Having encountered the git executable issue before, at the July 2015 Davis workshop[1] we (1) put this in the install instructions (2) checked students' laptop on arrival and at lunch prior to the git session, and (3) briefed our helpers on this issue and how to catch and fix it quickly. This worked fine, as opposed to a previous workshop where we had to pause as about a third of learners needed a fix.
Yes, it does seem that the parallel approach has both conceptual mapping challenges and would require additional time. One aid to mapping would be to show how the RStudio GUI's visual state tracks changes executed in the command line (e.g., the checkbox next to a file will be checked when it is staged). For the reverse, one would want to type the equivalent command of an action executed in the GUI. [1]: http://dib-lab.github.io/2015-07-06-ucdavis/ On Wed, Mar 2, 2016 at 11:17 AM Kate Hertweck <[email protected]> wrote: > I have limited experience teaching Git to novice programmers using content > similar to that included in GitHub Guides (https://guides.github.com). I > also have taught the RStudio Git integration in a few workshops. Both of > these issues hit problems regarding additional installs (I've had problems > with students working with both the native Git GUI and finding a Git > executable through RStudio), which is pretty demoralizing to students when > it occurs mid-lesson. I'm also very hesitant to teach combined GUI/command > line exercises, because mapping the activities occurring in one to the > other is surprisingly difficult for learners. Then again, I also use git > almost exclusively from the command line. Despite this, I would be > interested in developing the RStudio git integration lesson so students can > follow along more easily on their own (perhaps following a workshop). > > On Wed, Mar 2, 2016 at 9:47 AM, Noam Ross <[email protected]> wrote: > >> I learned to use git via the RStudio GUI. It is not the most powerful >> interface, but it was enough for most tasks I needed as a beginner: >> committing, reviewing history, pushing/pulling from Github, etc. As I >> became comfortable with the concepts and had more advanced needs I switched >> to the command line. >> >> For R-based SWC/DC workshops I have taught and seen, we often teach git >> in the command line and then afterwards briefly demo the fact that similar >> tasks can be accomplished in the RStudio IDE, which learners already have >> installed. Has anyone attempted to teach with the GUI first, or somehow >> teach them in parallel? I hypothesize learners might be more likely to use >> git immediately with this approach. >> >> Two other thoughts on this approach: >> >> 1. While we aim to teach script-ability, scripting git commands is >> rare/advanced use. >> 2. I have gotten feedback that teaching the RStudio git GUI is hard to >> follow. This seemed to be because (1) this was a less well-developed >> lesson - more a demo at the end of the main git lesson, and (2) learners >> could not follow along via SWC notes or the live command history we shared >> via dropbox. So such lesson might require a screenshot-heavy set of >> accompanying lesson notes. Guidance might come from the DC experience with >> OpenRefine and Excel lessons. >> >> - Noam >> >> On Wed, Mar 2, 2016, 7:54 AM Konrad Hinsen <[email protected]> >> wrote: >> >>> On 02/03/16 10:30, Juan Nunez-Iglesias wrote: >>> >>> > I hope we all migrate soon to UIs built on top of git, such as gitless >>> > <http://gitless.com/>. But it's gonna be slow and painful, because of >>> > the huge momentum that git has. >>> >>> I considered using gitless instead of git for my recent course for >>> French PhD students >>> (https://github.com/khinsen/FdV-Computer-Aided-Research-2016). The two >>> main reasons I ended up sticking to plain git were >>> >>> 1) Gitless requires additional installation, whereas plain git was >>> already available on all machines. >>> >>> 2) There is excellent SWC teaching material for Git, but not for gitless. >>> >>> So, yes, SWC is becoming a part of the "system inertia" for me! >>> >>> A problem I see with gitless is that it is technically compatible with >>> git, but not operationally. For a given local repository, you use either >>> git or gitless. For everyone who knows some git commands, that means >>> unlearning. >>> >>> What I use myself is Magit (http://magit.vc/) within Emacs. It's a much >>> more reasonable UI for git, but it's fully compatible with git (it >>> actually calls git under the hood), so you can mix it with command line >>> work. But I wouldn't consider Magit for teaching because I am not >>> looking forward to doing "introduction to Emacs" first! >>> >>> Konrad. >>> >>> >>> >>> _______________________________________________ >>> Discuss mailing list >>> [email protected] >>> >>> http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org >>> >> >> _______________________________________________ >> Discuss mailing list >> [email protected] >> >> http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org >> > > > > -- > Kate L. Hertweck, Ph.D. > Assistant Professor, Department of Biology > The University of Texas at Tyler > 3900 University Blvd., Tyler, TX 75799 > Email: [email protected] > Office: HPR 109, 903.565.5882 > https://www.uttyler.edu/biology/ >
_______________________________________________ Discuss mailing list [email protected] http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org
