Date: Thursday, July 21, 1pm EDT (-4 UTC)
Prefer: return=representation; "info:apix/image-manip
".). Explore?Design overview and Extension Definition & Binding docs
Overview doc was reviewed and edited. A separate topic for API-X and URI emerged from the discussion around the overview doc (see agenda item #3) → completed the review. Further edits and updates will have to be done via pull requests.
Extensions Definition and Binding doc has been wordsmithed and is available/ready for comment. Commenting will be closed by the end of the week.
There’s a section that is incomplete, and immature, we will check it as incomplete into GitHub and flesh that out as Pull Request.
Andrew: has the EDB doc received the adequate amount of attention for the closing deadline? Is this pace effective?
Aaron: Overview doc is inherently more attractive to the wider community audience. So other documents may have a more limited audience.
Finalizing design doc at the pace of 1/week basis. → we’ll keep the pace.
Review of other design docs:
Execution and Invoking Service doc will be done first (next week) to give people more time to digest Service Discovery and Binding doc, which is scheduled to be finalized in the week after next week (week of 8/1).
URIs and API-X doc (emerged from discussion on the overview doc):
While this is not on the schedule to be reviewed and commented on, it would be valuable to add this to the list of design doc to be reviewed and agreed, as well as committing this to GitHub → others agree
We will start working on this document now, while people are engaged. Next week is a soft deadline
Unknown User (acoburn) thinks much of URI discussion is out of scope as design, is more of a deployment issue than it is a design issue → others agree, think this discussion is worth having and helps lay out options for API-X from a systems perspective.
API-X chooses to pass Host header to Fedora as an implementation decision. CLAW re-writes URIs and exposes opaque IDs. URI doc mentions both strategies. Should make it clear that this is an implementation decision, and both approaches are valid (with different sets of plusses and minuses, and different operational consequences).
Alternative method of exposing services using Prefer header:
Unknown User (acoburn) does not have favorable opinion for this method out of practical concern in implementing functionality, only brought it up as an option based on the comments from Rob Sanderson.
Suffix approach as documented is really easy and straightforward to implement
Collisions with existing LDP resources not a significant concern. Using reserved characters reduces risk of collision to near 0. If the issue is observed in practice, can mitigate by sending HEAD request (or otherwise verifying existence, non-existence of conflicting LDP resource)
Non-developer overview doc:
Ruth Duerr has not been available to follow-up on comments. By the next meeting, we’ll aim to have a more polished version of the doc to be reviewed by the group.
TODOs summary:
Group to review Extension Definition and Binding & URIs docs
Aaron to wordsmith Execution and Invoking Services & Service discovery and Binding docs