Time
S
Nick
Message
02:52 jri joined #dataverse
04:38 nanozz joined #dataverse
06:52 jri joined #dataverse
07:59 jri joined #dataverse
08:16 jri joined #dataverse
08:43 poikilotherm joined #dataverse
08:46 juancorr joined #dataverse
09:57 nanozz left #dataverse
11:42 pdurbin joined #dataverse
11:54 jri joined #dataverse
12:04 jri_ joined #dataverse
12:34 poikilotherm
Morning pdurbin
12:34 poikilotherm
Welcome to another week of madness ;-)
12:36 poikilotherm
Thx for fixing #5349 :-)
12:37 poikilotherm
Lets hope that MrK has some time in spare to tackle this. Just left a review on it, too.
12:45 poikilotherm
pdurbin I remembered what I dislike about IRC : when you loose your connection, it is a burden to readup what happened in between. This hit me a few times now when on WLAN or mobile.
12:52 MrK joined #dataverse
12:58 jri joined #dataverse
13:01 jri joined #dataverse
13:05 jri_ joined #dataverse
14:08 pdurbin
poikilotherm: yes, but iqlogbot helps us read up by logging messages, of course. IRCv3 will have a "chathistory" extension: https://github.com/ircv3/ircv3-specifications/pull/349 . Sorry for the inconvenience.
14:08 donsizemore joined #dataverse
14:08 pdurbin
MrK: I hope my changes to your flyway branch look ok to you.
14:13 MrK
Ah yeah you just merged with develop.
14:13 pdurbin
MrK: yes, and I also make the diff smaller. Fewer whitespace changes. Long lines still long. :)
14:14 MrK
Ah yeah I don't have problem with that, it's Intellij formating during push.
14:17 pdurbin
MrK: if you could avoid reformatting entire files, it would be appreciated. Thanks again for the pull request! I'm hoping it'll be discussed after standup today or tomorrow.
14:18 MrK
Yeah I will uncheck the option, poikilotherm added some comments which I will adress after I'm done with my tasks.
14:18 pdurbin
Great! Thanks!
14:18 pdurbin
I learned that DSpace uses Flyway. "The database schema used by DSpace is initialized and upgraded automatically using Flyway DB ." https://wiki.duraspace.org/display/DSDOC6x/Storage+Layer
14:19 pdurbin
I don't see any guidance for DSpace developers though. I guess Flyway "just works" and is magic. :)
15:05 pameyer joined #dataverse
15:06 poikilotherm
pameyer pdurbin : I just created https://github.com/IQSS/dataverse/issues/5361
15:07 * poikilotherm
is crossing fingers this makes sense to you guys
15:09 pameyer
pdurbin poikilotherm - web based installer?
15:09 poikilotherm
That's a dream of pdurbin... :-D
15:10 poikilotherm
It might be cool, but it needs more work... ;-)
15:10 poikilotherm
And there are some difficulties to tackle first, like reloading and stuff
15:12 pameyer
reminds me of having the force piwiki to be installable non-interactively :(
15:13 pameyer
but yeah, it's probably down the road
15:14 poikilotherm
force? piwiki?
15:15 pameyer
was looking into piwik; my usual assumption is that anything I'm going to use seriously has to be something where I can check the install into config management
15:16 pameyer
piwik support for that was "call us" - ended up being run manual install once, do surgery to make templateable and disable phoning home
15:16 pameyer
it's mameo (or something close to that spelling now)
15:16 poikilotherm
Matomo
15:16 poikilotherm
;-)
15:17 pameyer
yup :)
15:18 poikilotherm
IMHO there should be both options
15:18 poikilotherm
Actually when the bootstrapping is done from within the code, that would make the DevOps stuff easier... You don't need to call the installer
15:18 poikilotherm
And of course this can be extended further
15:19 Sherry joined #dataverse
15:19 poikilotherm
But changing runtime config most certainly should not be part of the boostrapping
15:20 pameyer
I'm generally in favor of making devops type stuff easier ;) but sometimes you do have to duct tape things
15:20 Sherry
UVa's Dataverse (V. 4.8.6) has stopped twice a day over the weekend (it also stopped each of the prior three days).
15:21 Sherry
What can we send (log-wise) for someone to look at tthings for us?
15:22 pameyer
Sherry : that sounds like things are getting worse
15:23 pameyer
glassfish log is one of the usual recommendations (but make sure you don't post it publicly, or scrub the DOI credentials). thinking on what other system logs would be helpful...
15:24 Sherry
We have done nothing with the dataverse app (no upgrades, no database updates) in months. But just last week it's been "falling over" (as my sysadmin says) at least once a day.
15:24 pameyer
I use that phrase too - although usually with some additional adjectives/adverbs
15:25 pameyer
two first possibilities I'd think of - more crawler/bot/unauthenticated traffic, or increase in users doing more upload/ingest type stuff
15:26 pameyer
apache (or httpd) logs might be helpful for the first one
15:26 pameyer
do you know the OS and distribution of your server(s)?
15:28 Sherry
Will have to ask about OS
15:28 pameyer
ok - it would make a difference for what system logs are available, and what they're called
15:29 pdurbin
Sherry: please email your server.log file to support
15:29 pameyer
if you're asking your sysadmin, it would also be worth asking if systat logs are availble
15:29 Sherry
Thanks, Phil and Pete Our sysadmin is packaging them and will get them out later today.
15:30 pdurbin
Sherry: did you show him or her http://munin.greptilian.com/greptilian.com/server3.greptilian.com/index.html yet?
15:31 Sherry
Will do now.
15:32 pdurbin
Thanks. Again, that's my home server but hopefully it'll prompt a discussion of monitoring stategies that they have already put into place for you.
15:33 pdurbin
poikilotherm: how many minutes or hours until you leave today?
15:35 poikilotherm
About 15
15:37 pdurbin
!
15:37 poikilotherm
pdurbin would you like me to open an issue + pr for the graphviz stuff?
15:38 poikilotherm
That might fit in the timeslot left for today...
15:38 poikilotherm
(at least the issue)
15:39 pdurbin
poikilotherm: oh. Maybe. I was thinking that we (you :) ) could maybe write a little bit more on the http://guides.dataverse.org/en/4.9.4/developers/documentation.html page to explain that there are two options for creating images. Maybe pros and cons. Just an idea.
15:39 poikilotherm
:-D
15:39 poikilotherm
Alright, THAT won't fit in the timeslot ... ;-)
15:40 pdurbin
poikilotherm: I'm also wondering if you have any feedback on my feedback. :) The half feedback I gave you at http://irclog.iq.harvard.edu/dataverse/2018-11-30#i_80663
15:40 pameyer
dot -Tsvg $foo > $figure ;)
15:40 pdurbin
I have more feedback to give on the second half of your post. But I thought I should pause and listen first. :)
15:41 poikilotherm
Right. Feedback following
15:42 poikilotherm
About the microservice vs monolith: actually I think this might have been misleading. It depends on what definition you read for microservices. Maybe lets stick with SOA
15:43 poikilotherm
It doesn't really matter if Dataverse is a monolith or not, you guys keep integrating more components and services, so that at least makes a SOA
15:43 poikilotherm
SOA = service oriented architecture
15:43 pdurbin
ok
15:43 poikilotherm
Of course you could split up Dataverse more, but that's not important right here, right now.
15:43 pdurbin
agreed
15:43 pameyer
centos7 sysstat looks like it's too old; doesn't have the new file descriptor stuff
15:44 pdurbin
pameyer: what about sar?
15:44 pameyer
sysstat is the package for sar
15:44 pdurbin
oh :)
15:44 pameyer
I was really excited to see that file descriptors were in the sysstat docs for a moment there
15:45 poikilotherm
About people using Docker/Kubernetes: you are right that no current production instance is using it. I just had the impression that future production instances like DataverseEU, FZJ, and more might be an indication that there is a tendency to have both ways!
15:45 poikilotherm
I meant not more = majority, just more = more than currently 35 prod. installations
15:46 pdurbin
poikilotherm: sorry to interrupt but before you go, I forget if you're in Berlin or not but Slava and Jonas are having a Dataverse meeting there this week: https://github.com/IQSS/dataverse/issues/4772#issuecomment-443412252
15:46 poikilotherm
I am with you on wording about "should" and "must"
15:46 poikilotherm
Nope, other side of Germany... ;-)
15:46 pdurbin
ah, bummer, oh well
15:47 pdurbin
poikilotherm: does Red Hat's interest in running Dataverse on OpenShift count as interest in Kubernetes?
15:47 poikilotherm
About understanding of adoption of Microprofil APIs: I might have misinterpreted this...
15:47 poikilotherm
Of course it does!
15:47 pdurbin
ok
15:47 pameyer
docker/kubernetes/etc does seem to be the way things are going.
15:48 poikilotherm
And I think they actually did the groundwork stuff to get this stuff inside IQSS view of how people use Dataverse
15:48 pameyer
I'd be curious to know where the tradeoff point for advantages vs complexity increases is for dataverse though
15:49 poikilotherm
About your agony: will try my best :-)
15:49 poikilotherm
And that's it for today for me. Sorry guys, gotta go now
15:49 pameyer
have a good night
15:50 poikilotherm
Maybe I can come back later, depends on my workload for the evening... ;-)
15:51 pdurbin
I guess I want to hear more about the groundwork.
18:21 pameyer
pdurbin: I checked wikipedia over the weekend, and a payara is a fish with fangs
18:43 pdurbin
heh, that sounds right
20:00 poikilotherm joined #dataverse
20:01 poikilotherm
Good evening dear IQSS devs and others
20:02 pameyer
good afternoon, european devs ;)
20:04 pdurbin
the Sun's getting real low
20:04 poikilotherm left #dataverse
20:05 poikilotherm joined #dataverse
20:05 poikilotherm
Well, you should have about 1.5 - 2 hours of sunlight left, shouödn't you?
20:05 poikilotherm
-ö+l
20:06 poikilotherm
pdurbin pameyer : any feedback on #5361 at standup/...?
20:06 pdurbin
about an hour until sunset, an hour and a half until dusk
20:07 pdurbin
standup is not where we talk about new issues
20:07 pdurbin
that happens when we estimate issues on wednesday afternoon
20:07 poikilotherm
Alright :-)
20:07 poikilotherm
Then I keep patient till Wednesday
20:07 pdurbin
but only if the issue has been flagged as "ready for estimation"
20:07 pdurbin
which is something only Danny does
20:07 poikilotherm
*G*
20:07 pdurbin
if you email him and ask nicely
20:08 pdurbin
so, no, we didn't talk about it
20:08 poikilotherm
Maybe we should just have a VC about 5292 and related... :-D
20:08 pameyer
I'm on the wrong side of the river to know about standup today
20:08 pdurbin
pameyer comes to standup twice a week
20:08 poikilotherm
Yeah you said he is only with you twice a week or so?
20:09 pdurbin
yeah
20:09 poikilotherm
I think you mentioned that during talking about Docker builders on your floor ;-)
20:09 poikilotherm
(or the absence of em)
20:09 pdurbin
poikilotherm: we have our usual community call tomorrow. Want to ask there?
20:09 poikilotherm
About 5292 and related?
20:09 pdurbin
we have the community call every other week
20:09 pdurbin
yeah
20:10 pameyer
is this one the first w\ the new call-in option?
20:10 poikilotherm
I could try, but most certainly I will not be available at 6pm
20:10 poikilotherm
Kids + work on house stuff...
20:10 pdurbin
it's at noon our time
20:10 poikilotherm
Yeah
20:11 pdurbin
pameyer: no idea what the call in options will be
20:12 poikilotherm
Will ask my government later when she plans to be home tomorrow evening and maybe then I will put my plans aside and join ;-)
20:12 pdurbin
heh, ok
20:14 poikilotherm
pdurbin you asked about writing stuff about images for docs
20:14 poikilotherm
Anything specific?
20:14 pdurbin
Oh, I was just spitballing ideas.
20:14 * poikilotherm
seeks cover
20:14 pdurbin
The release notes will link to closed GitHub issues.
20:15 * poikilotherm
raises his head
20:15 pdurbin
And it would be nice if one of the issues was something like, "As a documentation writer, I'd like to create images with inline markup." Or something. Fixed! In the next release.
20:15 pdurbin
I hope that makes sense.
20:15 poikilotherm
Sure!
20:16 poikilotherm
That was on my mind when I commented on your comment on #5288#
20:16 pdurbin
Oh, also, the doc/release-notes folder is for how to overcome breaking changes (sql update, solr schema update), not for advertising features.
20:16 poikilotherm
???
20:17 pdurbin
I lost you. Sorry.
20:17 poikilotherm
Then I completely missunderstood the docs about that
20:17 * poikilotherm
reads it again
20:17 poikilotherm
Gimme a sec
20:17 pdurbin
If the README is doc/release-notes needs to be updated, please feel free to make a pull request.
20:17 pdurbin
I wrote it. All errors are mine.
20:19 poikilotherm
Ok, just to make sure I get you right this time:
20:20 poikilotherm
We are talking about https://github.com/IQSS/dataverse/blob/develop/doc/sphinx-guides/source/developers/making-releases.rst#id18
20:20 poikilotherm
Right?
20:22 pdurbin
Yeah, I'm talking about what I wrote there as well as https://github.com/IQSS/dataverse/blob/092ea01af583670aebbf66fae659c293ca6a037d/doc/release-notes/README.md
20:23 poikilotherm
Hmm ok...
20:24 poikilotherm
Then maybe a proper definition of release notes is missing
20:24 pdurbin
Maybe.
20:24 pdurbin
Another thought is that I'd love to have a list of features in the guides somewhere. Especially since https://dataverse.org/software-features is so out of date.
20:25 poikilotherm
In my world, release notes tell me about new features to be aware of, bugs that were fixed and things to be aware of, as "there be dragons". Maybe something like an upgrade-barometer
20:25 pameyer
depends on if "release notes" means "cool new features" or "stuff that will break unless you update something"
20:25 pameyer
or both
20:25 poikilotherm
Like Gitlab Inc does, like Atlassian does etc
20:25 pdurbin
poikilotherm: I could forward you a thread about our list of features and my desire for it to be updated more often if you want.
20:26 poikilotherm
Yeah pameyer, you name it: its a definition thing
20:26 poikilotherm
And currently, the docs aren't precise about that - that's why I added a release note for #4690 in the first place... ;-)
20:27 poikilotherm
pdurbin: sure, go ahead :-)
20:32 pdurbin
poikilotherm: actually, instead of forwarding a thread a I'll just say that https://dataverse.org/software-features could be updated to include Dropbox integration, OAuth login (ORCID, GitHub, Google), Widgets, OAI-PMH, support for Zotero. Endnote, and Altmetrics. And other stuff I'm not thinking of. If we move the list to the guides we can use pull requests to maintain it.
20:34 pameyer
schema.org json-ld, bibtex
20:35 pdurbin
yeah
20:36 pdurbin
I mean, this stuff is scattered across the guides but it would be nice to have a single list, I think.
20:36 pdurbin
Maybe with nice images. I don't know.
20:36 pameyer
I didn't know dataverse supported altmetrics - my knowledge of it is limited to it being down when I'm trying to do perf measurements
20:39 donsizemore joined #dataverse
20:40 poikilotherm
pdurbin: Yeah, that page has a lot of missing things...
20:41 pdurbin
but a pull request won't fix it
20:41 poikilotherm
Yeah. But how does this relate to the release notes part?
20:42 poikilotherm
When I look at the GitHub release notes, those are pretty undetailed
20:43 pdurbin
We could do a better job of highlighting new features. Sometimes we make a blog post.
20:44 poikilotherm
Have you looked at how Gitlab des this?
20:44 poikilotherm
https://about.gitlab.com/releases/
20:44 poikilotherm
Always companioned by a larger blog post
20:44 poikilotherm
E. g. https://about.gitlab.com/2018/11/22/gitlab-11-5-released/
20:44 pdurbin
very nice
20:44 pdurbin
lots of structure
20:44 pdurbin
clean
20:45 poikilotherm
And always pointing to docs
20:45 poikilotherm
It is always a pleasure to read the blog posts
20:45 poikilotherm
But to get a quick overview, what has changed, the releases page is pretty neat
20:46 pdurbin
https://blog.github.com/changelog/ is pretty nice for knowing what new features GitHub has added.
20:49 poikilotherm
I feel pretty lost in that list.
20:49 poikilotherm
It seems more like a list of commits, rendered in some fancy way
20:49 poikilotherm
While looking at the dates, it seems they add stuff every day
20:50 poikilotherm
And that makes it less readable. Also there is no weight of things.
20:50 poikilotherm
At Gitlab, the order of the entries is important, as big changes are listed at the top
20:50 poikilotherm
But ok - this is a Changelog, not a release notes
20:52 pdurbin
yeah
20:55 poikilotherm
Alright - so to get a baseline on the GraphViz stuff: I will create the issue about inline images
20:56 poikilotherm
What parts of the docs do you want me to extend about images?
20:57 pdurbin
http://guides.dataverse.org/en/4.9.4/developers/documentation.html please
20:57 pdurbin
I was thinking... one advantage of the old approach is that the you can link to the png in GitHub.
20:58 pdurbin
With the Graphviz version, there's no image to link to.
20:58 pdurbin
Tradeoffs.
20:58 pdurbin
Nice to have choices.
20:58 poikilotherm
Link from where?
20:59 pdurbin
poikilotherm: I linked to a few PNGs on GitHub here, for example: https://groups.google.com/d/msg/dataverse-community/YjU38pFl3v4/Ju6Y86L_AQAJ
21:00 poikilotherm
Ok in case this is linking to stuff already in develop - wouldn't it be an easy option to have the compiled docs deployed for multiple versions and develop branch?
21:00 pdurbin
yeah, it would
21:00 poikilotherm
For develop this is a trivial Travis job..
21:01 pdurbin
it was just a thought
21:01 pdurbin
yeah, it exists already
21:01 pdurbin
I can go click "build now" if you want.
21:01 poikilotherm
Oh and if you don't want to mangle with own servers - isn't there GitHub pages for this?
21:02 pdurbin
yes
21:03 pdurbin
poikilotherm: oh, did you see my reply about chathistory and IRCv3?
21:03 poikilotherm
Yes, I saw that.
21:03 pdurbin
cool
21:04 pdurbin
I think there's a service called "IRC Cloud" or something that would give you persistence and scrollback. I believe it's a paid service, though.
21:04 poikilotherm
https://github.com/ircv3/ircv3-specifications/milestone/4
21:04 poikilotherm
This seems like it will take years, not days till this is available
21:05 poikilotherm
And when the spec is live, there will be a need for implementations client and server side
21:05 pdurbin
Maybe. I thought I heard that some IRC clients and servers have implemented IRCv3 stuff on an experimental basis, at least.
21:07 pdurbin
I don't know if freenode supports chathistory or not.
21:12 poikilotherm
https://github.com/sphinx-doc/sphinx/issues/3382
21:14 pdurbin
poikilotherm: what did you mean about the "IQSS view of how people use Dataverse"?
21:15 poikilotherm
Maybe its just my bad english - will try to elaborate more in hope you guys get me right...
21:15 poikilotherm
If I got it right, Dataverse was invented by IQSS
21:15 pdurbin
heh, yes
21:15 poikilotherm
IQSS is a institute focusing on social sciences and the needs of researches within that domain
21:16 pdurbin
true
21:16 poikilotherm
Then, one shiny day, you guys made a decision (dunno how that happenend, but this doesn't really matter) that Dataverse might do MORE
21:16 poikilotherm
That researches out there are interested in repos
21:17 poikilotherm
I am not sure if this has been first only a "here fellow social science domain researches, look what we did, maybe it helps", but the word spread
21:18 poikilotherm
And then it reached not only social science, nowadays there is a broader spectrum of domains interested in using Dataverse
21:19 poikilotherm
Just as the RDM stuff got momentum in a lot of countries, domains and researcher communities, the development was hit by that momentum, too
21:19 pdurbin
poikilotherm: in 2006 there were only two installations of Dataverse (then called "DVN" for Dataverse Network), and they were both social science institutes (IQSS at Harvard and Odum at the University of North Carolina): https://slides.com/mercecrosas/dataversecommunity2018/fullscreen#/5
21:19 poikilotherm
And not only on a level of "how to do XYZ in Dataverse" being a question from a researcher, it gained momentum on a technical level
21:20 poikilotherm
People started to ask for technical features
21:20 pdurbin
more people, sure
21:20 poikilotherm
And I am pretty sure, that many of those features came from domains outside of social sciences
21:21 andrewSC joined #dataverse
21:21 pdurbin
sure, pameyer's rsync feature for example
21:21 poikilotherm
Yeah!
21:21 poikilotherm
Big data!
21:21 poikilotherm
And more to come
21:22 poikilotherm
And with all those features, a lot of admins started to adopt dataverse
21:22 pdurbin
hopefully, did you see plot in the slides above? :)
21:23 poikilotherm
As time moved on, admins started to adapt new techniques like Docker and eventually Kubernetes. This of course started slowly and not everywhere, but this is still growing. Containers are here to stay
21:23 pdurbin
well
21:23 pdurbin
oh
21:23 pdurbin
I see what you mean.
21:23 pdurbin
you mean sysadmins
21:24 pdurbin
sysadmins are playing with Docker more these days, yes
21:24 pdurbin
and they talk a lot about Kubernetes
21:24 pdurbin
I'm not sure how much Kubernetes is used in production.
21:25 poikilotherm
I don't know if pameyer created docker-aio before or after the Redhat Openshift people started their work. Anyway, I get the impression that altough you core devs at IQSS are not into Docker, all those input from others widened your view and you started to adapt Dataverse to this stuff
21:25 pdurbin
after, I can repost the timeline :)
21:26 pdurbin
I don't use Docker every day. No.
21:26 pdurbin
I have it installed. That's something.
21:26 poikilotherm
Honestly, I don't think nowadays there is a way around it. If you guys want Dataverse to grow and spread, you most likely need to adapt further
21:26 pdurbin
I run docker-aio maybe once a month. I like it.
21:26 pdurbin
we need to evolve
21:27 poikilotherm
And IMHO there are little chances to escape the influence it wll have on you and your workflows
21:27 pdurbin
no escape
21:27 poikilotherm
When I started my proffessional education in 2006, there was no word about all this
21:28 pdurbin
sure, Docker is new, containers are new
21:28 poikilotherm
Nowadays I pretty much enjoy the benefits. The ecosystem of Docker has been significantly getting more mature
21:28 pdurbin
Java is old.
21:28 pameyer
docker's new; containers are old
21:28 pdurbin
But not as old as Python. :)
21:28 pdurbin
pameyer: old-ish :)
21:28 poikilotherm
pameyer: Yes! Jails from BSD for the win :-D
21:28 pdurbin
BSD jails are old. :)
21:28 pdurbin
jinx
21:29 poikilotherm
I remember around 2008 at university we playes with OpenVZ and Jails
21:29 pameyer
sorry - didn't mean to derail things ;)
21:29 poikilotherm
And look where we are now :-D
21:29 pdurbin
We used to run Bind in chroot.
21:30 poikilotherm
Oh man - that really s****d
21:30 poikilotherm
All those device nodes you had to create
21:30 poikilotherm
And then crossing fingers you did everything right
21:30 pdurbin
There was an RPM for it. Not so bad.
21:30 poikilotherm
We did it hard core on GEntoo :-D
21:30 poikilotherm
Well, we were yung
21:30 pdurbin
dang
21:31 poikilotherm
Good old days... Not so far away anyway ;-)
21:32 poikilotherm
I hope my "IQSS view on people" is now a bit clearer :-)
21:32 poikilotherm
I dunno if you agee with me or if this is just my stupid brain - just an outside view on you guys ;-)
21:32 pdurbin
You think we're a bit stuck in the past.
21:33 pdurbin
Java. PostgreSQL. Solr.
21:33 poikilotherm
That's not a bad ting
21:33 poikilotherm
No no, don't get me wroong
21:33 pdurbin
Where's the Mongo?
21:33 poikilotherm
Java is supercool
21:33 poikilotherm
Postgres too
21:33 pdurbin
Where's the node
21:33 poikilotherm
Solr is a bit oldschool, Elastic has far more devs and power, but ok
21:34 poikilotherm
Mongo might be easier for the data model when coming to metadata. That was why I picked it for PeerPub
21:34 poikilotherm
But I understand that back in the old days, there was no NoSQL
21:34 poikilotherm
So that's fine with me
21:34 poikilotherm
And not to forget - Postgres knows JSON these days
21:35 pdurbin
Yeah, a lot of this stuff didn't exist when we started.
21:35 pdurbin
IRC was around. :)
21:35 poikilotherm
Dataverse is a cool tool!
21:35 poikilotherm
I don't know if you have resources to refactor
21:35 pdurbin
nope
21:35 pdurbin
You don't get grants to refactor code.
21:35 poikilotherm
As I learned someday, in software dev business, you can do what you want. The day will come you need to refactor
21:36 poikilotherm
(Or start over)
21:36 poikilotherm
That's the problem with research software ;-)
21:36 poikilotherm
And as I am very aware of this, coming from a bigger research institute, I am definitly ok with this
21:36 pdurbin
We just did a big refactor from DVN 3.x to Dataverse 4. It took forever.
21:37 poikilotherm
Well, you have some very smart people around that built a truly unique app
21:37 poikilotherm
I really like your approach on Metadata flexibility
21:37 poikilotherm
Thats a unique thing I cannot emphasize enough
21:38 poikilotherm
That was the key factor for us to pick Dataverse
21:38 pdurbin
Nice. That's why I'm happy we finally wrote some docs on metadata flexibility: http://guides.dataverse.org/en/4.9.4/admin/metadatacustomization.html
21:38 poikilotherm
:-)
21:39 pdurbin
and the next release will include info on how to load them up
21:39 poikilotherm
Great!!!!
21:39 pdurbin
pameyer figured it out without docs. Amazing.
21:40 poikilotherm
I hope you guys foregive me that I am pushing the new stuff...
21:40 pdurbin
I like the new stuff. An external contributor got us from Ant to Maven.
21:40 poikilotherm
Cool :-D I digged into this a bit, but definitly not all the way through
21:40 poikilotherm
Hopefully pameyer wrote some of the docs ;-)
21:40 pdurbin
Usually, "new stuff" for us is the latest version of Java EE.
21:41 poikilotherm
;-)
21:41 pdurbin
DVN 3 was Java EE 6. Dataverse 4 was Java EE 7. Java EE 8 is out. Not sure what the next version will be called.
21:42 poikilotherm
Yeah, you mentioned that before :-)
21:42 pdurbin
"new stuff" is also the next version of Java that isn't EOL
21:42 pdurbin
newer versions of postgres and solr
21:42 poikilotherm
And no offense: I think you guys use just about 40% of EE 7
21:42 poikilotherm
But that's ok
21:42 pdurbin
yeah
21:43 poikilotherm
I hope I can give back some of my experience and knowledge into this project, as I truly believe in the power of OSS
21:43 pdurbin
poikilotherm: that reminds me, if you're on the other side of Germany, does that mean you're close to FOSDEM?
21:43 poikilotherm
We will benefit from your work on Dataverse for a long time, hopefuly
21:44 poikilotherm
Yes
21:44 pdurbin
Cool. I'd like to go to FOSDEM some day.
21:44 poikilotherm
:-)
21:44 poikilotherm
I was thinking about going there next year
21:44 pdurbin
I've started going to LibrePlanet. The past three years. It's local.
21:44 poikilotherm
Brussels about 2-3 hours from here
21:45 pdurbin
gotcha
21:45 poikilotherm
And I think I heared rumors about a RSE track
21:46 pdurbin
huh, cool
21:46 poikilotherm
Those people at SSI in Edinbourgh do a pretty good job
21:46 pameyer
the RSE stuff seemed to start picking up in Europe a few years back
21:46 pdurbin
I've heard good things about them.
21:47 poikilotherm
Yeah, in UK this is already a big thing. Slowly starting in Germany, too
21:47 poikilotherm
deRSE has been founded just a week or two ago
21:48 pameyer
"new stuff" is sometimes a game changer, and sometimes not worth the switchover effort
21:48 pameyer
at least, in my experience.
21:48 poikilotherm
That's true.
21:53 poiki-at-home joined #dataverse
21:53 pameyer
but there's also gradiations of "old stuff". aka - fortran77 code that a researcher's been cranking out new releases on for decades is one thing, old php + java applets is another thing
21:53 poiki-at-home
Gnrf, freenode just kicked me out
21:53 poiki-at-home
Need to read up logs first
21:53 poiki-at-home
(Hey pdurbin, here we go again... )
21:55 pameyer
don't think you missed much
21:55 poiki-at-home
Nope
21:55 poiki-at-home
My messages were not send...
21:56 pameyer
:(
21:57 poiki-at-home
As long as you are online, IRC is just fine
22:00 pameyer
IRC from tmux sessions has been reasonably stable for me
22:00 poiki-at-home
But just as long as you uplink is stable...
22:00 poiki-at-home
+r
22:01 pameyer
true
22:12 poiki-at-home
pdurbin hopefully this makes you happy: https://github.com/IQSS/dataverse/pull/5366
22:12 poiki-at-home
Please add review comments if and where I should elaborate more
22:23 pdurbin
poikilotherm: looks great! I moved it to QA. You *could* mention plantuml but whatever. :)
22:24 poikilotherm
plantuml needs a special plugin and further deps installed
22:24 pdurbin
yeah
22:24 poikilotherm
By the way: Gitlab is doing a great job at rendering plantuml in the brower
22:24 poikilotherm
Is there sth. similar in GItHub?
22:24 pdurbin
poikilotherm: oh, and it's probably not freenode's fault that you're getting disconnected. There's a memory leak in the version of Shout that I installed: https://github.com/IQSS/chat.dataverse.org/issues/3
22:25 poikilotherm
I am using an IRC client
22:25 poikilotherm
Not the webpage ;-)
22:36 poikilotherm
Are you guys using bouncers?
22:40 poikilotherm
Thinking about setting up ZNC on my home server...
22:41 poikilotherm
Anyway. I will leave for getting some sleep now. Have a good night guys :-)
