Time
S
Nick
Message
02:40 shauna joined #dataverse
05:14 axfelix joined #dataverse
08:02 jri joined #dataverse
11:18 pdurbin
bencomp: I have no idea why this was marked as spam but I just approved/released it: https://groups.google.com/d/msg/dataverse-community/3yfbb606EJU/kAU2fnWJCwAJ
12:45 bencomp
hm, strange
12:46 bencomp
but thanks for approving it :)
13:29 pdurbin
sure. I whitelisted your address too
13:31 bencomp
pdurbin: did you see https://github.com/IQSS/dataverse/pull/2818#issuecomment-169929055 ?
13:32 bencomp
I bet lwo would be interested in discussing a plugin system too
13:39 pdurbin
I did see that.
13:39 pdurbin
What sort of plugin would lwo work on?
13:41 bencomp
maybe not him personally, maybe Slava. Remember the presentation at the DCM?
13:43 pdurbin
oh sure. yeah, let's ping slava if he jumps in here
13:43 pdurbin
I think what might be nice is to start a list of potential plugins
13:44 bencomp
plus the interfaces that each might need - file access, metadata access, read-only, read-write, etc.
13:44 bencomp
got to go do other things now, back to lurk mode ;)
13:45 pdurbin
right. would file storage abstraction be a plugin? using S3 instead of the file system? maybe
13:47 bencomp
that would be a great example
13:48 bencomp
local file system would be the "default plugin"
13:49 bencomp
Handle or DOI or URN or XYZ could also be plugins
13:49 pdurbin
sure. I'm just thinking that technically these would be very different plugins
13:49 pdurbin
front end vs. back end infrastructure
13:50 bencomp
true
13:50 bencomp
focus on one thing first
13:50 bencomp
:)
13:50 pdurbin
what fun is that? ;)
13:56 bencomp
possibly the fun of feeling good after you get things done
13:57 bencomp
;)
13:57 pdurbin
heh
14:06 jri joined #dataverse
14:16 jri joined #dataverse
15:11 pameyer joined #dataverse
15:12 pdurbin
pameyer: hey. so I have a new doc about process and requirements I'd like to share with you
15:12 pameyer
pdurbin: great
15:13 pdurbin
2016-01-07-dataverse-project-management.pdf
15:13 pameyer
pdurbin: I'm guessing this is related to how to figure out and document requirements in a way that works well?
15:14 pameyer
where?
15:15 pdurbin
please check your email. just sent it
15:15 pameyer
just got it
15:16 pdurbin
I took what to me are the key slides from yesterday's talk and put them in a single PDF that will print on one page (double sided).
15:17 pdurbin
pameyer: since the grant we're on is for 3 years I'd like you to think about how to break the work up into projects since each "unit of integration" (pull request) is a project.
15:17 pdurbin
does that make sense?
15:18 pameyer
it does
15:19 pameyer
looking like it's unfortunate I didn't make it into the training class about project management
15:20 pdurbin
heh
15:20 pdurbin
well this is all new. you can help us figure it out :)
15:20 pameyer
it is - the problem is figuring out the problem...
15:20 pdurbin
:)
15:21 pameyer
but if I end up coming up with a project / BRD/RFD that's too close to "make the magic black box do $x" - please tell me about where the problems are :)
15:23 pdurbin
heh. sure. So to me the first project is changing the Dataverse code enough to have sufficient feature parity with https://data.sbgrid.org so that we can cut over DNS to a new system running Dataverse rather than the legacy/existing system. The problem is... I'm still fuzzy on what the features even are. :)
15:23 pdurbin
and maybe we could slip in a few new features. I don't know
15:24 pdurbin
but we should be careful to not try to do too much in the first project. think Minimum Viable Product (MVP)
15:24 pameyer
the current features (and a few that we're about to push), should be easy enough for me to nail down
15:24 pdurbin
ok. good
15:24 pameyer
I *try* to keep the complexity down - but not always successfully
15:24 jri left #dataverse
15:25 pdurbin
so what would we even call the first project?
15:25 pameyer
and there's always the situation where what seems like the lowest-complexity to me isn't globally lowest-complexity
15:26 pameyer
do we want to consider finding the "pain points" a project (aka - it's not software, may not correspond to a branch)
15:26 pameyer
?
15:26 pdurbin
I want to know all the pain points. Each pain point should probably be a GitHub issue.
15:26 pdurbin
oh wait, "it's not software"?
15:27 pdurbin
everything is software to me ;)
15:27 pameyer
:)
15:27 pameyer
sadly I haven't been able to figure out how to version control my thought processes
15:27 pdurbin
:)
15:28 pameyer
How about "documenting current implementation functionality sufficiently that it's understandable to people that didn't write it"?
15:28 pameyer
but a shorter name
15:29 pdurbin
pameyer: well, all of our docs are in the Dataverse source tree so a GitHub issue is fine for that: http://guides.dataverse.org/en/4.2.2/developers/documentation.html
15:32 pameyer
pdurbin: ok, that sounds like the way do go
15:33 pdurbin
we even have a label for documentation: https://github.com/IQSS/dataverse/labels
15:33 pdurbin
26 open issues!?! d'oh!
15:33 pameyer
pdurbin: I may not be able to get started on it today, we got reviewer/editor comments on a paper. but this is next on my list
15:34 pameyer
so still only double-digits - sounds not too bad
15:35 * pdurbin
calms down
15:35 pameyer
pdurbin: separate branch for "feature parity" docs to start with?
15:36 pdurbin
pameyer: I just got access to the list of projects. there's one called "Upload Large Data Files" that's associated with our grant.
15:38 pdurbin
pameyer: sorry, I don't understand what you mean about docs. Docs about what?
15:38 pdurbin
User Guide? Installation Guide?
15:39 pameyer
pdurbin: docs about features of our current implementation
15:39 pdurbin
hmm
15:40 pdurbin
well, we won't be documenting your legacy app in the Dataverse docs, of course
15:41 pameyer
(facepalm) - excellent point
15:42 pdurbin
I *think* this is all BRD territory.
15:42 pdurbin
what the business requirements are
15:43 pameyer
so this reduces to the problem of "where do BRDs live?"
15:43 pdurbin
hopefully soon there will be a template or example for us to follow
15:45 pdurbin
I believe the BRDs will be written collaboratively in Google Docs. Once approved, they will be converted to Markdown or RST and committed to the Dataverse source tree on GitHub. Maybe under the Dev Guide. That's my understanding anyway.
15:46 pdurbin
this is hard without an example to look at. let me check if there is one yet
15:46 pameyer
ok. markdown + git seems easier for collaborative editing than google docs (at least for me); but other people have different opinions
15:47 pdurbin
well, Google docs are nice for more fluid editing early on. but yes, I hear you
15:48 pameyer
pretty much anything is better than emailing MS word docs around...
15:48 pdurbin
heh
15:48 pdurbin
amen
15:49 pameyer
* pameyer resists ranting about the failings of MS word
15:50 pdurbin
pameyer: looking again at the PDF I sent you... maybe we should focus more on the FRD rather than the BRD... not sure
15:50 pdurbin
"functional requirements"
15:51 pdurbin
I just want us to start writing down requirements somewhere.
15:52 pdurbin
:)
15:52 pameyer
I think you're right
15:52 pameyer
BRD = retroactively write FRD for current implementation
15:53 pdurbin
right
15:53 pdurbin
I mean, I'm thinking backward from the pull request anyway. What's in the pull request? What functionality has been added? Do the tests pass?
15:54 pameyer
the current implementation is too minimal of an MVP to have much in the way of tests :<
15:54 pdurbin
that's fine. we can write new tests. they'll go in the pull request
15:55 pameyer
at least, I don't consider a few "def test1()" , "def test2()" run from main (and commented out in normal operations) to be much
15:55 pameyer
which test framework does DV use?
15:55 pdurbin
I forget if I showed you this: Improve automated testing · Issue #2746 · IQSS/dataverse - https://github.com/IQSS/dataverse/issues/2746
15:56 pdurbin
we use JUnit and REST Assured
15:56 pameyer
I don't think you showed me 2746 before
15:56 pdurbin
ideas for that issue are very welcome
15:57 pameyer
are you thinking about tests for current/django/legacy?
15:57 pdurbin
how about we port whatever tests you have
15:58 pdurbin
standup meeting. back in a bit
16:00 pameyer
enjoy
16:16 pdurbin
back
16:16 pdurbin
pameyer: are you familiar with the concept of a "user story"?
16:17 axfelix joined #dataverse
16:17 pameyer
pdurbin: I've heard the phrase before, but that's pretty much it
16:18 pdurbin
that's kind of what I was going for at https://trello.com/c/z961tZzk/32-resume-large-downloads
16:19 pdurbin
it's not quite a user story, which would be more like "Alice's download stops halfway so she resumes it by..."
16:21 pdurbin
anyway, user stories are nice since they're so concrete
16:21 pameyer
it's sounding like the point of a user story is to make sure the requirements are a match for what users are likely to do with it - is that close?
16:21 pdurbin
right. exactly
16:22 pameyer
for this case, corresponding entry in FRD would be "user is able to resume partially completed download"?
16:22 pameyer
or BRD?
16:23 pdurbin
not sure. let's just call it a requirements doc for now. and let's use git and markdown if that's easier for both of us
16:23 pameyer
that works for me
16:23 pameyer
minimum viable documentation
16:23 pdurbin
pameyer: want to make a repo under https://github.com/sbgrid ?
16:24 pameyer
ok - but I don't have access to that org yet
16:25 pameyer
but that's a fixable problem
16:25 pdurbin
oh oh. I can make it under IQSS if that's easier
16:25 pameyer
I can make one under https://github.com/pameyer and move it later
16:26 pameyer
the guys I need to talk to about it are both off today
16:26 pdurbin
sounds great. I could even enable the wiki at https://github.com/IQSS/dataverse which is backed by git. but let's go with a repo under your account for now
16:27 pdurbin
I'm thinking it's just a single markdown file for now. And by the way, this is awesome: https://pypi.python.org/pypi/grip/
16:29 pameyer
https://github.com/pameyer/sbdg_requirements0
16:29 pdurbin
awesome. do you want me to make pull requests? or just give me access?
16:30 pameyer
I was thinking to give you access - but trying to figure out how to
16:30 pameyer
and you should have access now
16:35 pdurbin
yep, I was able to push. perfect. thanks
16:37 pdurbin
pameyer: I don't mean to keep you from your paper comments
16:38 pameyer
I'm multi-tasking with the low complexity comments - about to get into the ones that require more thinking
16:38 pameyer
pdurbin: let's pick this up monday
16:39 pdurbin
sounds great
17:24 axfelix joined #dataverse
17:47 metamattj joined #dataverse
17:49 xvx joined #dataverse
20:22 metamattj joined #dataverse
