Re: Minutes of LDPNext Bi-Weekly Meeting #2

Cody et al.--

This is great stuff! Thanks for digging into the If-Match problem, in particular. I'm sorry that so many Fedora folks weren't able to 
attend (the meeting collided with a user group meeting), but let's take the conversation forward here on the email list. I'd like to check a few points recorded in the minutes:

1) "Fedora community (some or many) have been using a set of RDF libraries written in Ruby. And there seems to be some concern that you’re not guaranteed to get triples in a certain order. So - you can have the same “semantic” representation of a resource, but not a byte-by-byte sameness."

It's not to do with Ruby (or any other language in use). The reference implementation of Fedora is currently written in Java, for example, using Apache Jena as the RDF framework. The remark about semantic vs. byte-for-byte sameness is half of the heart of the matter, or as Pierre-Antoine Champin was quoted, "...the order of triples or blank node labels are not significant in RDF and may vary across serializations." and the requirement for strong-ness in the ETags used with If-Match is the other half. But I think the rest of the minutes make clear that everyone understands that.

2) "An possible alternative such as using (`If-Unmodified-Since`) was proposed."

This move was explored by the Fedora community, and it's a useful technique, but it seems to run into a different "spec vs. spec" problem:

https://tools.ietf.org/html/rfc7231#section-7.1.1.1

defines datetimes as used in HTTP (and for If-Unmodified-Since) as precise down to seconds. We have sections of our community (particularly those exploring the management at scale of research data) for whom that's not enough. So If-Unmodified-Since is definitely part of the picture, but we currently suspect that more is needed.

Personally, I think the requirement on If-Match to use strong ETags is really unfortunate at best, and I don't think it's crazy to argue that the strength (even quality) of available comparison functions is going to vary with the negotiated representation, so that it doesn't make sense to require any particular type of comparison (i.e. byte-for-byte) universally. But my opinions are not going to change the HTTP spec. {grin}

---
A. Soroka
The University of Virginia Library

> On Sep 22, 2016, at 5:26 PM, Cody Burleson <cody.burleson@base22.com> wrote:
> 
> Team,
>  
> Please be advised that we conducted our second bi-weekly web meeting today. There were not many attendees, but we took advantage of the time to chew on the “If-Match and Weak ETags” issue together. We do not have any concrete outcomes for that particular issue yet, but we think it was helpful to discuss it. We discussed some other things, but the “If-Match and Weak ETags” issue was the main focus.
> 
> Notes, as well as a meeting recording can be found here:
> 
> https://www.w3.org/wiki/LDP_Next_CG_Meeting_2016-09-22
> 
> Thank you, 
>  
> Cody
>

Received on Friday, 23 September 2016 14:32:42 UTC