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

Using timezone: Eastern Standard Time
<cbeer>hi ff_logger10:53
<awoods>ermadmix: I believe we are already doing a similar hashing/hierarchy structure in F4... but only if the user does not provide a specific id on ingest.10:56
<ermadmix>ajs6f: re 10:25, yes thanks for your patience as I wade through it, starting to understand the approach you took to it now, and it seems to basically do what it sets out to do. There are some things, lingering comments,etc, but I'm hesitant to touch, is there anything you expect me to do before merging?10:57
<bljenkins>Project fcrepo-kitchen-sink build #702: SUCCESS in 6 min 36 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/702/
* osmandin leaves
<ajs6f>ermadmix: If you have lingering doubts, comments, get 'em into Github and I will address them immediately.10:58
* osmandin joins11:01
standup meeting now
barmintor, standup?11:04
<barmintor>mikeAtUVa: no
not today
<mikeAtUVa>barmintor: fair enough...
* osmandin leaves11:06
<pivotal-bot>Osman Din finished "Introspect the bson output for modeshape schematics and report back how friendly the output is" https://www.pivotaltracker.com/story/show/4901279911:08
<cbeer>escowles: how brief is briefly?11:10
<escowles>cbeer: three or four testing runs (~30 minutes or so)11:11
<cbeer>hm. longer than any timeout at least.
<escowles>i made some changes (using names instead of IP addresses) and then it stopped working, and didn't work when i changed them back11:12
<cbeer>i also thought we decided to kill the body11:13
<pivotal-bot>Chris Beer added comment: "Or, do we keep the final slash for "container" resources?11:18
https://dvcs.w3.org/hg/ldpwg/raw-file/default/ldp-..." https://www.pivotaltracker.com/story/show/60883814
Chris Beer added comment: "(we aren't producing relative URIs to child resources.. but maybe we ought to as well..)" https://www.pivotaltracker.com/story/show/6088381411:19
Chris Beer estimated "Update produced fixity result URIs" as 1 point https://www.pivotaltracker.com/story/show/6206329211:20
Chris Beer started "Update produced fixity result URIs" https://www.pivotaltracker.com/story/show/62063292
Chris Beer added comment: "I think I originally suggested using fcr:fixity in that URI, but we don't have the UriInfo in the fixity RDF ..." https://www.pivotaltracker.com/story/show/6206329211:25
Chris Beer added "Add PREMIS vocabulary to fixity results reporting" https://www.pivotaltracker.com/story/show/6224356211:29
* osmandin joins11:31
awoods: it's my understanding that I don't need to do another PR for ATOM11:33
* kaarefc leaves11:34
* kaarefc joins
<cbeer>ajs6f: how much PREMIS do you know?11:35
<ajs6f>cbeer: No more than I've been forced to learn.11:36
But I know people wo know soething of it.
osmandin: Correct— hold off.
<cbeer>ajs6f: i'm trying to figure out what the right way to express our fixity state property is
i suspect it's in http://id.loc.gov/ontologies/premis.html#EventOutcomeInformation
<osmandin>ajs6f: ok, thx
<ajs6f>cbeer: that's beyond my immediate knowledge— you're guess would be at least as good as mine. can you talk to someone in your archives who is into this kind f thing?11:37
<awoods>osmandin/ajs6f: yes, hold-off on ATOM
<ajs6f>awoods: No, _you_ hold off on Atom!
* awoods holding off on Atom11:38
<ajs6f>Nobody moves, nobody gets hurt.
<awoods>osmandin: any news on a bson utility for exposing the ispn persistence files? Is this something that will go into a github project?11:39
osmandin: https://www.pivotaltracker.com/story/show/6207150411:40
<pivotal-bot>feature: Create simple utility that takes ModeShape filesystem binary artifacts (i.e. fcrepo4-data directory as input) and generates textual key-value representation of a particular file (node) or the whole dir (tree). (started) / owner: Osman Din
<osmandin>awoods: I referenced it on the wiki. I'm still working on it, adding basic functionality or features etc...11:41
<awoods>osmandin: I suspect you are planning on adding the project to github?11:42
<pivotal-bot>Mike Durbin added comment: "I created nearly empty pages (with just an excerpt tag to populate the summary on the parent "Features" page..." https://www.pivotaltracker.com/story/show/62234746
<osmandin>awoods: It's in development @ https://github.com/osmandin/modetrace11:43
I'll update the git, and mention the branch in the ticket11:44
<awoods>osmandin: Great. I was just checking. It sounds like you are "on it".11:45
<osmandin>awoods: Yes, that's what I'm working on. I also 'finished' the doc. ticket for your review.11:46
<awoods>osmandin: It is on the list11:47
<pivotal-bot>A. "Horbulaco" Soroka added "Kill legacy (Atom) JMS format" https://www.pivotaltracker.com/story/show/6224623411:56
<ajs6f>afk for lucnh11:59
* ajs6f leaves
<osmandin>afk 3012:00
<pivotal-bot>Eric James added "create user documentation of external search in 'features' section of wiki" https://www.pivotaltracker.com/story/show/6224909012:28
Eric James started "create user documentation of external search in 'features' section of wiki" https://www.pivotaltracker.com/story/show/62249090
Andrew Woods added comment: "@osmandin, this looks very good. I have some inline comments and page-level comments. Please put this ticke..." https://www.pivotaltracker.com/story/show/4901279912:39
Andrew Woods rejected "Introspect the bson output for modeshape schematics and report back how friendly the output is" https://www.pivotaltracker.com/story/show/4901279912:40
Andrew Woods added comment: "@escowles, Please put this ticket back in the "Finish"ed state when you have addressed the comments." https://www.pivotaltracker.com/story/show/6055694812:41
<awoods>mikeAtUVa: Are "transactions" going to make the "Features" list?12:45
<pivotal-bot>Osman Din started "Introspect the bson output for modeshape schematics and report back how friendly the output is" https://www.pivotaltracker.com/story/show/4901279912:48
<awoods>mikeAtUVa: Let's add "Policy Driven Storage" to the feature list as well.12:49
<cbeer>anyone know why DefaultGraphSubjects uses the RESTAPI_NAMESPACE to generate subject URIs?12:53
* ermadmix leaves13:02
<pivotal-bot>Esme Cowles added comment: "@awoods: I've addressed your comments, moved the page into the Performance section and renamed it: https://w..." https://www.pivotaltracker.com/story/show/6055694813:03
Esme Cowles finished "Single-node F4 Ingest Benchmarks" https://www.pivotaltracker.com/story/show/60556948
<cbeer>escowles: ^^^ it looks like you changed DefaultGraphSubjects to use the RESTAPI_NAMESPACE instead of info:fedora13:11
any idea why? (it's breaking my fixity results uri changes..)
<escowles>cbeer: i don't remember changing DefaultGraphSubjects at all...13:12
<escowles>cbeer: oh -- sorry, i thought you meant recently -- i did update a bunch of predicate refs when we changed the namespaces, so it's completely possible i changed one to the wrong namespace by accident13:14
it didn't break the tests at the time -- has something else changed recently that combined with that to make it fail?
<cbeer>escowles: I'm working on code to generate hash-URI fixity results, and we end up with invalid URIs like http://fedora.info/#/path/to/node#fixity-1234513:15
i'm tempted to change it back to info:fedora
<escowles>cbeer: changing back to info:fedora for those URIs seems fine to me -- they're not predicates13:16
* ermadmix joins13:26
<pivotal-bot>Andrew Woods delivered "configuration documentation " https://www.pivotaltracker.com/story/show/6117357813:29
Andrew Woods accepted "configuration documentation " https://www.pivotaltracker.com/story/show/61173578
Andrew Woods accepted "Single-node F4 Ingest Benchmarks" https://www.pivotaltracker.com/story/show/6055694813:30
Andrew Woods accepted "Use new namespace for fedora:versioningPolicy" https://www.pivotaltracker.com/story/show/62191958
Andrew Woods accepted "fcr:transform exception" https://www.pivotaltracker.com/story/show/61942638
<escowles>mikeAtUVa/awoods: my thinking for https://www.pivotaltracker.com/story/show/62234746 was to create a ticket for each Features page, with a link to the existing Beta Development design page, so they could be used to track who was supposed to be writing what -- does that sound useful?13:41
<pivotal-bot>feature: Inventory Beta Design Documents for Documentation Draft (started) / owner: Esme Cowles
<mikeAtUVa>escowles: that sounds great13:47
escowles: we might need to be heavy-handed to get them all assigned and done, but I'm prepared for that.13:48
* mikeAtUVa lets his temporary title of scrum master go to his head...
<escowles>mikeAtUVa: yes, you'll wielding a whip before the week's out
<mikeAtUVa>escowles: this work (ticket creation and assignment) could either be done under your ticket, or mine (https://www.pivotaltracker.com/story/show/62181402)... should we flip for it, or do you have a preference.13:50
<pivotal-bot>feature: Create "Documentation > Features" tree in wiki. (started) / owner: Mike Durbin
<escowles>mikeAtUVa: either way is fine with me -- i just didn't want to create a bunch of tickets unless that were going to be useful13:51
i've done the rest of that ticket and I could work on roughing out some content for the developer's guide and administrator's guide instead13:53
* ajs6f joins
<mikeAtUVa>escowles: the other alternative is that you or I complete that work under the banner of another ticket... or that we group a few of them together.13:54
escowles: neither of which is great... I think we should just make the 14 tickets.13:56
<escowles>i agree: 14 tickets seems like the best approach, since they will probably be assigned to different people
* osmandin leaves13:57
* osmandin joins
* github-ff joins
[fcrepo4] ajs6f created pr/197 (+1 new commit): http://git.io/XvNcCA
fcrepo4/pr/197 38b731a ajs6f: Removing unnecessary inner classes in event filters
* github-ff leaves
* github-ff joins13:58
[fcrepo4] ajs6f opened pull request #199: Pr/197 with addition (master...pr/197) http://git.io/yemVPw
* github-ff leaves
<escowles>mikeAtUVa: so i'll go ahead and create the tickets, sound good?
* github-ff joins13:59
[fcrepo4] ajs6f closed pull request #199: Pr/197 with addition (master...pr/197) http://git.io/yemVPw
* github-ff leaves
* github-ff joins
[fcrepo4] ajs6f pushed 1 new commit to master: http://git.io/3sTUBQ
fcrepo4/master b3952fd ajs6f: Merge pull request #142 from futures/DeleteCXFTrash...
* github-ff leaves
<mikeAtUVa>escowles: Cool, thanks! Probably 1 or 2 point tasks?14:00
<pivotal-bot>A. "Horbulaco" Soroka started "Kill legacy (Atom) JMS format" https://www.pivotaltracker.com/story/show/62246234
<escowles>i'd say most of them are 2-pointers because the design pages aren't even close to user-ready
<mikeAtUVa>escowles: Agreed. It took me a while just finding out what documentation we had and making sense of it in some cases.14:02
<pivotal-bot>Mike Durbin added comment: "The Features page is present, a stub for each subpage exists and use of the excerpt tag is demonstratied. E..." https://www.pivotaltracker.com/story/show/6218140214:03
Mike Durbin estimated "Create "Documentation > Features" tree in wiki." as 1 point https://www.pivotaltracker.com/story/show/62181402
Esme Cowles added "Feature Documentation: Admin Search" https://www.pivotaltracker.com/story/show/6225688414:07
Esme Cowles added "Feature Documentation: Batch Operations" https://www.pivotaltracker.com/story/show/6225706614:09
Esme Cowles added "Feature Documentation: Clustering" https://www.pivotaltracker.com/story/show/6225715014:10
Esme Cowles added "Feature Documentation: Triplestore" https://www.pivotaltracker.com/story/show/6225725814:12
Esme Cowles added "Feature Documentation: Workspaces" https://www.pivotaltracker.com/story/show/6225735814:13
* github-ff joins
[fcrepo4] gregjan opened pull request #200: Authz fcrepo webapp (master...authz-fcrepo-webapp) http://git.io/TyEf_A
* github-ff leaves
<pivotal-bot>Esme Cowles added "Feature Documentation: Properties CRUD" https://www.pivotaltracker.com/story/show/6225745614:14
* travis-ci joins14:15
[travis-ci] futures/fcrepo4#1379 (pr/197 - 38b731a : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/commit/38b731aa3f62
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/15242651
* travis-ci leaves
<pivotal-bot>Gregory Jansen added comment: "https://github.com/futures/fcrepo4/pull/200 This PR adds the roles form to any node UI. I am still tryin..." https://www.pivotaltracker.com/story/show/61850482
Esme Cowles added "Feature Documentation: Authorization/Authentication" https://www.pivotaltracker.com/story/show/6225756614:16
Esme Cowles added "Feature Documentation: Backup & Restore" https://www.pivotaltracker.com/story/show/6225762614:17
<gregjansen>I need a velocity ninja. Trying to detect #if a $topic is of a particular mixin, namely fedora:resource..
* travis-ci joins
[travis-ci] futures/fcrepo4#1381 (master - 8ffbb36 : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/efce262e5cdf...8ffbb36f714f
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/15242740
* travis-ci leaves
<ajs6f>gregjansen: I can't help you, but do think that the phrase "velocity ninja" sounds kind of cool.
<pivotal-bot>Esme Cowles added "Feature Documentation: REST API" https://www.pivotaltracker.com/story/show/6225770014:18
<gregjansen>ajs6f: I enjoyed typing it
what I have: $rdf.find($nodeany, $topic, $helpers.asNode($rdfLexicon.HAS_MIXIN_TYPE), $helpers.asNode("http://fedora.info/definitions/v4/rest-api#resource"))
<ajs6f>gregjansen: I think cbeer took some pains to pack a set of helpers into the velocity context, but I don't know what they were/are called.
* travis-ci joins
[travis-ci] futures/fcrepo4#1382 (master - b3952fd : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/8ffbb36f714f...b3952fd26107
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/15242758
* travis-ci leaves
<pivotal-bot>Esme Cowles added "Feature Documentation: Federation" https://www.pivotaltracker.com/story/show/6225780614:19
<gregjansen>ajs6f, cbeer: found the helpers, cannot make it match. I suppose I will need to dump some triples and see what is up..
<mikeAtUVa>gregjansen: I found adding a method to the helper (and then unit test) for simple questions is clear and doesn't tax one's understanding of velocity... ie, I added "getVersionDate"14:21
gregjansen: then at least your debugging as you get it to work is java code and it takes velocity out of the mix
<gregjansen>mikeAtUva: nice, that's what I'll do.. but I just finished my velocity attempt, so let me try it..14:23
<pivotal-bot>Esme Cowles added "Feature Documentation: External Search" https://www.pivotaltracker.com/story/show/62258114
<mikeAtUVa>gregjansen: if it works that'll save you the taks of writing a unit test;)14:24
<pivotal-bot>Esme Cowles added "Feature Documentation: Namespaces" https://www.pivotaltracker.com/story/show/62258184
Esme Cowles added "Feature Documentation: Content CRUD" https://www.pivotaltracker.com/story/show/6225836614:25
<bljenkins>Project fcrepo-fixity-corrupter build #534: SUCCESS in 1 min 47 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/534/
<pivotal-bot>Esme Cowles added "Feature Documentation: Versioning" https://www.pivotaltracker.com/story/show/6225862814:28
<bljenkins>Project fcrepo-jms-indexer-pluggable build #336: UNSTABLE in 6 min 9 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/336/14:30
<pivotal-bot>Esme Cowles added comment: "I've created 14 tickets for the feature pages: several are basically done and just need review and excerpts,..." https://www.pivotaltracker.com/story/show/62234746
Gregory Jansen finished "Expose AuthZ Service/API on Object/Datastream views in F4 webapp" https://www.pivotaltracker.com/story/show/6185048214:34
Gregory Jansen added comment: "https://github.com/futures/fcrepo4/pull/200 access roles API being added here" https://www.pivotaltracker.com/story/show/6201990214:35
Gregory Jansen finished "Add access roles API to fcrepo-webapp" https://www.pivotaltracker.com/story/show/62019902
<gregjansen>thanks mikeAtUva, I ended up adding string literal support to view helpers for this14:41
soliciting opinions - the access roles form currently shows example json (it isn't converted to taking turtle/n3 yet) would it be preferable to do a xhr for the current json (if any)14:43
<ermadmix>random question- in fcr:namespaces, the namespace http://fedora.info/definitions/v4/repository# has a preferredNamespaceUri of http://www.jcp.org/jcr/1.0. Is that correct? All the other preferredNamespaceUri match the namespace14:50
<bljenkins>Project fcrepo-fixity-corrupter build #535: SUCCESS in 1 min 24 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/535/14:51
* github-ff joins
[fcrepo4] ajs6f created KillAtomJMS (+1 new commit): http://git.io/7GOVXA
fcrepo4/KillAtomJMS c4d51fa ajs6f: Killed off legacy (Atom) JMS format
* github-ff leaves
<ajs6f>ermadmix: It's probably getting mapped wrong somewhere. They are cerainly distinct.
<pivotal-bot>A. "Horbulaco" Soroka added comment: "https://github.com/futures/fcrepo4/pull/201" https://www.pivotaltracker.com/story/show/6224623414:52
* github-ff joins
[fcrepo4] ajs6f opened pull request #201: Killed off legacy (Atom) JMS format (master...KillAtomJMS) http://git.io/Hsyceg
* github-ff leaves
<pivotal-bot>A. "Horbulaco" Soroka finished "Kill legacy (Atom) JMS format" https://www.pivotaltracker.com/story/show/62246234
<bljenkins>Yippie, build fixed!14:56
Project fcrepo-jms-indexer-pluggable build #337: FIXED in 6 min 57 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/337/
* gregjansen leaves15:01
<ermadmix>ajs6f: ok, not going to track this down now, I'll create a ticket for it15:03
* ajs6f leaves15:05
* ajs6f joins15:07
<pivotal-bot>Eric James added "fcr:namespaces preferredNamespaceUri for http://fedora.info/definitions/v4/repository# is http://www.jcp.org/jcr/1.0 r..." https://www.pivotaltracker.com/story/show/62261778
Eric James edited "fcr:namespaces preferredNamespaceUri for http://fedora.info/definitions/v4/repository# is http://www.jcp.org/jcr/1.0 ..." https://www.pivotaltracker.com/story/show/6226177815:08
* gregjansen joins15:10
<bljenkins>Project fcrepo-fixity-corrupter build #536: SUCCESS in 1 min 24 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/536/15:14
Project fcrepo-kitchen-sink build #703: FAILURE in 55 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/703/15:15
Project fcrepo-jms-indexer-pluggable build #338: FAILURE in 3 min 1 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/338/15:16
<pivotal-bot>A. "Horbulaco" Soroka added "Unit tests for indexer framework" https://www.pivotaltracker.com/story/show/6226251215:17
A. "Horbulaco" Soroka added "More integration tests for the named-fields indexer gear" https://www.pivotaltracker.com/story/show/6226275815:20
A. "Horbulaco" Soroka accepted "Write tickets for improving test coverage on indexer gear." https://www.pivotaltracker.com/story/show/6215617415:21
A. "Horbulaco" Soroka started "Write tickets for improving test coverage on indexer gear." https://www.pivotaltracker.com/story/show/62156174
A. "Horbulaco" Soroka added comment: "https://www.pivotaltracker.com/story/show/6226251215:22
https://www.pivotaltracker.com/story/show/62262758" https://www.pivotaltracker.com/story/show/62156174
<ajs6f>mikeAtUva: You know, it occurs to me that what the confusion you expressed the other day ("Why aren't iterators of triples considered to have an 'order'?) might be a matter of conflation in the JDK itself. Scala, by contrast to Java, distinguishes sequences that have a defined order (the type Seq) from sequences that do not (the type Traversable). I can't find a Java semantic match to Scala's Traversable. Scala goes further and distinguishes "lazy" Travers15:27
* osmandin leaves15:28
<ajs6f>s (the type Stream)from "strict" Traversables. Java doesn't seem to do that in the language API, much an equivalent to Stream. Ah, well.
<mikeAtUVa>ajs6f: that does seem to capture the logical distinction between the two issues... most of the stock java.lang.Iterable implementations have consistent (even if undefined) order but the contract doesn't imply that behavior.15:30
<ajs6f>mikeAtUva: Agreed. But the very ambiguity is problematic.
* ajs6f leaves15:35
<barmintor>ajs6f: List
* ajs6f joins15:37
<barmintor>ajs6f: List
<ajs6f>barmintor: Isn't order at the very heart of the List abstraction?15:38
<barmintor>ajs6f: yes, its what distinguishes it from its supertype, Iterable
<ajs6f>barmintor: Perhaps. But a List has size(). What we're looking for is something that has no size and no order.15:39
A stream.
<barmintor>ajs6f: Iterable
or Enumeration15:40
<ajs6f>That's the ambiguity mikeAtUVa remarked. Technically, Iterators have no order. But even experienced Java devs like mikeAtUva don't consider that. It ought to be brought out and explicitly guaranteed.
barmintor: ENUMERATION!?! No way.
* barmintor shrugs
has no size or guarnateed order15:41
<ajs6f>Enumerations are specifically instances from a finite pool.
A finite set of instances.
A stream of RDF triples hardly meets that description.
<barmintor>"An object that implements the Enumeration interface generates a series of elements, one at a time."
<ajs6f>Well, there's another problem with Enumeration: http://docs.oracle.com/javase/7/docs/api/java/util/Enumeration.html#hasMoreElements()().15:42
<escowles>you know things have gone south when you start considering Enumeration...
<ajs6f>There's no just no good reason we aren't using Scala NOW!15:43
<barmintor>if experienced Java developers are trained to expect all Iterators to have predictable order, it's probably because the various Hash implementations didn't take advantage of the performance benefits their non-commitment to order offered until SDK 7 (afaik)15:46
but the predictability is *supposed* to be dependent on the underlying collection15:47
<ajs6f>barmintor: No argument. But I don't think mikeAtUVa is weird.
barmintor: Right, the dependence is problematic. I mean, it's fine, but there ought to be a way to say, "There is no dependence. In fact, there is no order!"15:48
Because sometimes there _isn't_ an underlying colelction.
<barmintor>right: Iterator makes no promise about order
<ajs6f>Right, but that's not made explicit. There's such a thing as lying by omission.:)15:49
I want a notice is RED CAPITAL LETTERS!
_And_ I want an Iterator that _does_ make (configurable) guarantees about order.
<mikeAtUVa>barmintor: that an iterator's order is deterministic is something I may occasionally lazily assume, although in this case my flawed lazy assumption was that if I put things on a stream in an order, they'd be rendered by that stream in that order later down the road.15:50
<barmintor>Well, one might suggest that, since Iterable<E> can be a target of foreach, that Iterator is not the issue
<ajs6f>mikeAtUVa: We should add the guarantee to RdfStream.
<barmintor>mikeAtUVa: Right, but only if it's a List or a Queue, right?15:51
<ajs6f>barmintor: Eh? For-each doesn't imply order, to my mind.
For-each just implies a complete traversal.
<barmintor>ajs6f: Exactly. It makes it clear that order is a characteristic of the collection by removing the explicit Iterator syntax
<ajs6f>No, a complete traversal most certainly does not occur in a defined order.15:52
<mikeAtUVa>ajs6f: eh, it may even be true in our case (I didn't dig down into why my test seemed to show things in a diferent order very infrequently) it doesn't make sense int he world of RDF
<escowles>what kind of order could we guarantee for the RDF stream? and is there any use case for needing or wanting that order?
<barmintor>ajs6f: It does if the collection has a defined order
<ajs6f>barmintor: BUT THERE IS NO COLLECTION!
<escowles>well, we could make FedoraResource extend Collection<Triple>...15:53
<barmintor>We seem to be in total agreement.
<ajs6f>barmintor And again, a black-letter reading of the spec supports you, but mikeAtUVa is a good example of a "reasonable person" who _didn't_ make the right assumption.
<mikeAtUVa>escowles: it'd be nice to display in the HTML the list of versions in chronological order.... we can do that by providing more triples... it's not a big deal, just a large theoretical discussion spawned from my own carelesness;)
<ajs6f>barmintor: We're in agreement that the spec doesn't make the guarantee. We're not in agreement about how people normally read the spec.
* mikeAtUVa does not want to be the poster child for "reasonble person".15:54
<escowles>mikeAtUVa: ok, i see -- versions could be listed in order if they were returned from the iterator in order, etc...
<ajs6f>escowles: We can't do that, because we don't have an underlying Collection and cant support the Collection methods.
The whole point of the RDFIteration thing was to avoid supporting the Collection contract.
<barmintor>ajs6f: I guess what I suggest is that the responsibility for documenting predictable order (or lack thereof) is on the Iterable impl (which I gather is FedoraResource?)
<ajs6f>(Which usually demands bringing the collection into heap.)
barmintor: I agree that we should make our lack-of-guarantee much more explicit.15:55
mikeAtUVa: We should add the guarantee to RdfStream."
I should have said the "non-guarantee".
<mikeAtUVa>ajs6f: that's good or I might be tempted to throw really important triples near the front of the iteration to improve performance;)15:57
<ajs6f>escowles/mikeAtUva: For a specific function (like ordering Versions) we can sophistimacate the RDF to support order. We don't have to do that by ordering triples.
<escowles>as long as we don't use rdf:List to order them, that sounds good to me15:58
<ajs6f>mikeAtUVa: We may very well need to be able to call out subgraphs as being of especial importance, but that's a matter for abstractions on _top_ of RdfStream.
(Or simply other methods that produce subgraphs as RdfStreams, like "the version stuff", "the property stuff", "the datastream stuff", what-have-ye.15:59
<ajs6f>escowles/mikeAtUVa: For ordering versions (which have a simple and canonical order), can't we use triples already present in the stream to order? Don't the versions themselves appear in triples that express their birtdays?
<barmintor>sophistimacate sophistimacate sophistimacate
<ajs6f>It's like sophistificatorating, but more sophisticated.
mikeAtUva: How about a trivial PR that adds a comment to the Javadocs on RdfStream explaining why people shouldn't do what you did?16:02
<mikeAtUVa>ajs6f: they do have dates in most cases.... and that's what I did for the unit tests, but they aren't always sortable and there's still some work to be done presenting just a meaningful set.16:04
ajs6f: er, integration tests... not unit tests...
<ajs6f>mikeAtUva: Okay. That sounds like work we want to do anyway (regularizing the serialization of dates, etc.)16:05
<mikeAtUVa>ajs6f: oh, no the dates are regular and can be simply sorted (using string comparisons) but not all *versions* are sortable...
<ajs6f>? There are different kinds of versions?16:06
<mikeAtUVa>ajs6f: in the current form, versioning a content node doesn't change the modification date of the datastream node so there are two versions with the same date...
<ajs6f>Urg. I don't feel like we have ever really completely understood the difference between a datastream node and the content node underneath it.16:07
<mikeAtUVa>ajs6f: there's still some work to do figuring out how to best expose subgraph versioning (which is what we have, though we only checkpoint at the "relevant" level)... I punted on that.
<ajs6f>So we version nodes (because JCR does) but also portions of the repo?16:08
<mikeAtUVa>Well, based on cnd, that version might include properties of that node (or not) and might include child nodes (or not).16:09
<ajs6f>I feel like we're driving a car with an engine that is too big for the suspension.
<mikeAtUVa>ajs6f: I half expect the acceptance tester for that feature to kick it back to me to at least ask for clarification if not simplification.
ajs6f: eh, it's iterative... our code and our understanding;)16:10
<ajs6f>mikeAtUva: Well, as long as it meets our local use cases.
I mean, the -community's- use cases.16:11
Of course.
<mikeAtUVa>ajs6f: yeah, of course.
* osmandin joins16:13
<ajs6f>afk bb16:20
* ajs6f leaves
* ajs6f joins16:25
ermadmix: Doing okay with that indexer PR?16:32
cbeer: Looks like Marmotta was trying to participate in the LDP group:
but maybe fell by the wayside?
Do you want to pursue that on our behalf (since you've done more than anyone else in that field)?
<cbeer>ajs6f: yep. i think they're one of the groups that's waiting for a final spec before doing anything
<ajs6f>I can help.
<cbeer>ajs6f: what do you want me to pursue?
<ajs6f>cbeer: Waiting for the final spec is a good way to be sure that it won't meet our needs.
cbeer: First task: figure out how to participate. :)16:34
Looks like there's a WG with regular meetings.
<cbeer>yep. and Stanford, at least, is a w3c member
<ajs6f>Nothing says GETTING THINGS DONE like a regular meeting.
<cbeer>i sent them some of our feedback on their discussion list
Seriously, we should leverage that W3C membership.16:35
I can't believe I just used the word "leverage".
<cbeer>yeah, but then I have to be on more conference calls :P
<ajs6f>&^(&^(& me.
That's why I promised to help.
I LOVE conference calls. No one knows if you're even wearing trousers on a conf call.16:36
<pivotal-bot>A. "Horbulaco" Soroka added comment: "Relative URIs scare me. They implicitly make guarantees about the resource from which they are rel..." https://www.pivotaltracker.com/story/show/6088381416:37
A. "Horbulaco" Soroka added comment: "https://dvcs.w3.org/hg/ldpwg/raw-file/default/ldp-bp/ldp-bp.html#include-a-trailing-slash-in-conta..." https://www.pivotaltracker.com/story/show/6088381416:38
A. "Horbulaco" Soroka added comment: "Can we include the trailing slash in _published_ URIs but still dereference without them?" https://www.pivotaltracker.com/story/show/6088381416:39
* gregjansen leaves16:40
<ajs6f>cbeer: This reminds me that we once talked about trying to represent ourselves to the JCR group as well. Did we ever think of a good avenue for that?
<cbeer>ajs6f: nope. because why do it ourselves when we can make rhauch do it for us? :P16:41
<ajs6f>cbeer: I get that, but we don't _always_ agree with rhauch.
And if we said the same thing as rhauch in a convo, that helps rhauch _and_ us.16:42
Strength in illusory numbers.
<cbeer>man, it would be nice if our tests reliably passed16:46
<ajs6f>PATCHES… oh, that joke just doesn't have the same savor weeks before a release.16:47
<cbeer>i'm getting a pretty consistent JMS error, though
<ajs6f>Is it the Atom serialization error?16:48
Oh, wow.
Did y'all just see the latest message on the email list?
Someone wants to do _exactly_ what I worried someone would want to do.
Hang an F4 repo over a set of node primary types that _aren't_ the ones we're using.16:49
And _that's_ why we stick to mixins, ladies and jellybeans.
I have no idea whether this person's idea makes any sense at all.
But that won't stop me from being glad we went down the mixin road.
Thank you, "Stenfano", for validating my weird obsession with elegant type systems.16:50
* barmintor sprains an ocular nerve
<cbeer>escowles: do you know if mcritchlow and ajs6f ever talked about stanbol? I lost an email from Matt right after DLF, and I wonder if I should still reply16:51
<ajs6f>barmintor: It's okay, just repeat after me: ajs6f was right. Crazy, annoying, even _wrong_, but still right.
cbeer: mcritchlow?
<escowles>cbeer: i haven't heard anything about it16:52
ajs6f: my boss
<ajs6f>escowles: I thought you work for Declan?>
<escowles>ajs6f: well, the buffer between me and declan...
<cbeer>escowles: ok. i'll dredge up that thread. and, ajs6f, here's your chance to convince UC that stanbol is the way forward
<ajs6f>Fair enough. Declan is pretty intense.
<escowles>ajs6f: i reported directly to declan for a year or so when my previous supervisor had.... issues -- it was fun, but occasionally pretty intense16:53
<ajs6f>Stanbol is nothing more nor less than a really useful set of tools for people who have aggregations of metadata and want to participate in the SemWeb.
escowles: Bet you drank some pretty amazing beer.16:54
<escowles>ajs6f: that does tend to happen -- though i think i probably only saw him once that year (i was in the netherlands for half of it)
<cbeer>ajs6f: beats me what the context is.. i suspect it comes out of our hydra/rdf meeting where i tried to convince people we were solving problems Stanbol has already tackled
<ajs6f>escowles:Oh, you poor bor.
<escowles>ajs6f: i know, it was terrible -- westmalle by the case as the local grocery store, etc.16:55
* barmintor thinks to himself: "Po'boys are sandwiches"
<ajs6f>I have never seen anything that could convince Rubyists that anyone has ever solved a problem before in a way that they could just reuse.
barmintor: You must thinking of either hoagies, grinders, or subs.16:56
<barmintor>ajs6f: Not if they have oysters, or catfish
<ajs6f>cbeer: I'll take up the conv on emai.
* ksclarke leaves
<ajs6f>barmintor: true that.
We do see catfish "sandwiches" around here, but I think that's because of the general confusion about our culinary traditions between the mountains and the sea.16:57
<escowles>i didn't know that po'boys went that far north16:58
<ajs6f>escowles: Welcome to the age of electronic communication, when your dialect has less to do with the place of your birth and more to do with the consistency of your media diet.16:59
* kaarefc leaves17:00
<cbeer>grr. all i want is a successful build.
<ajs6f>cbeer; If it makes you any happier (and the JMS errors are from the Atom problems), I have a PR to kill that stuff (pursuant to this morning's standup conversation).17:02
I'm out for the day. See y'all later.17:04
* ajs6f leaves
* osmandin leaves
* tecoripa leaves17:05
* tecoripa joins17:08
* tecoripa leaves17:09
* github-ff joins17:11
[fcrepo4] cbeer created fixity-uris (+2 new commits): http://git.io/qCRDGw
fcrepo4/fixity-uris 7528f0b Chris Beer: update fixity responses to use PREMIS vocabularies and hash-uri subjects
fcrepo4/fixity-uris fa47399 Chris Beer: Use info:fedora/ as the default resource namespace
* github-ff leaves
<barmintor>plus ca change17:12
<pivotal-bot>Chris Beer added comment: "https://github.com/futures/fcrepo4/pull/202" https://www.pivotaltracker.com/story/show/62063292
Chris Beer finished "Update produced fixity result URIs" https://www.pivotaltracker.com/story/show/6206329217:13
* github-ff joins
[fcrepo4] cbeer opened pull request #202: Fixity uris (master...fixity-uris) http://git.io/lXv--g
* github-ff leaves
* github-ff joins17:22
[fcrepo4] cbeer created namespace-jcr-fix (+1 new commit): http://git.io/IF9ZfQ
fcrepo4/namespace-jcr-fix fbdf974 Chris Beer: Use the translated RDF namespace URIs when adding a namespace's preferred URI triples
* github-ff leaves
<pivotal-bot>Chris Beer added comment: "https://github.com/futures/fcrepo4/commit/fbdf974373bb" https://www.pivotaltracker.com/story/show/62261778
Chris Beer added comment: "https://github.com/futures/fcrepo4/pull/203" https://www.pivotaltracker.com/story/show/6226177817:23
* github-ff joins
[fcrepo4] cbeer opened pull request #203: Use the translated RDF namespace URIs when adding a namespace's preferre... (master...namespace-jcr-fix) http://git.io/JnRDSA
* github-ff leaves
<pivotal-bot>Chris Beer estimated "Update produced fixity result URIs" as 2 points https://www.pivotaltracker.com/story/show/62063292
Chris Beer finished "fcr:namespaces preferredNamespaceUri for http://fedora.info/definitions/v4/repository# is http://www.jcp.org/jcr/1...." https://www.pivotaltracker.com/story/show/62261778
Chris Beer estimated "Add PREMIS vocabulary to fixity results reporting" as 1 point https://www.pivotaltracker.com/story/show/6224356217:24
Chris Beer started "Add PREMIS vocabulary to fixity results reporting" https://www.pivotaltracker.com/story/show/62243562
Chris Beer finished "Add PREMIS vocabulary to fixity results reporting" https://www.pivotaltracker.com/story/show/62243562
Chris Beer added comment: "Part of https://github.com/futures/fcrepo4/pull/202" https://www.pivotaltracker.com/story/show/62243562
Chris Beer edited "Add PREMIS vocabulary to fixity results reporting" https://www.pivotaltracker.com/story/show/62243562
Chris Beer deleted "MODE-2098" https://www.pivotaltracker.com/story/show/6105106017:26
* travis-ci joins17:27
[travis-ci] futures/fcrepo4#1385 (fixity-uris - fa47399 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/7528f0be8677^...fa473992f45f
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/15252882
* travis-ci leaves
* travis-ci joins17:35
[travis-ci] futures/fcrepo4#1387 (namespace-jcr-fix - fbdf974 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/commit/fbdf974373bb
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/15253335
* travis-ci leaves
<bljenkins>Project fcrepo4 build #1518: UNSTABLE in 27 min: http://ci.fcrepo.org/jenkins/job/fcrepo4/1518/17:38
Yippie, build fixed!18:10
Project fcrepo4 build #1519: FIXED in 31 min: http://ci.fcrepo.org/jenkins/job/fcrepo4/1519/
Project fcrepo-fixity-corrupter build #537: SUCCESS in 1 min 47 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/537/18:12
Yippie, build fixed!18:16
Project fcrepo-kitchen-sink build #704: FIXED in 4 min 30 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/704/
Yippie, build fixed!18:17
Project fcrepo-jms-indexer-pluggable build #339: FIXED in 7 min 31 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/339/
* ermadmix leaves18:19
* ksclarke joins21:30
* github-ff joins22:02
[fcrepo4] ajs6f pushed 1 new commit to master: http://git.io/CQ9erA
fcrepo4/master 169a3e4 ajs6f: Merge pull request #203 from futures/namespace-jcr-fix...
* github-ff leaves
* travis-ci joins22:13
[travis-ci] futures/fcrepo4#1389 (master - 169a3e4 : ajs6f): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/b3952fd26107...169a3e4fd77d
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/15263770
* travis-ci leaves
<bljenkins>Project fcrepo4 build #1520: UNSTABLE in 1 hr 6 min: http://ci.fcrepo.org/jenkins/job/fcrepo4/1520/23:08
Chris Beer: Use the translated RDF namespace URIs when adding a namespace's preferred URI triples

Generated by Sualtam