Time
S
Nick
Message
02:16 garnett joined #dataverse
03:50 axfelix joined #dataverse
04:29 ata_t joined #dataverse
04:35 garnett joined #dataverse
04:58 yan___ joined #dataverse
07:30 ata_t1 joined #dataverse
08:25 bencomp joined #dataverse
09:27 bencomp1 joined #dataverse
12:14 pdurbin
nice. someone new looking at the code: https://github.com/IQSS/dataverse/issues/2706
12:34 bencomp
pdurbin: indeed. and a good catch too
12:35 pdurbin
I wonder if he/she has measured how expensive it is.
12:36 bencomp
good question too
12:43 ata_t joined #dataverse
14:29 ata_t joined #dataverse
15:11 pdurbin
bencomp: I just left a comment on that issue
15:13 bencomp
pdurbin: I noticed :)
15:15 bencomp
I added a comment to 2653: https://github.com/IQSS/dataverse/issues/2653#issuecomment-152190073
15:15 pdurbin
I saw that. Great comment.
15:18 bencomp
I didn't yet respond to mheppler's update in https://github.com/IQSS/dataverse/issues/2232#issue-84184258 but it's another "?!" moment for me
15:22 pdurbin
bencomp: huh. can you reproduce that issue? It's quite old.
15:23 bencomp
pdurbin: as long as it's open, I assume it's still an issue
15:24 pdurbin
no, maybe it was fixed
15:24 bencomp
how should I know?
15:25 pdurbin
bencomp: that's why I'm asking you if you can reproduce it
15:26 pdurbin
this is something the community can help with. if they want to
15:26 pdurbin
no pressure :)
15:28 pdurbin
Someone from the community could try it on https://demo.dataverse.org and indicate if the bug still exists or not as of the version in the bottom right: v. 4.2 build 92-d6443eb
15:28 pdurbin
I hope this makes sense.
15:29 pdurbin
I can write this up under https://github.com/IQSS/dataverse/blob/master/CONTRIBUTING.md if it's helpful.
15:30 ata_t joined #dataverse
15:31 bencomp
pdurbin: it is part of my grand plan to get you (plural) to get your stuff together - manage your product, address issues in your issue list, doing the stuff that is described in 650+ TODOs and @todos
15:31 bencomp
noone even seems to care about duplicate issues
15:32 pdurbin
bencomp: are you telling me you don't see any role for the community in helping us? :(
15:32 bencomp
and kcondon is supposed to manually test every proposed solution, it seems
15:35 pdurbin
I like to think the community wants to help. There's plenty of evidence that they are helping.
15:35 bencomp
of course the community wants to help
15:36 bencomp
but you can't assume that "the community" will go through the issue list to clean it up
15:36 pdurbin
even if we ask them to?
15:36 pdurbin
(not that I've asked yet)
15:39 pdurbin
it's just an idea
15:39 pdurbin
to have the community go through old issues and see if they've been fixed already
15:40 pdurbin
maybe it's a bad idea
15:40 bencomp
that does sound like a bad idea
15:40 pdurbin
fair enough
15:40 bencomp
"we've not been able to keep our issue list up to date, could you help us?"
15:41 pdurbin
right. something like that
15:41 bencomp
you could start by closing issues as duplicates when someone suggests it :)
15:42 bencomp
and do not shy away from updating issue titles to make them clearer
15:43 bencomp
even if someone outside the team created the issue
15:43 bencomp
(would be nice to check if they agree)
15:53 pdurbin
bencomp: I'm checking first. I just left a comment here: https://github.com/IQSS/dataverse/issues/2700
15:55 bencomp
pdurbin++
15:56 bencomp
it should help to have a clear view of what is actually problematic
15:57 pdurbin
yeah
15:58 bencomp
in this case you suggest replacing "data sets" with "files" - does that really tell you what the issue is about?
15:58 bencomp
being extra critical here ;)
16:24 pdurbin
bencomp: what title would be better?
16:27 bencomp
a title that represents the core of the problem (or symptom) or goal functionality so that you don't need to read the description to understand
16:27 bencomp
:) in this case....
16:29 bencomp
"Dataverse should provide persistent URLs for files", for example
16:30 garnett joined #dataverse
16:43 pdurbin
bencomp: makes sense. maybe he'll see this chat and consider that as a title
16:44 pdurbin
or you could leave a comment
16:46 bencomp
done
16:48 bencomp
pdurbin: I started reading the book "Refactoring" by Marting Fowler. I think you would like it :)
16:48 bencomp
s/Marting/Martin/
17:46 pdurbin
yeah, I like that guy
18:15 pdurbin
I liked http://martinfowler.com/bliki/MonolithFirst.html and http://martinfowler.com/articles/dont-start-monolith.html
18:20 garnett joined #dataverse
18:37 ata_t joined #dataverse
18:51 ata_t joined #dataverse
19:05 ata_t joined #dataverse
19:18 ata_t joined #dataverse
19:38 ata_t joined #dataverse
19:41 bencomp joined #dataverse
20:03 garnett joined #dataverse
20:03 bencomp
pdurbin: look at you closing stale pull requests :)
20:04 bencomp
are you ready to look at new ones or shall I wait a bit longer?
20:08 pdurbin
bencomp: if you write tests and keep them small, sure :)
20:11 pdurbin
bencomp: what I really want are pull requests that fix user-facing bugs
20:11 pdurbin
bencomp: do you feel like working on this one? https://github.com/IQSS/dataverse/issues/2710
20:12 pdurbin
it's a fresh issue! only an hour old :)
20:12 ata_t joined #dataverse
20:12 ata_t joined #dataverse
20:13 bencomp
pdurbin: thanks, but no thanks
20:13 bencomp
still can't work on coding for Dataverse except in my own time
20:16 pdurbin
bummer
20:16 bencomp
and to prevent that fixing any new bug takes more than a week I would suggest to work on code quality :)
20:17 pdurbin
hmm
20:17 pdurbin
I'm all for code quality but I don't think it's a cure all.
20:17 bencomp
break up 300+ line methods into readable bits
20:17 bencomp
then I can quickly propose a change
20:18 bencomp
for example
20:21 pdurbin
that makes sense
20:24 bencomp
e.g. DownloadInstanceWriter#writeTo is not quite 300+ lines, but ~200 with lots of branches
20:24 bencomp
I can't make any change
20:24 bencomp
because it's extremely hard to test properly
20:25 pdurbin
I'm not familiar with that class but I believe you.
20:25 bencomp
and as Fowler writes in Refactoring: before you refactor, make sure to write a test so you know that your refactoring doesn't change behaviour
20:25 bencomp
8-)
20:26 bencomp
B-)
20:28 pdurbin
yes! exactly!
20:29 pdurbin
bencomp: this is why your big pull requests are scary! not enough tests in the code
20:32 bencomp
pdurbin: you're right
20:35 bencomp
not enough tests is also why I am amazed by how bugs like The Bug https://github.com/IQSS/dataverse/issues/2648 are being "fixed"
20:35 bencomp
and then "fixed for reals"
20:37 pdurbin
bencomp: I do want your suggestions in that issue for refactoring to live on. That would be a great pull request!!
20:38 bencomp
pdurbin: it has to be a mindset
20:39 bencomp
ok, enough work stuff for me for today
20:42 pdurbin
bencomp: a pleasure chattign with you, as always
20:44 bencomp
glad you still think that way, pdurbin. I feel the same - and the discussion is far from over ;)
20:56 pdurbin
:)
21:48 pdurbin
I attempted to answer http://economics.stackexchange.com/questions/8922/examples-of-applied-micro-paper-with-r-code-and-data-in-public-repository
21:57 axfelix joined #dataverse
