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

Using timezone: Eastern Standard Time
* ksclarke leaves00:52
* jonathangee leaves06:59
* jonathangee joins
* edInCo joins07:23
<pivotal-bot>A. "Cartroef" Soroka added comment: "https://wiki.duraspace.org/display/FF/The+Fedora+4+object+model" https://www.pivotaltracker.com/story/show/5157176708:41
* ajs6f joins08:42
cbeer: Did you already do this?
https://www.pivotaltracker.com/story/show/51426639
<pivotal-bot>feature: Parameterize Dataset HTML response (unstarted) / owner: A. "Cartroef" Soroka
A. "Cartroef" Soroka added comment: "We went with Apache Oltu instead." https://www.pivotaltracker.com/story/show/4545824908:43
A. "Cartroef" Soroka edited "Add an OAuth2 service to the REST API's JAX-RS container using Spring Security OAuth2" https://www.pivotaltracker.com/story/show/45458249
* ajs6f leaves08:53
* ksclarke joins09:23
* ajs6f joins11:06
<pivotal-bot>A. "Cartroef" Soroka added comment: "https://issues.jboss.org/browse/MODE-1995" https://www.pivotaltracker.com/story/show/5347902911:34
* ksclarke leaves11:43
* ksclarke joins11:59
* edInCo leaves12:01
* nbanks joins12:09
* mjg_ joins12:22
I scoured the wiki looking for an explanation but to no avail: can someone please explain what a "managed external datastream" is? is that just a managed external datastream in a customized location?12:24
<cbeer>eddies? ^12:26
<ajs6f>aawoods? ^^
<mjg_>er, I meant to type: is that just a managed datastream in a customized location? an object that sits outside the Fedora-configured objectStore but can have audits, versioning, checksums, etc. handled by Fedora?12:28
<cbeer>ask ff-tech? the steering group deserves a hard time for making a list of features without a scope, use case or narrative to go with them12:30
<mjg_>I am happy to be the designated hard time-giver.12:31
does the steering group monitor ff-tech?
<cbeer>if they don't, they should. afaik, it's the only public-facing mailing list we have right now
<aawoods>The term "managed external datastream" refers the content that is managed by fedora as a federated/projected datastore.
<cbeer>aawoods: in non-fcrepo terms, does that mean what mjg_ thought it did?12:33
<mjg_>cbeer: done.12:34
I'm afraid I'm not sure what a projected datastore is.12:35
<ajs6f>mjg: Remember that jazz about aiming a repo over non-managed storage and still being able to use the API to access/mutate that content?12:36
<mjg_>I probably would know if I spent more time at the Chicago dev house talking about fcr4 and less about beer.
<ajs6f>eddies liked to bill it as "instant ingest"...
Fedora 4: Don't you wish you could remember?
<mjg_>ajs6f: you mean, "BLACK MAGICK?"
yes, OK, this rings a bell.12:37
<cbeer>ajs6f: is that really the same thing? i figured instant ingest meant it eventually made it into the repo, but was immediately available.
<ajs6f>mjg: was that a beer? eddies and barmintor and cbeer drink some pretty obscure stuff.
<cbeer>and this meant.. it'd stay where you put it, i guess
<mjg_>managed datastreams within non-managed storage, then? that seems like what I was getting at. ish.
<ajs6f>cbeer: "instant ingest" = projection + workflow that we haven't built yet.
<mjg_>ajs6f++12:38
<aawoods>cbeer: The answer for now is limited to what we get from modeshape (and you may be in the best position to answer that). As for what the term will mean, I would like to say that it will include audit, versioning, checksums, etc... but there is obviously work to be done in defining that list.
<ajs6f>Thanks, mjg, but save the karma for when we actually build the workflow. :)
<mjg_>I'm an architect; I give karma for ideas and talk, not for "building!"
<ajs6f>Putting my architect's hat on, I'll be happy to accept it.12:39
<cbeer>aawoods: ok. i've put you on-record as saying that.12:41
<mjg_>aawoods: so if I am comfy having everything managed by fcr4 (fc4? fcrepo4? ff? what are the cool kids calling it? I need short forms), then I don't care about managed external datastreams?
<aawoods>mjg_: Arguably the content is managed by fedora in both cases. I think it comes down to how you want to get your content into fedora (ingest via API or via projection)... and it also depends on what features (audit/versioning/etc) fedora will be able to provide for projected content.12:44
<ajs6f>mjg: Projection also allows for considerable transparency into the persistence layer. By fulfilling the projection contract, you can bring an almost arbitrary kind of storage under some portion of fcrepo's available services.12:47
Some folks have been concerned to maintain (in some form) the traditional Fedora promise that persistence is "human-readable" (and therefore durable) and this is one of the ways we're creating that possibility.12:48
<pivotal-bot>Chris Beer added "MODE-1997" https://www.pivotaltracker.com/story/show/5389466712:51
Chris Beer started "MODE-1997" https://www.pivotaltracker.com/story/show/53894667
<mjg_>I'm having a hard time imagining a use case where you'd want the cost of Fedora without benefits like audit, versioning, and checksums.12:53
<cbeer>(aawoods: in mjg_'s defense, I just created the external managed datastream page after you defined what it meant in here.)
<ajs6f>The case in point is where you already have Fedora in play, and projection allows to reuse services you've already created.
* github-ff joins12:55
[fcrepo4] yqjiang opened pull request #103: Add stomp to jms (master...master) http://git.io/vdvCKw
* github-ff leaves
<mjg_>OK, thanks all. helpful.13:02
<cbeer>ajs6f: i've been struggling through the LDP doc some more13:11
i think we should change the sparql-update endpoint into a PATCH method
<ajs6f>cbeer: Is spec'd by LDP? I'm find with that.13:12
<cbeer>and reserve POST/PUT for providing the whole document
they're wishy-washy about PATCH.
<ajs6f>That does seem consonant with the HTTP semantics.
<cbeer>4.7.1 LDPR servers may implement HTTP PATCH to allow modifications, especially partial replacement, of their resources [RFC5789]. No minimal set of patch document formats is mandated by this document.
<ajs6f>Unless we want to get into defining objects as LDP containers.
Which I think we will have to at least talk through.
<cbeer>yeah. i'm hoping they get around to editing the information about containers before then13:15
<ajs6f>I haven't read it carefully—do they not do a good job of talking through the HTTP side of things?13:36
* mjg_ leaves13:48
* github-ff joins14:22
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/WXRZfg
fcrepo4/master 4326cea Chris Beer: remove unnecessary methods from JcrRdfTools
* github-ff leaves
* eddies leaves14:26