[PROJ] RFC4: last chance to comment before motion

[Howard Butler]

> On Jan 6, 2020, at 7:34 PM, Even Rouault <even.rouault at spatialys.com> wrote:
> 
> On lundi 6 janvier 2020 16:21:36 CET Even Rouault wrote:
>>> Even in the HARN case, I think it's good for people to be able to know
>>> that "foo-1.0" is broken and "foo-1.0.1" is what they should use,
>>> instead of "I have foo-1.0; is that the ok version or not?"
>> 
> 
> After discussing this topic of grid versionning with Howard, this seems a bit 
> of over-engineering for now, at least for the remote grid fetching part. We 
> might reconsider this in the future, for example if grid producers themselves 
> adhere to stricter versionning.

My worry is PROJ applying its own version naming scheme to grid files will have the effect of the PROJ project creating a new set of canonical names, when they should really be originating from the agencies or the authority.  It isn't PROJ's scope to fix the sins of the grid providers or the authority(s), and if we assume responsibility, we now need to version'ify every grid filename, we need to provide an API to provide a pointer to the "latest" file, and we have to keep a ledger of changes.

> The static snapshots of the git repository (to be determined if we go to a 
> single one or some other subdivision scheme) will remain the main tools to 
> have grid versioning for now.

We have also enabled S3 object versioning on the bucket that hosts the CDN data. Versions aren't public for now, but this will give us protection to rollback in case of a mishap or if it is clear in the future that we need to engineer something to handle the issue. A legitimate worry about expanding this approach is the lock-in for S3/CloudFront, which I agree we should seek to minimize. 

Howard