(Dav and Greg, you weren't as off-list as you might have thought you were.)
I find something like Google Docs where people make edit suggestions that someone else "merges" work quite well. For larger changes people quickly get the hang of leaving a comment with why they think it is a good change. It works very well for documents and kinda well for slides. Even with people who know git and GitHub! I've never investigated if there are any ways to enforce this workflow or later look at who did what. That never came up as a use-case. The only downside is that it is yet another tool I like using which is owned by the same single supplier. T On Mon, Oct 8, 2018 at 8:44 PM Dav Clark via discuss < [email protected]> wrote: > Thanks Greg... And know that I've been meaning to get in touch but I'm > just doing too damn much. Also, you getting a job at rstudio likely means > you're not available for the specific thing I was going to ask you about: > doing experiments using the activity records in Gigantum to diagnose > changes in learners' use of Jupyter (e.g., iterating on variations on the > "same" cell, using help, etc.). > > You've probably seen some of what I've written about Gigantum, but in > essence, it's a replacement for a lot of the management that you'd > traditionally do with a CLI: git, installing packages, starting jupyter, > etc. > > One thing that IS probably now of interest to you is that we're looking > into managing RStudio the same way right now, but due to the > less-open/extensible architecture, it's more of a struggle than I'd like > (specifically, we need to proxy activity and intercept it so that we can > construct an activity record, auto-git-commit and stuff like that). We > haven't reached out to RStudio folks more broadly, but now that I'm writing > I realize this is probably silly. But I welcome your input on this in any > case. > > As for the wiki thing - I think you're right that a wiki-ish thing is > probably about the right starting point, and your requirements seem good. > And, once we're talking about wikis, we're not too far from "cells" and I > wonder if perhaps a more humane replacement for git would finally let us > use Jupyter-like documents in a way that makes sense (e.g., having a > friendlier agent working with us on version control might finally make > Jupyter documents' JSON structure an asset as opposed to a > conflict-creating liability). > > Interestingly, your first requirement (all branches come off of master) is > built in to our simplified git tooling in Gigantum. I think it's a stretch > to implement this wiki-ish thing (for now) in Gigantum, but it still might > be productive to consider a standard git-ish workflow design across tools. > > Anyway, I'm probably rambling. Just got back from a wedding in Odessa, > Ukraine! Still quite tired / jet lagged. > > I'll keep you posted if I come across any solutions that seem to make > sense, and please do the same! > > Best, > Dav > > On Mon, Oct 8, 2018, 2:24 PM Greg Wilson <[email protected]> wrote: > >> [offlist] >> >> Hi Dav, >> >> I've been trying for years to find a wiki that supports what you want - >> my belief is that we just (for some value of "just") need: >> >> - every branch comes off master and is merged back to master >> >> - every branch allows multiple updates, and allows people to view/review >> the diff between its current state and master >> >> - every branch can contain changes to multiple files, but is atomic for >> commit >> >> This gives us what pull requests give us that commenting on Google Docs >> or writing in an etherpad doesn't, without the baggage of Git. If you know >> of something like this, I'd be grateful for a pointer. >> >> Cheers, >> >> Greg >> On 2018-10-08 2:10 p.m., Dav Clark via discuss wrote: >> >> I just had a couple of meeting with a few different folks, ranging from >> "mindfulness science" to "innovation in open science" to "history of >> science and religion". All of them have a need for something very much like >> a carpentry workshop: a focused skills boost that can get students and/or >> researchers up and running. >> >> I would love to have a workflow that mimics something like the GitHub / >> Bitbucket pull request workflow. BUT, I think wrapping your head around >> git + web services as a collaborative document production workflow is HARD >> (bordering on pathology). >> >> So, the question is whether anyone has found something that allows for a >> pull-request style multi-author workflow for prose, but that doesn't >> require the use of git. Ideally something that has actually worked for >> collaboration with at least some authors who have no understanding of git. >> >> I would love to see the methodologies developed for the Carpentries' >> curricula spreading out through different disciplines, spanning things like >> basic grammar for the humanities, transparent / reproducible / open >> scientific practices, etc. >> >> Many thanks! >> Dav >> >> *The Carpentries <https://carpentries.topicbox.com/latest>* / discuss / > see discussions <https://carpentries.topicbox.com/groups/discuss> + > participants <https://carpentries.topicbox.com/groups/discuss/members> + > delivery > options <https://carpentries.topicbox.com/groups/discuss/subscription> > Permalink > <https://carpentries.topicbox.com/groups/discuss/T95f755e418d1f2ac-M16fbdb111493c36df59642b4> > -- Let us be your hub hero https://hubhero.net ------------------------------------------ The Carpentries: discuss Permalink: https://carpentries.topicbox.com/groups/discuss/T95f755e418d1f2ac-Mf97dab5b61657911fd521943 Delivery options: https://carpentries.topicbox.com/groups/discuss/subscription
