Ultimately, the responses are very mixed, but the important thing is
that some of us can reliably get folks up and running quickly with a
consistent environment, even if they hit a snag with VirtualBox (which
is what we're using now for local installs). I need to be fairly
disciplined about this, as once folks are in the "trying to get software
to work on my laptop" mode, it's very hard to convince them to give up
(for the time being).
There's some nuance in doing all of this, though, and providing the same
setup can leave folks frustrated or very grateful depending on how it's
presented. I can't reliably communicate how to do that (and, honestly,
some of what I do may be mere "incantation"!). But part of what I
*think* works, is emphasizing that this is a full-featured, batteries
included environment (including some tricky installs, even for linux),
and that after we're done the training, we're happy to help configure
their laptop however they'd like.
Best,
D
On Sun, Oct 26, 2014 at 10:57 AM, Greg Wilson
<gvwil...@software-carpentry.org
<mailto:gvwil...@software-carpentry.org>> wrote:
Thanks for the background, Dav - how are Windows users reacting to
all this?
Cheers,
Greg
On 2014-10-26 1:48 PM, Dav Clark wrote:
As Aron suggested, we are actively considering the use of Docker
now at Berkeley. In particular, boot2docker 1.3 has made some nice
changes that allow for easier mapping from inside the inner
container all the way to the host level (though it's still not
there yet for port mapping). I apologize if I was part of the
telling you "to get lost" crowd. I do have a set of concerns that
are somewhat articulated above, but it boils down to the following
two use-cases:
1. Having an installation process that relies on familiar and
reliable GUI operations to get started with the new programming
environment.
2. Having a more-or-less already configured VM / account for some
users in "the cloud" (where, "the cloud" might be my development
server in D-Lab).
Docker may be useful for (2) now, but I think it's still not worth
the extra work beyond either setting up your own multi-user
server, or simply having a straight-up cloud VM (i.e., not
Dockerized, but rather an EC2 AMI or DigitalOcean droplet, etc.).
A lot of /instructors/ are not necessarily skilled sys-admin / ops
types (like me, for example ;). So even though we are insulating
our students in (2), we still should keep things simple for our
instructors.
But I agree that Docker is getting to where it's likely to be more
useful than difficult soon, and I certainly don't want to
discourage anyone from exploring how to make that easier! Perhaps
it's important to distinguish between development efforts and
current suggestions for provisioning student environments?
FWIW, the above-mentioned pioneer, Carl Boettiger, has been
submitting changes to the standard VM we're using at Berkeley to
show us the way forward with Docker:
https://github.com/dlab-berkeley/collaboratool/pull/90
We certainly welcome folks to contribute / fork / extend our work
there. We've made strong arguments about why it makes sense to use
Packer as the BASE configuration tool (gloss: it can provision
pretty much anything). But nothing is set in stone.
Shine on, you crazy Dockers,
D
On Sun, Oct 26, 2014 at 10:07 AM, Aron Ahmadia <a...@ahmadia.net
<mailto:a...@ahmadia.net>> wrote:
Hi Trevor and Greg,
My understanding from interacting with the tmpnb server [live
demo at https://tmpnb.orge] is that it is a zero-install
approach. The installation/maintenance load is reshouldered
onto system administrators, and the users can focus on getting
work done.
I think that assumption [1] There's no room in the
schedule..., should probably be [1A There's no room in the
schedule] OR [1B The learners will be working from this cloud
infrastructure in the future]. It looks to me that a number
of groups, from the D-Lab at Berkeley to JuliaBox at MIT, to
DIT4C at University of Melbourne, are moving at full steam on
providing this back-end support.
Does this make sense?
A
On Sun, Oct 26, 2014 at 4:31 PM, W. Trevor King
<wk...@tremily.us <mailto:wk...@tremily.us>> wrote:
On Sun, Oct 26, 2014 at 12:20:39PM -0400, Greg Wilson
wrote:
> 5. The learners are already familiar with Linux,
explicitly want to
> learn it, or it's an authentic task worthy of a lesson
in its own
> right.
Docker is trying to branch out beyond Linux. For example,
they're
working with Microsoft on a Docker engine [1] and native
client [2]
for Windows Server. Not that many of our students are
likely to show
up with the server-flavor installed on their laptop, but
still, Docker
is aiming to be a generic deployment framework like
existing virtual
machines, but without bundling kernels. I don't know
how many
scientists use Windows Server boxes for their research,
but I'm often
surprised by support for non-free OSes ;).
Cheers,
Trevor
[1]:
http://blog.docker.com/2014/10/docker-microsoft-partner-distributed-applications/
[2]:
http://azure.microsoft.com/blog/2014/10/15/new-windows-server-containers-and-azure-support-for-docker/
--
This email may be signed or encrypted with GnuPG
(http://www.gnupg.org).
For more information, see
http://en.wikipedia.org/wiki/Pretty_Good_Privacy
_______________________________________________
Discuss mailing list
Discuss@lists.software-carpentry.org
<mailto:Discuss@lists.software-carpentry.org>
http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org
_______________________________________________
Discuss mailing list
Discuss@lists.software-carpentry.org
<mailto:Discuss@lists.software-carpentry.org>
http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org
--
Dav Clark
Data Scientist
UC Berkeley D-Lab
dlab.berkeley.edu <http://dlab.berkeley.edu>
510-664-7000 <tel:510-664-7000>
_______________________________________________
Discuss mailing list
Discuss@lists.software-carpentry.org
<mailto:Discuss@lists.software-carpentry.org>
http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org
--
Greg Wilson
Software Carpentry |http://www.software-carpentry.org/
--
Dav Clark
Data Scientist
UC Berkeley D-Lab
dlab.berkeley.edu <http://dlab.berkeley.edu>
510-664-7000
_______________________________________________
Discuss mailing list
Discuss@lists.software-carpentry.org
http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org
