclass and rest is setting up things. The generic architecture inrwfsis useful in principle, but it is painful to developfmi... The class inheritance is unclear even for me and figuring what gets called and when is obscure. I've been struggling with this for years and I just don't thing the generic nature ofrwfsjustifies the complexity. Especially so sincergdalhas improved a lot in terms of WFS access sincerwfswas first created andsfis making things much easier. Much of the WFS-specific functionality that was inrwfsis just not needed anymore. Thus, I'm thinking about doing a re-write of the wholefmipackage and retiringrwfscompletely.Any thoughts?