Research talk:Ideas/Example

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search

Virtual collaboration space: To wiki or not to wiki?[edit]

Okay, sure, seems like a non-question. Given who we are and what we're doing, the most obvious solution is to use a wiki as a central hub, and link to off-wiki tools like data and code repos, google groups, research group webspaces, etc from there. BUT many Wikipedia researchers aren't super-comfortable contributing within the classic Wikimedia MediaWiki environment (i.e. here). So wiki-based workflow could present a substantial barrier to entry for any but the usual suspects. Also means that if we wiki, the burden of upkeep will probably fall on those who are most comfortable. If they don't take up the mop and keep it at hand, it could impede the success of the project. Jtmorgan (talk) 08:25, 10 August 2013 (UTC)

A media wiki instance or space is probably fine for the moment (especially as long as we don't expect everyone to contribute equally). Also, my intuition is that it will be important to make it easy to coordinate light-weight collaborations (e.g. meeting planning, sharing readings) somewhat separately from the more involved sorts of collaboration involved in designing, executing, and publishing projects. Between those two, I think we just need to solve the easy one (lightweight stuff) for now. Aaronshaw (talk) 08:50, 10 August 2013 (UTC)
I agree. I think that our best strategy right now is to put together the most lightweight system that will serve our needs, but remain open to the idea that it may become apparent that we need something more. One of the things I'm most concerned about is keeping us in a space where non-academics (and non-wmf) will feel comfortable participating. Keeping the primary space on a wikimedia wiki is a good way to keep that barrier low. --EpochFail (talk) 01:59, 11 August 2013 (UTC)

Where to host?[edit]

Where do we host our virtual collaboration space? On Wikimedia servers? On the servers of some well-funded university research group? Again, may seem like an obvious answer, but wanted to raise it anyway since the choice we make will impact who participates and how. Jtmorgan (talk) 08:25, 10 August 2013 (UTC)

WMF Labs hosting should be an option. And I like this overall idea! 00:59, 11 August 2013 (UTC)
I'd like to suggest that our first coordination spaces live within Wikimedia wikis (like here on meta) and mailing lists (we may want to add another to the wikimedia set). However, when it comes to analysis resources, I'd like to suggest that we try to take advantage of WMF labs for two reasons:
  1. WMF labs has direct access to a set of slave databases that replicate Wikimedia projects
  2. It's "free" (as in beer) and resizable resource since it is a cluster from which we can request additional resources when necessary
There's also a separate hadoop cluster at the WMF that's under heavy development and may be available for some more CPU intensive analyses. --EpochFail (talk) 02:15, 11 August 2013 (UTC)

Research mentorship and community involvement[edit]

I'm posting this as a topic for discussion rather than a project goal because it's a potentially-significant expansion of Epochfail's original scope: what about encouraging community members involvement in performing and writing research, rather than just providing inputs and consuming findings? That allows the professional researchers to fill a mentorship role in some cases, and learn from community members in others. Could also help teach generally-applicable methods and research thinking to interested community members who are new to the research biz. Jtmorgan (talk) 09:38, 10 August 2013 (UTC)

Why hasn't wiki-research-l filled this role?[edit]

moved from Grants:IdeaLab/Labs2#Why_hasn.27t_wiki-research-l_filled_this_role.3F Jtmorgan (talk) 00:19, 11 September 2013 (UTC)

If there's really such a missed opportunity for collaboration, why has it not already taken place on wiki-research-l, the mailing list for Wiki researchers? Answering this question will be essential to the success of this project. --EpochFail (talk) 01:42, 10 August 2013 (UTC)

I'm not sure that a mailing list on its own is centralized or persistent enough to sustain the kind of collaboration we're trying to foster. We probably need additional infrastructure. Also every project needs maintainers to rouse the rabble, get out the vote, and clean up afterwards. Maybe no one is vested enough in the research-l list to consistently play that role. Jtmorgan (talk) 08:36, 10 August 2013 (UTC)
+1 - I think this project is awesome and my hunch is that to do better than a mailing list, with both on-wiki and offline components, it will take folks who have some time to community organize (rabble rouse, cheerlead, orchestrate, and cleanup) in order to make it persistently awesome. I'm curious whether the folks on the participants list here have this time? Siko (WMF) (talk) 16:48, 26 August 2013 (UTC)
Yep, so if we're going to make this happen, best to scope out some specific roles and responsibilities. And to do that effectively, we probably need an actionable, near-term goal, like putting on a particular event or collaborating around a specific deliverable with an actual due date. (me/ not volunteering, just bikeshedding.. for now) Jtmorgan (talk) 00:30, 28 August 2013 (UTC)
Although a big portion of our daily communications are now happening online, but the need to physical contact and in person conversations in order to share and peruse research ideas is undeniable. Although the wiki-research-l mailing list has been playing an important rule in bridging different researchers and projects at the level of Q/A and literature sharing, but more focused and intensive projects with certain goals need a more coherent venue of communication supported by a well-designed infrastructure. Taha Yasseri (talk) 08:38, 10 August 2013 (UTC)

In some ways I find w-r-l too narrowly focused, in other ways too general. It also focuses a lot on slow-developing projects or results, rather than things that can be tested, evaluated, and repeated over days and weeks.

  • Narrowness: it's mainly about wikimedia. Despite the more general wiki-title. And not necessarily about the set of interesting things noone understands; but a discussion of the set of things that are currently already being researched. Wikimedia projects share lots of policies, narrative, thought-patterns. I'm interested that there isn't 10x more work done on the cases where active intervention in wikis, or community sites of any kind, are known to have a significant short-term and persistent impact.
  • Broadness: it addresses research ideas, and thought-experiments, more than specific data-sets and historical features. The latter we could focus on as a collective, with hundreds of views and approaches, all trying to converge on a shared understanding. More like evaluating a large public scientific data set in any other science.
  • Speed: some social science work requires buy-in, private evaluation, private analysis, and finally some public discussion. In contrast, some of the best data-driven social discussions on wikis are ones that are driven by a few different folks running their own analysis/stats on shared data, comparing results, sharing draft visuals, ID'ing and correcting classes of errors. The slow kind may support academic discourse, but doesn't really make it easy to either chip in a few days' cranking-time (if you're interested sporadically) or to try to reproduce the work on your own.

A good data-analysis project, that encouraged public analysis, shared datasets, and separated ientification of features of interest from evaluating and understanding them, would address many of these gaps. SJ talk  00:33, 14 September 2013 (UTC)

wiki data vs. WikiData[edit]

Re: "for enabling better collaboration around wiki data in the future" Can we make "wiki data" more clear? Because it could be heard as WikiData (which differs only in spelling). Jodi.a.schneider (talk) 14:32, 23 September 2013 (UTC)

"involved in conducting research of research on Wikipedia, wikis, and other kinds of open collaborative communities"[edit]

What about "conducting research on and doing analysis of"? This might be more open to community members (who are explicitly included, but who might not identify as people "conducting research"). Jodi.a.schneider (talk) 14:39, 23 September 2013 (UTC)