Log of the #fcrepo channel on chat.freenode.net

Using timezone: Eastern Standard Time
* ff_logger_ joins04:00
* jonathangee leaves05:42
* ff_logger_ leaves05:47
* jonathangee joins05:50
* dwilcox joins06:43
* coblej joins08:38
* peichman joins08:58
* mikeAtUVa leaves09:02
* whikloj joins
* acoburn joins09:20
* mikeAtUVa joins
* acoburn leaves09:22
* coblej leaves09:59
* coblej joins10:00
* coblej leaves
* coblej joins10:01
* coblej leaves10:07
* awoods joins
* coblej joins
* bseeger joins10:12
* coblej leaves10:28
* coblej joins
* mikeAtUVa leaves10:35
* coblej leaves10:38
* coblej joins
* coblej leaves10:51
* coblej joins
* coblej leaves10:54
* kefo joins10:57
* ajs6f joins10:58
<whikloj>:( no more smooth jazz10:59
* coblej joins
* coblej leaves11:00
* coblej joins
* acoburn joins
* dshalvi joins11:01
* coblej is here
* yinlin joins
* escowles joins
* escowles is here11:02
<yinlin>is here, no mic
* ajs6f is here
<acoburn>*is here*
<yinlin>that's me
no mic, soory
<kefo>CAn you still register for Code4lib? It usually maxes out.11:09
OR at least often maxes out.
* dhlamb joins11:11
<escowles>i can confirm that hydra is using If-Unmodified-Since, since 11/21/2016: https://github.com/projecthydra/ldp/commit/36520b2a78744f0610d75702384330c43bbf8e2111:16
* mikeAtUVa joins11:28
<whikloj>escowles: yes yes yes
<bseeger> is dialing in…
that's never stopped me11:37
* peichman leaves11:54
* peichman joins
<mikeAtUVa>regarding import/export: I'm happy to lead/work on a concerted effort to pare down and implement the minimal viable product from our point of view in the second half of March or in April.11:56
Doran: after the first planning meeting on that sprint we should circle up on finishing migration utils.11:58
awoods: indeed.12:00
<ajs6f>Are they paying for travel?12:02
<acoburn>awoods: all of these things will be challenging in migrating content between Fedora-on-Modeshape and Trellis: https://github.com/trellis-ldp/trellis/wiki/Comparison-of-Trellis-and-Fedora412:03
<ajs6f>got to run12:04
<awoods>acoburn: we will want address those differences... and import/export is an appropriate context for doing that.12:05
<acoburn>awoods: some of those differences are features
<awoods>acoburn: which do you have in mind?12:06
the lack of referential integrity
the rules around PUT-on-create12:07
the use of link headers in resource creation
the lack of server-managed triples
the ability to create non-container LDP-RSs
being strict about certain LDP properties and their cardinality12:08
for example
awoods: that is, when the LDP spec says MUST I see that as a MUST; Fedora-on-Modeshape doesn't12:09
i.e. for certain ldp properties on direct/indirect containers
that is, in Fedora-on-Modeshape, you can create an indirect container w/o a ldp:membershipResource property — it won't work properly, but you can create it; in trellis you can't12:10
that's just one example12:11
awoods: plus, the policies on resource constraints are pluggable in trellis, so another deployment could have an entirely different set of rules governing "what is a valid RDF graph for a resource"12:15
* dwilcox leaves12:42
* dwilcox joins12:43
* dwilcox leaves
* dwilcox joins12:44
* dwilcox leaves
* dwilcox joins
* dwilcox leaves12:45
* yamil joins12:51
* coblej leaves12:57
* yinlin leaves13:02
* coblej joins13:17
* ajs6f leaves13:32
<awoods>acoburn: In cases where the LDP spec says MUST, fedora-on-modeshape should be doing that.14:03
acoburn: the other cases, we should address one by one, and update the spec as required.
<acoburn>awoods: IMO, the spec should be silent on some of those matters14:04
<awoods>acoburn: import/export is a reasonable bar for implementation inter-op
<acoburn>awoods: honestly, implementation interop is not a goal I have
<awoods>acoburn: sure, and not all of the points mentioned are blockers
acoburn: what would you consider your goal?14:05
<acoburn>awoods: writing a reliable repository platform that works with existing clients
awoods: which is different than import/export interop14:06
<awoods>acoburn: sure, I am talking about the spec
<acoburn>awoods: the question is whether the spec needs to define every behavior or whether the spec allows for variation14:07
awoods: for example: server managed triples
awoods: I like that the spec is silent on that matter
awoods: also: the spec is silent on single-subject-restriction14:08
awoods: I will not have that restriction
awoods: but that means that some resources couldn't be exported from trellis into Fedora-on-Modeshape
awoods: and I think the spec should be silent on that matter14:09
<awoods>acoburn: We probably differ on some points, but it is helpful to know your thoughts.14:10
<acoburn>awoods: my overall thought is that for anything topic on which the spec is silent, there is a tacit understanding that variation is possible14:11
<awoods>acoburn: agreed
<acoburn>awoods: in almost every such case, I am choosing a very different behavior from the existing Fedora software14:12
awoods: which should be permissible
awoods: if it isn't permissible, then that behavior should be part of the spec
<awoods>acoburn: agreed14:13
acoburn: I believe the spec is still in-progress
<acoburn>awoods: also, as an FYI, these are the IRIs I have needed to define in addition to what is in the Fedora ontology: https://acdc.amherst.edu/ns/trellis
awoods: that's the definition of a "draft specification"14:14
<awoods>acoburn: would it make sense to move some of those into the fedora ontology?
* dwilcox joins
<acoburn>awoods: only if they are part of the spec
awoods: for instance, I like being able to define the constraints in an external document, but another impl may have a different set of ldp:constrainedBy links14:15
* th5 joins
<acoburn>awoods: unless you want to define all the ways in which a Fedora server can be constrained (which I think would be a bad idea)14:16
awoods: btw, I like the way I'm producing the RDF/HTML ontology, if you're interested in copying it: https://github.com/trellis-ldp/trellis-ontology/blob/master/scripts/ttl2others.py14:18
awoods: bseeger wrote most of that
awoods: it means I can produce static documents for a web server and respond with TTL, JSON-LD, XML and HTML based on Accept header14:19
awoods: and the ontology itself is written in TTL — the other serializations are generated by a script
* ajs6f joins14:24
* ajs6f leaves14:28
* jjtuttle leaves14:32
* dshalvi leaves14:58
<awoods>acoburn: re:rdf/html ontology... thanks for the tip.15:01
* coblej leaves
* mikeAtUVa leaves15:02
<acoburn>awoods: I'm using nginx, and the config to support conneg is super easy
awoods: but I suspect it's the same for httpd15:03
awoods: you can try it with curl on the ontologies I've got there: https://acdc.amherst.edu/ns/
<awoods>acoburn: I will give it a spin15:04
* bseeger leaves
* bseeger joins15:06
* escowles leaves15:08
<acoburn>awoods: https://gist.github.com/acoburn/bf2426799e1a53e4f1d83ef8d23de18715:18
* coblej joins15:20
* bseeger leaves15:23
* bseeger joins15:46
* bseeger leaves
* bseeger joins15:47
* acoburn leaves15:48
* bseeger leaves15:52
* bseeger joins15:56
* bseeger leaves16:00
* th5 leaves16:17
* yamil leaves16:44
* coblej leaves17:15
* peichman leaves17:27
* travis-ci joins17:55
ualbertalib/fcrepo4-oaiprovider#70 (fcrepo4-oaiprovider-4.2.1 - c78d5f6 : Piyapong Charoenwattana): The build passed.
Change view : https://github.com/ualbertalib/fcrepo4-oaiprovider/compare/fcrepo4-oaiprovider-4.2.1
Build details : https://travis-ci.org/ualbertalib/fcrepo4-oaiprovider/builds/207185447
* travis-ci leaves
* whikloj leaves18:02
* dwilcox leaves18:07
* dwilcox joins18:09
* kefo leaves18:31
* mikeAtUVa joins20:55
* dwilcox leaves21:20
* mikeAtUVa leaves21:26
* peichman joins22:34
* peichman leaves22:35
* awoods leaves23:00

Generated by Sualtam