Log of the #duraspace-ff channel on chat.freenode.net

Using timezone: Eastern Standard Time
* jcoyne joins08:23
* jcoyne leaves08:28
* dwilcox joins08:33
* dwilcox leaves09:08
* dwilcox joins09:11
* mikeAtUVa joins09:13
* ksclarke joins09:14
* ermadmix joins09:20
* jcoyne joins09:33
<pivotal-bot_____>A. "Ropsamingo" Soroka added comment: "Maybe there's a ticket for MODE here, asking for a way to clear projection config from the repo b..." https://www.pivotaltracker.com/story/show/6730811009:51
* scossu joins09:55
* github-ff joins10:12
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/Xq043w
fcrepo4/master cd363f9 Kai Sternad: Fix PATCH not working in Transactions...
* github-ff leaves
<pivotal-bot_____>Andrew Woods added comment: "Resolved with: https://github.com/futures/fcrepo4/commit/cd363f97eea67d169f01b79c93517e0c0b9a75ca" https://www.pivotaltracker.com/story/show/6626285610:13
Andrew Woods added comment: "Resolved with: https://github.com/futures/fcrepo4/commit/cd363f97eea67d169f01b79c93517e0c0b9a75ca" https://www.pivotaltracker.com/story/show/64058980
Andrew Woods delivered "Adding properties to object in a transaction has no effect" https://www.pivotaltracker.com/story/show/64058980
Andrew Woods delivered "PATCH not working in Transactions" https://www.pivotaltracker.com/story/show/66262856
Andrew Woods added comment: "@stefanoc, would you mind verifying this fix?" https://www.pivotaltracker.com/story/show/6405898010:14
* scossu leaves
<pivotal-bot_____>Andrew Woods added comment: "A MODE ticket is probably in order. I would like to dig under the covers a bit more to create a clearer pic..." https://www.pivotaltracker.com/story/show/6730811010:17
Andrew Woods added "Invisible children in Transactions" https://www.pivotaltracker.com/story/show/6775443810:23
Andrew Woods added "Unhandled exception on 404 within Transactions" https://www.pivotaltracker.com/story/show/6775468610:25
Andrew Woods added "Invisible properties in Transactions" https://www.pivotaltracker.com/story/show/6775504410:28
Andrew Woods edited "Unhandled exception on 404 within Transactions" https://www.pivotaltracker.com/story/show/67754686
Andrew Woods edited "Invisible children in Transactions" https://www.pivotaltracker.com/story/show/67754438
Andrew Woods edited "Invisible properties in Transactions" https://www.pivotaltracker.com/story/show/67755044
* travis-ci joins10:36
[travis-ci] futures/fcrepo4#1634 (master - cd363f9 : Kai Sternad): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/1288353eee36...cd363f97eea6
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/21018501
* travis-ci leaves
* jcoyne leaves10:37
* ajs6f joins10:54
* github-ff joins10:56
[fcrepo4] escowles pushed 1 new commit to upload-to-projection: http://git.io/pJ9Siw
fcrepo4/upload-to-projection b044ffe Esmé Cowles: Trapping errors when broken links to non-existent repository content (e.g. versions associated with federated filesystem)
* github-ff leaves
<pivotal-bot_____>Esme Cowles added comment: "I think the key issue here is that the filesystem federation is mapping directories to fedora:resource, whic..." https://www.pivotaltracker.com/story/show/6766365810:58
<escowles>ermadmix: standup?11:00
<pivotal-bot_____>Andrew Woods started "Test access control on federated filesystems" https://www.pivotaltracker.com/story/show/6755753211:07
Andrew Woods edited "Test access control on federated filesystems" https://www.pivotaltracker.com/story/show/67557532
<ajs6f>awoods: Is there a day/time of which you were thinking for the "special topic" call?11:23
<awoods>ajs6f: the sooner the better.11:24
<ajs6f>Thursday 2PM ET?
<awoods>ajs6f: I have a meeting then, but Thursday 12 - 2pm is free.11:26
<ajs6f>awoods; What about 4PM, after the sprint close?
<awoods>ajs6f: 4pm is good11:28
<ajs6f>Then I'll send a message, if you think we're enough.
* travis-ci joins
[travis-ci] futures/fcrepo4#1635 (upload-to-projection - b044ffe : Esmé Cowles): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/39c5e16da413...b044ffedcbea
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/21021931
* travis-ci leaves
<awoods>ajs6f: send a note and see what happens.11:29
<ajs6f>The afulness of ReadyTalk, I suppose?
<awoods>ajs6f: we will likely not have more than 10 people on the call... so g-hangout could be an option.11:30
<ajs6f>awoods: _Fine by me_.
* scossu joins11:32
* github-ff joins11:39
[fcrepo4] ajs6f pushed 1 new commit to GenericFcrepoTransform: http://git.io/2lRfQA
fcrepo4/GenericFcrepoTransform cb541f1 ajs6f: Improved message for mismatched type exception
* github-ff leaves
<pivotal-bot_____>A. "Ropsamingo" Soroka started "Make Transform and TransformFactory in fcrepo-transform use generics instead of reflection" https://www.pivotaltracker.com/story/show/67483878
A. "Ropsamingo" Soroka finished "Make Transform and TransformFactory in fcrepo-transform use generics instead of reflection" https://www.pivotaltracker.com/story/show/67483878
<ajs6f>awwods: ^^^ all yours
* scossu leaves12:00
* travis-ci joins12:01
[travis-ci] futures/fcrepo4#1636 (GenericFcrepoTransform - cb541f1 : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/7f5a343f12c6...cb541f1413ce
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/21025372
* travis-ci leaves
* edInCo joins12:22
* scossu joins12:52
* scossu leaves13:03
* ermadmix leaves13:07
* scossu joins13:08
* barmintor joins13:11
* scossu leaves13:13
* scossu joins13:14
* dwilcox leaves13:23
* dwilcox joins
* tecoripa joins13:25
* fasseg joins
<tecoripa>ajs6f: do you mind if I make a few edits to your proposal, to clarify some points?13:26
<fasseg>Hey guys! I have a question about fedora 3: When using the API-M method export I can�'t see any difference between the econtext values public and migrate..What am I missing there?
<ajs6f>Every edit costs you a beer.13:27
<tecoripa>no, every edit costs YOU a beer
<ajs6f>fasseg: I think it's got something to do with external dses, but let me check.
<ajs6f>tecoripa: Bbbbbbut~! How did you do that? You turned it around on me!
* dwilcox leaves13:28
<tecoripa>ajs6f: what is "monotonically less-restrictive"?
<ajs6f>tecoripa: Each additional of a type always means LESS restriction, never more restriction.
fasseg: • migrate: selecting the migrate context will ensure that the export file is encoded in a manner appropriate for migrating an object from one Fedora repository to another. The important feature to note here is that when a Fedora object's datastream contains a URL that is relative to the repository in which the object resides, the migrate option will ensure that the URL remains relative in the new repository. Any repository-re
• public: selecting the public context will ensure that the export file is encoded in a manner appropriate for processing the digital object outside the context of a Fedora repository. The file will contain callback URLs for all content stored inside the Fedora repository. The assumption is that this file must not contain any internal identifiers that are only meaningful inside a Fedora repository. Instead, all internal identifiers t
• archive: the archive context is not available in Fedora 2.0. In the future this option will be enabled to produce an export file that is self-contained, meaning that all datastream content is within the export file. The file will serve as a self-contained archive of the digital object. All XML content will be directly inlined within the XML export file. All binary content will be base64-encoded within the export file. There will be
Sorry, all.
<tecoripa>OK, that's what I thought...13:29
<ajs6f>fasseg: That's from the 2.x docs, but I think it's all still true.
<tecoripa>here comes the first beer you owe me...
<fasseg>ajs6f: Thanks: SO I should se a difference in the URIs when using public and migrate
<ajs6f>fasseg: I think so. You should see "fedora.local.server" show up.13:30
<tecoripa>ajs6f: what is a "pattern of recordation"?
<ajs6f>tecoripa: The manner in which a typed resource is persisted in order that when it is recalled, it maintains type information.13:31
<pivotal-bot_____>Esme Cowles added comment: "Removing mix:versionable from the fedora:resource definition fixes the problem. I can now create folders an..." https://www.pivotaltracker.com/story/show/67663658
<ajs6f>tecoripa: E.g., with metadata, with mixins, in some external database...13:32
* ermadmix joins
<tecoripa>ajs6f: ok, I understand... ka-ching, here comes my second beer
<ajs6f>tecoripa: You'll bankrupt me!
<tecoripa>ajs6f: depends on how soon I pass out.13:33
<ajs6f>awoods: Forget about those doughnuts I keep bugging you for. You should start bringing beer to the sprint close meetings.13:34
<fasseg>and another question from the folks at the training day today, which I couldn't answer completely was if it's possible to version Datastreams of Control Group E. I supsected that it's possible and fedora 3 will just add a new datastreamversion element into the FOXML, but I wasn't sure. Can you guys confirm that?13:35
<tecoripa>ajs6f: primary-type/mixin-type: is primary-type your standard inheritance model (classes, sub classes, etc.?)13:36
<ajs6f>tecoripa: Primary type is a single-inheritance hierarchy, if that's what you mean...?13:37
<tecoripa>ajs6f: yes. I'll give you a pass on that beer.
<awoods>good news, escowles.13:40
re: https://www.pivotaltracker.com/s/projects/684825/stories/67663658
<escowles>awoods: i just figured out that mix:referenceable was being added by mix:versionable, so most of the test failures are actually b/c of mix:referenceable being gone13:41
so i'm adding that back in to fedora:resource and removing most of my test updates
<tecoripa>ajs6f: associating SPARQL queries to a mixin: what is the mechanics of that? do the queries become attributes of the type?
<ajs6f>tecoripa: Undetermined as yet. I'm not going to do much design work until the architecture gets some acceptance, and if it doesn't, then I won't do any at all. The type itself is persisted in system space, so we neither can nor should alter it. We could store the queries in fedora:system portion of the repo and just name them after the types.13:43
tecoripa: Or we could introduce a further layer of indirection, if we thought it useful.
<tecoripa>ok, understood.13:44
ajs6f: last question: what is disjoint?
<ajs6f>Either one or the other but not both...
A pair of disjoint types is a pair such that a resource can be one or the other but not both.13:45
And you can see the generalization to n types.
Think set theory.
<tecoripa>right, okay.
is that the principle that prevents multiple inheritance for primary types?13:47
<ajs6f>tecoripa: No.13:48
Well, sort of.
Primary types are not all disjoint to each other.
<tecoripa>ok... only some are...
<ajs6f>E.g. if I have a type A and a subtype B, and a resource in the subtype, it is in the type A.13:49
But sure, if I have a type A and a type B (neither a subtype of each other) then they are disjoint, but the mechanism is simply that a resource can have only _one_ primary type recorded.13:50
<tecoripa>I think I see. I'm trying to unpack the sentence "It is inherent in the JCR notion of "primary-type" that primary types are disjoint within their hierarchy."
<ajs6f>That "one primary type written down per node" thing is the mechanism.
* dwilcox joins
<ajs6f>So that sentence means "A node can have only one primary type recorded" -> "the only primary types a node can be in are the one that is recorded for it and any supertypes of that type" -> "primary types are disjoint with all other primary types except their own super- and sub- types" ->"It is inherent in the JCR notion of "primary-type" that primary types are disjoint within their hierarchy."13:52
<tecoripa>ajs6f: I'm trying not overthink this: is disjointness within a hierarchy the same principle that prevents classes from having more than one parent class in java, for example? I'm trying to picture this somewhat less abstractly.13:53
<ajs6f>Actually, that is a confusing way of saying it. Maybe it would have been better to jsut say "primary types are disjoint with all other primary types except their own super- and sub- types" and hve left it at that.
tecoripa: I'd rather say that the JCR thing is relying on the same principal that allows Java _objects_ to have a single "chain" of classes and superclasses.13:54
<tecoripa>right, okay. good distinction.
<ajs6f>Node -> Java Object , Node type -> Java type, primary type -> Java class or abstract class, mixin -> Java interface.
Not principal.
<tecoripa>I'm going to try to give a little more explanatory text on the notion of disjointness, for tyhose of us without a CS or logic background.
<ajs6f>tecoripa: Good on ya.
<tecoripa>thanks for the romper-room example -- that's what I pictured, but I wanted to make sure I understood it correctly.
<ajs6f>tecoripa: This is all coming out of the ARTIC use case and similarly powerful ones.13:56
I was trying to get at the notions underneath.
There's nothing romper room about type theory.
<tecoripa>ajs6f: does this make sense:? "Disjointness is the principle that inheritance passes through a single line, in a tree hierarchy;  in practice, this means that JCR nodes can only have a single primary type (although that primary type may inherit from a supertype, etc.)"13:58
The fact that "JCR nodes can  only have a single primary type" is independent of that.
<ajs6f>They're two different things. Let me see if I can find the Java analog.
So there are three different things going on here, and we have to distingish them.14:00
1) SIngle-line inheritance.
2) Disjointness
3) Singly-recorded types
(on instances)
<ajs6f>And in the case of JCR, from the first and third, in some cases (in the case of primary types) we can deduce the second.14:01
But you can have the second independently.
COnsider the OWL construct disjointWith.
That is a pure expression of disjointness. It's got nothing to do with hierarchy.14:02
<tecoripa>ok. I'm doing some browsing...
<ajs6f>It's the simple assertion that a thing may be A or B or neither, but it's never both A and B.
<tecoripa>and disjoint simply means no overlap, correct?
Think Venn diagram.
<ajs6f>Types are the sets (or proper classes) of the things that have that type.14:03
And sometime they overlap in membership, and sometimes they don't.
<tecoripa>and if they don't, then they're disjoint.
<ajs6f>Sometimes they overlap completely, and that's owl:equivalentClass14:04
If they don't, those types are disjoint. Yes.
<tecoripa>and disjointness is an artifact of the single-line inheritance...?
<ajs6f>That _and_ the fact that a node can have but one primary type recorded for it.
Let's pretend that wasn't the case.
<tecoripa>(light goes on)14:05
<ajs6f>So we have a hierarchy of types.
And they have single-line inheritance.
But we happily assign three of them to one node, because we can now do that in our magical fantasy world.
And now those types are _not_ disjoint.
They overlap.
ok, let me read your note on disjointness again14:06
<ajs6f>But we have both conditions, and together they allow us to deduce disjointness.
<tecoripa>and see if i understand the implications of that.
<ajs6f>If you do, let me know. I'd love to know what any of that gobletyggook means.
<tecoripa>okay, try this on for size (not sure if this is making thing clearer or not): "y virtue of the fact that JCR graphs are trees (which means nodes have a single line of inheritance), and that JCR only allows a single primary type to be specified for a given node, all JCR primary types are disjoint:  that is, nodes of a given primary type can never be associated with another primary type not in the same line of inheritance."\14:14
<ajs6f>tecoripa: Sounds a hell of a lot clearer than what I wrote.14:15
afk bbl14:16
<tecoripa>ajs6f: ok thanks. I'll edit and save now.\
<ajs6f>Hang on!
<tecoripa>and cancel you beer debt to me.
<ajs6f>One correction:
It's not "the fact that JCR graphs are trees (which means nodes have a  single line of inheritance)". It's "the fact that JCR _types_ form a tree (which means _types_ have a  single line of inheritance)".14:17
That's the single-line inhertiance that matters here.
<ajs6f>Nodes cold inherit from twelve diff parent nodes and that wouldn't change the disjointness of primary types.14:18
Nodes cold be in an arbitrary graph.
<tecoripa>yes, right...
<ajs6f>Okay, afk bbl for reals.
<tecoripa>it';s the type hierarchy that matters here
It's _always_ the type system that matters. It's the ONTOLOGY!
* ajs6f leaves
* github-ff joins14:38
[fcrepo4] escowles created default-unversioned (+1 new commit): http://git.io/KmPaTQ
fcrepo4/default-unversioned 01790aa Esmé Cowles: Removing mix:versionable from fedora:resource definition, which disables automatic versioning by default
* github-ff leaves
* github-ff joins14:40
[fcrepo4] escowles opened pull request #268: Removing mix:versionable from fedora:resource definition (master...default-unversioned) http://git.io/3Hnwag
* github-ff leaves
<pivotal-bot_____>Esme Cowles added comment: "Adding mix:referenceable to the fedora:resource definition (instead of mix:versionable) greatly reduced the ..." https://www.pivotaltracker.com/story/show/6766365814:41
Esme Cowles finished "Uploading to federated filesystem via REST API breaks browse (but not content)" https://www.pivotaltracker.com/story/show/67663658
Esme Cowles added comment: "Now that 67663658 is fixed, I can add an IT for copying between the repo and federated filesystem, and test ..." https://www.pivotaltracker.com/story/show/6715451014:43
* github-ff joins14:58
[fcrepo4] escowles pushed 1 new commit to default-unversioned: http://git.io/Ax1vtQ
fcrepo4/default-unversioned d6acff5 Esmé Cowles: Adding integration tests for copying between repo and federated filesystem
* github-ff leaves
<pivotal-bot_____>Esme Cowles added comment: "I've added ITs for copying between the repo and federated filesystem -- added to PR: https://github.com/futu..." https://www.pivotaltracker.com/story/show/6715451014:59
* travis-ci joins15:05
[travis-ci] futures/fcrepo4#1638 (default-unversioned - 01790aa : Esmé Cowles): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/commit/01790aafcbff
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/21039471
* travis-ci leaves
* fasseg leaves15:12
* fasseg joins15:21
* travis-ci joins15:22
[travis-ci] futures/fcrepo4#1640 (default-unversioned - d6acff5 : Esmé Cowles): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/01790aafcbff...d6acff564141
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/21040958
* travis-ci leaves
* tecoripa leaves15:41
* dwilcox leaves15:55
* dwilcox joins
* fasseg leaves15:57
* dwilcox leaves16:00
* fasseg joins16:08
* ermadmix leaves16:11
* ajs6f joins16:19
* ajs6f leaves16:23
* ksclarke leaves16:57
* dwilcox joins
* dwilcox leaves17:03
* fasseg leaves17:07
* mikeAtUVa leaves17:14
* dwilcox joins17:15
* dwilcox leaves17:27
* dwilcox joins17:28
* dwilcox leaves17:32
* barmintor leaves17:36
* dwilcox joins17:40
* dwilcox leaves17:45
* dwilcox joins
* dwilcox leaves17:50
<pivotal-bot_____>Stefano Cossu added comment: "Tested with properties and children, confirm that it works." https://www.pivotaltracker.com/story/show/6405898017:53
Stefano Cossu accepted "Adding properties to object in a transaction has no effect" https://www.pivotaltracker.com/story/show/6405898017:56
* gregjansen leaves18:02
<pivotal-bot_____>Andrew Woods accepted "PATCH not working in Transactions" https://www.pivotaltracker.com/story/show/6626285618:27
* fasseg joins18:28
* fasseg leaves18:32
<pivotal-bot_____>Andrew Woods added comment: "@escowles, could you please update the documentation under the Versioning feature to make clear that mix:ve..." https://www.pivotaltracker.com/story/show/6766365818:34
Andrew Woods added comment: "Documentation to update: https://wiki.duraspace.org/display/FF/Versioning" https://www.pivotaltracker.com/story/show/67663658
* github-ff joins18:40
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/v3R4cw
fcrepo4/master 45a169f ajs6f: Simplify type of LDPath transform result...
* github-ff leaves
<pivotal-bot_____>Andrew Woods added comment: "Resolved with: https://github.com/futures/fcrepo4/commit/45a169fc2f787086fe587ae15eafa1f232994615" https://www.pivotaltracker.com/story/show/67483878
Andrew Woods delivered "Make Transform and TransformFactory in fcrepo-transform use generics instead of reflection" https://www.pivotaltracker.com/story/show/67483878
* github-ff joins18:41
[fcrepo4] awoods deleted GenericFcrepoTransform at cb541f1: http://git.io/cnh4FA
* github-ff leaves
* travis-ci joins18:57
[travis-ci] futures/fcrepo4#1642 (master - 45a169f : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/cd363f97eea6...45a169fc2f78
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/21056029
* travis-ci leaves
* edInCo leaves19:03
* scossu leaves19:13
* dwilcox joins20:36
* dwilcox leaves21:05
* dwilcox joins
* dwilcox leaves21:10
* scossu joins22:22
<pivotal-bot_____>Andrew Woods accepted "Make Transform and TransformFactory in fcrepo-transform use generics instead of reflection" https://www.pivotaltracker.com/story/show/6748387822:37
Andrew Woods edited "Make Transform and TransformFactory in fcrepo-transform use generics instead of reflection" https://www.pivotaltracker.com/story/show/67483878
* dwilcox joins23:06
<pivotal-bot_____>Andrew Woods added comment: "Pending minor code review comments." https://www.pivotaltracker.com/story/show/67663658
Andrew Woods rejected "Uploading to federated filesystem via REST API breaks browse (but not content)" https://www.pivotaltracker.com/story/show/67663658
* dwilcox leaves23:10

Generated by Sualtam