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

Using timezone: Eastern Standard Time
* ajs6f joins07:05
* dhlamb joins08:39
* awead joins08:49
* whikloj joins08:59
* acoburn joins10:11
* esm_ joins10:13
<f4jenkins>Project fcrepo4-T2 build #324: UNSTABLE in 5 min 23 sec: http://jenkins.fcrepo.org/job/fcrepo4-T2/324/10:53
* jgpawletko joins10:55
* dhlamb leaves11:04
<awoods>acoburn/ajs6f/escowles/ruebot/whikloj/all: Before creating JIRA tickets and Epics, I put together a skeleton of technical priorities to help focus development. Please comment/update as you see fit:11:10
Once we have a common vision, we can port it over to JIRA.
<ajs6f>awood: Can you unpack the phrase "Fedora as API" for me?11:11
<ruebot>awoods: are the top level items epics, and the lowest sprints?11:12
* github-ff joins11:13
[migration-utils] ruebot opened pull request #36: Add fcrepo as parent, update pom, and add license statement to all fi… (master...FCREPO-1654) http://git.io/v3Den
* github-ff leaves
<ruebot><3 bash scripts11:14
<awoods>ajs6f: unpacking: separate API from impl, TCKs, ontologies11:19
<ajs6f>awoods: Cool.11:20
<awoods>ruebot: The top level items could possibly be epics. The lower level items are just more broken out... potentially to the sprint level.11:21
acoburn/ajs6f/escowles/ruebot/whikloj/all: Also up for review: https://wiki.duraspace.org/display/FF/F4+GitHub+Organizations#F4GitHubOrganizations-RequirementsforEntrytofcrepo4-exts11:31
<ajs6f>awoods: Why the package name requirement?11:32
awoods: I think we need at least two owners.11:33
<awoods>ajs6f: there was some discussion on IRC yesterday afternoon about the package name... please give it a look and see if you feel differently.
<acoburn>ajs6f: the package name requirement had to do with making sure that the fedora project had control over the namespace in sonatype
<ajs6f>acoburn: Why would we care?11:34
<acoburn>ajs6f: if the project used some other namespace, and the owner left, but other projects depended on it, it might get messy
<ajs6f>acoburn: That's not my responsibility.11:35
<acoburn>ajs6f: it's not unrelated to needing at least two owners
<ajs6f>acoburn: The two owners thing is about the beginning. The namespace thing is about the ending.
<awoods>ajs6f: see http://irclogs.fcrepo.org/2015-08-13.html starting at 13:11
<ajs6f>"<whikloj> acoburn: and I think that gets to the heart of the matter. If these are extensions then they are not "required" for Fedora and if the maintainer drops them, it would be unfortunate but not the end of the world."11:37
"<acoburn> awoods: I think anything in -exts implies a certain level of commitment on the part of the fedora team
awoods: and anything in -exts also implies that we have control over releases"
What's in -ext is stuff for which the commiters do _NOT_ have committment.11:38
And for which we will _NOT_ be doing releases. Or else, what is the point of it at all?
<acoburn>ajs6f: yes, but someone needs to have commitment for those projects
<ajs6f>acoburn: yes, the people who put them in fcrepo-ext. If they drop the ball, the ball is dropped.11:39
<acoburn>ajs6f: I don't disagree
<ajs6f>acoburn: I'm sure as hell not picking it up.
<acoburn>ajs6f: that's what -archive is for, I believe11:40
<ajs6f>acoburn: So no problem. If the team responsbile for project X in -ext dies, we move the code to -archive and the matter is finished.
acoburn: No need to control namespaces.
<awoods>ajs6f: if the owners drop the ball... and someone else wants to pick it up... then what?11:41
<ajs6f>awoods: Fork.
<awoods>ajs6f: they need to change namespaces?
ajs6f: and impact all installations?
<ajs6f>awoods: Yep. That's the very meaning of a namespace.
<awoods>ajs6f: org.fcrepo is a namespace11:42
<ajs6f>awoods: yep. That's the nature of community-driven software. You choose to trust the groups that seem to meet your level of responsibility and care.
awoods; org.fcrepo is _my_ namespace. I don't want you sticking crap in it for which I didn't agree to take responsbility. I barely have time to deal with what's in there now, and you want to open the door to more?11:43
awoods: Ask barmintor how that ends.
<awoods>ajs6f: on the flipside, would you suggest no projects in fcrepo4-exts be in the org.fcrepo namespace?11:44
<ajs6f>awoods: Not unless the commiters take explicit and permanent reponsibility for them.
awoods: On a case-by-case basis.
<awoods>ajs6f: what is the criteria? two or more committers taking ownership?11:45
<ajs6f>awoods: the usual (+1/0/-1) voting with −1 = veto.
awoods: Two committers can own a project without all of the committers owing it.11:46
awoods: Committers do things without their committer hats on.
awoods: See asger's work on ECM for a great example.
<ruebot>what if we had componement managers/owners? that's how we handle it in islandora with our 40+ modules.11:47
<awoods>ruebot: yes, that is a requirement in the proposal.
<ajs6f>ruebot++ # this is fundamental to my argument. More people than committers / other people than committers need to take resposnsbilities.11:48
<ruebot>awoods: cool.
<acoburn>this is, quite possibly, a great way to get more people involved in fedora without having to dive into the core of it
<awoods>acoburn/ajs6f: Is it safe to assume that projects in fcrepo4-exts need to be released to maven central?
<ajs6f>Yes. "Scratch your own itch."
Instead of a selection itches listed on the strategic itch planning chart in Jira.
<awoods>acoburn/ajs6f: Is that a requirement?
<ruebot>...and if a given component does not have a manager, it is not apart of a release. we had to do that for our last release with a couple components.
<ajs6f>awoods: No. They may not even be JVm-resident gear.
<awoods>ajs6f: if they are JVm-resident gear, is it a requirement?11:50
<ajs6f>awoods: Possibly. I'd have to think about it. It's not obviously necessary, although it is obviously convenient.
<awoods>acoburn/ruebot: thoughts?11:51
<acoburn>awoods: about maven-central?
<awoods>acoburn: yes
<ajs6f>acoburn: I'm interested in any thoughts you may have about this extremely long Stack Overflow discussion:11:52
<acoburn>awoods: isn't publishing to maven-central synonymous with publishing a release
<ruebot>awoods: if i'm understanding this correctly, -ext would be a decentralized release, so a "component manager" would be responsible for maven-central/sonatype?
<awoods>ruebot: correct
<ajs6f>acoburn: A release can involve a lot more than artifacts-> MC.11:53
acoburn: It could entail docs management, demo site updating, community messaging, etc.
<awoods>acoburn: and it is also possible to publish maven artifacts to non-standard, but publicly accessible repositories.
<ruebot>awoods: then, those releases would correspond the matrix of versions that work together we'd still need to create11:54
<awoods>ruebot: presumably
<acoburn>awoods: it seems that the publishing decision should be in the hands of the owners
<ajs6f>ruebot: Create all together as a community effort.
<whikloj>Sorry I got in here late and had to catch up11:55
<ruebot>whikloj: don't worry. i'm still your voice. you're signed up to be a component manager of everything! :-P11:56
<ajs6f>whikloj: We're discussing whether to make you release manager for every Fedora project forever.
<awoods>acoburn: true, but it may be hard to have a community java module that is not accessible via maven... in the Fedora context.
<whikloj>ruebot: are you the voice of ajs6f too now?!
* ruebot gasps
* ajs6f coughs, shifts latex ruebot mask over his face nervously11:57
<ajs6f>We're all just facets of Fedo.
<whikloj>awoods: Is there any reason a module in -exts has to have the parent pom from fcrepo-parent? Technically I mean not philosophically.11:58
awoods: ^^From a technical perspective11:59
me take pretty one day
me talk pretty one day
<acoburn>whikloj: fcrepo-camel doesn't use the fcrepo parent pom12:00
<awoods>whikloj: no, but this is all swirling around module use and release
<whikloj>acoburn: exactly
<ajs6f>whikloj: many people will find it convenient, but I would expect others not even to use Maven.
<ruebot>...migration-utils was just updated to use the parent.12:01
<acoburn>ajs6f: just wait till someone starts using leiningen
<ajs6f>acoburn: I've stopped setting my hair on fire. have you?
<awoods>We want to provide some structure for consistency and guidance, but not so much that it is constraining.12:02
<whikloj>I think this all comes back to a difference between fcrepo4-exts as a repo and fcrepo4-exts as an extension of fcrepo4.
<ajs6f>ruebot:whikloj: we can't stop anyone who wants to from using the parent. It's a published artifact.
<whikloj>Extensions to fcrepo4 should have really no requirements placed on them.
<whikloj>awoods: you can
<ajs6f>whikloj: They don't. The requirements come in when you want to use Fedora community property (our Github accounts) to do your work.
<whikloj>awoods: oops, you can't control what someone somewhere writes and pubishes.
<ajs6f>whikloj: That is a smaller and more carefully defined group of extensions.12:04
<awoods>whikloj: correct, but you can control what ends up in fcrepo4-exts
<ajs6f>I would say "curated" but that word just seems kind of greasy nowadays from overuse.
<whikloj>ajs6f: I think that is the perfect word
fcrepo4-exts is a curated repo of extensions to fcrepo412:05
<ajs6f>whikloj: I'd like to curate some beer more than I'd like to curate some code.
<whikloj>so fcrepo4-exts has additional requirements
<awoods>whikloj: exactly. And we are trying to define the criteria for that curation.
<whikloj>I think points 1-4 are fine12:06
<ajs6f>If we're going to curate stuff, do we get to drink fancy red wine while we do it?
<whikloj>5, I go back and forth on
and 6 I think should be optional
<ruebot>ajs6f: i'll try and get diego to send some more wonderful chilean wine if so :-)12:07
<whikloj>ajs6f: https://www.youtube.com/watch?v=1GW22sAElpE
<ajs6f>whikloj: Ooh, that's a classic. That takes me back.
<acoburn>awoods: perhaps 1-4 are required and 5-6 are recommended?
<ajs6f>Man, Dave Foley and Kevin MacDonald look so young...12:08
<awoods>acoburn/whikloj: that is a possibility
<whikloj>awoods/acoburn: recommended but optional sounds good
<ajs6f>We could issue badges. Do 1-4 and you can come in the clubhouse, but do 5-6 and you get a sticker on your github repo.
"Fedo approved!"12:09
<whikloj>ajs6f: "another tahitian teehee for my friend, and waitress, this time no giggling"
<ajs6f>We could issue badges for bad habits too. "Shitty test coverage!" "Doesn't release regularly!" "Met them at a conference and they smell weird."
* ruebot snorts12:10
<awoods>acoburn/whikloj/ajs6f: although I think we need to be more clear on #5, and on whether artifacts need to be in maven central... or a public maven repo.
<ajs6f>whikloj: You and Youtube just destroyed any chance I had to get work done this afternoon.
<whikloj>ajs6f: me too, sadly
<ajs6f>awoods: There's a chance for non-coders to contribute. Design offensive and insulting badges for Fedora to hand out.12:12
<acoburn>awoods: wouldn't the availability of a project release (whether in maven-central or elsewhere) be related to item #1
<whikloj>acoburn: how do you mean ^^12:14
<awoods>acoburn: quite possibly, but not necessarily
<acoburn>whikloj: I mean if it's java, and it's not in some public repository, it's hard to add to your project and therefore hard to use
whikloj: I mean outside of, say, your own institution12:15
<ajs6f>acoburn: If the code is out there, it's possible. Inconveient, but possible.12:16
<whikloj>acoburn: ahhh right, then I think you are right. Unless they hand them out on USB-sticks at a conference
acoburn: ajs6f would design the labelling
<acoburn>whikloj: hopefully with a tahitian teehee on it
<ajs6f>"Special secret underground Fedora component SHH don't tell anyone!"
awoods: you want us to put our official vote in the wiki, or can I say "yes" to #1-4 and a provisional "yes" to 5 & 6, if they are optional but recommended.12:18
<awoods>whikloj: comments in the wiki are good. #5 is related to releasing, and therefore needs to be yes or no. I am not sure "optional" makes sense.12:22
whikloj/ajs6f/ruebot/acoburn: I also a requirement/recommendation needs to be added related to maven central.12:23
<ajs6f>awoods: At most a recommendation.
* ruebot commented
<awoods>ajs6f: for java projects, you would be ok if one found its way into fcrepo4-exts but was not released to maven central?12:24
<ajs6f>awoods: Yep.
awoods: I don't think it's likely.
<awoods>ajs6f: but allowable?12:25
<ajs6f>awoods: Sure.
awoods: An example would be scale testing gear.12:31
<awoods>ajs6f: or migration-utils
<ajs6f>awoods: Yeah, excellent example.12:32
<whikloj>awoods/ajs6f: I think you both have made the case for maven-central as a recommendation and not a requirement.
<ajs6f>We could set up our own repo. "Fedora Fleabag Fly-by-Night JAR Motel"12:34
<whikloj>ajs6f: I think I saw that movie
<ajs6f>whikloj: It featured Bruce Campbell, I believe.12:35
<acoburn>ajs6f: it's also featured here: http://bedbugregistry.com/
* esm_ leaves13:20
<ajs6f>I'm in the call.14:00
<acoburn>I'm also here
* bseeger joins14:01
<ajs6f>awoods: I'm here.14:03
It's logged, too.
* apb18 joins14:04
* iamtimmo joins14:07
* Dan_Davis joins14:08
<ajs6f>Sure, why not?14:09
currently interested in collecting use cases14:11
these services will be external, i.e. not in fedora core14:16
<ajs6f>That's because this is middleare ,not user-facing software.14:24
Use cases are for developers.
(These use cases, that is.)
<awoods>We can either add use cases on the wiki (with a given tag 'uc-api-ext') or in the google doc: https://wiki.duraspace.org/display/FF/Use+Cases14:27
<ajs6f>Just as longas it's clear that these are not use cases for the core itself.14:30
<ajs6f>Sure, why not?
should API Extension use cases be child pages of ^^ ?14:33
<ajs6f>awoods: +1
<acoburn>awoods: +1
<ajs6f>"You gotta keep 'em separated."
<awoods>Here is the parent use case page: https://wiki.duraspace.org/display/FF/Use+Cases+-+API+Extension+Architecture14:35
We conferenced the HELL out o that call.15:01
<awoods>ajs6f: indeed15:02
* esm_ joins15:13
<ajs6f>Oh, so many good KiTH sketches. Darn you, whikloj.15:50
<whikloj>ajs6f: "What did Danny tell me yesterday?"
<ruebot>awoods: give me a sec. i didn't try building before i updated the pom.xml -- i gotta resolve a couple dependencie conflicts for 1654.15:51
<awoods>ruebot: no rush
<whikloj>ajs6f: https://youtu.be/R15LkyO8gq8?t=1m49s
<ajs6f>whikloj: Oh, Danny. Weird Everyman superhero. And so many hidden talents: https://www.youtube.com/watch?v=cN8heAodh-k15:52
<whikloj>ajs6f: hahahaha, "its not a flag, let it touch the ground"
<ajs6f>whikloj: That is my wife's favorite KiTH sketch of all. "Always put salt in your eyes?"15:53
<whikloj>ajs6f: I love Scott Thompson, he is hilarious
<ajs6f>whikloj: You know he did a Danny Husk comic book?15:54
<whikloj>ajs6f: no way
<ajs6f>whikloj: http://www.amazon.com/Husk-The-Hollow-Planet-Danny/dp/1600108210
whikloj: Worth a gander.15:55
<ruebot>whikloj, ajs6f: Scott Thompson was on this week's Canadaland Show - http://canadalandshow.com/podcast/scott-thompson15:56
<ajs6f>ruebot: Thanks. That will certainly help me get back to getting work done.
<ruebot>awoods: ok. now it builds fine.15:57
ajs6f++ #not a problem :-P
<whikloj>ruebot/ajs6f: awesome!
* ruebot disappears15:58
* github-ff joins16:03
[migration-utils] awoods pushed 1 new commit to master: http://git.io/v3y5W
migration-utils/master 59470e4 nruest: Add fcrepo as parent...
* github-ff leaves
<whikloj>ruebot: sidebar - I love Jesse Brown as well16:04
* github-ff joins
[migration-utils] awoods closed pull request #36: Add fcrepo as parent, update pom, and add license statement to all fi… (master...FCREPO-1654) http://git.io/v3Den
* github-ff leaves
* esm_ leaves
* apb18 leaves16:05
<ajs6f>awoods: I feel like this perfectly represents our relationship: https://www.youtube.com/watch?v=uws_lwD3bQc16:07
awoods: )The first sketch, that is.)
* travis-ci joins16:09
fcrepo4-exts/migration-utils#89 (master - 59470e4 : nruest): The build passed.
Change view : https://github.com/fcrepo4-exts/migration-utils/compare/f419394be36f...59470e455f25
Build details : https://travis-ci.org/fcrepo4-exts/migration-utils/builds/75659403
* travis-ci leaves
* jgpawletko_away leaves16:10
* bseeger leaves16:33
* jgpawletko joins16:36
* ajs6f leaves16:38
* Dan_Davis leaves16:52
* whikloj leaves17:13
* jgpawletko leaves17:24
* jgpawletko joins17:25
* jgpawletko leaves17:39
* ksclarke joins17:44
* acoburn leaves18:16
* acoburn joins18:17
* the_mgt_ joins19:38
* the_mgt leaves19:43
* acoburn leaves19:58
* dwilcox joins21:16
* dwilcox leaves21:41