Re: GNU Guix Video Documentation

[Ricardo Wurmus <rekado@elephly.net>]

Hi Björn,

> I expected more like a screen cast: A life hacking session with
> comments on what's going on, why doing a step, opening a brower and
> show where to find docs. Of cause this has to be well prepared and some
> parts have to be cut out in order to be not too tedious (half an hour
> gcc log output during compilation is not so interesting for some
> people).

While screencasts can be useful, I don’t think they are the most useful
tool to convey ideas.  Much of what’s special about Guix is not the
command line user interface, but the underlying ideas.  These are better
illustrated, I think, with the help of graphics as we have been doing
for years when introducing Guix to new audiences.

One concern is also translations and future updates.  Recording a
terminal session with screencasting software makes it impossible for us
to easily translate the video.  When command line interactions are to be
shown I’d prefer to have a way to reproduce / regenerate the output in a
different locale automatically, i.e. using scripts.

We can easily mix what amounts to a narrated slideshow with scripted
command line sessions (cf asciicasts).  This can easily be automated, so
that we can rebuild the video and update it with minimal effort to
prevent it from getting stale.

Setting up this automation to some degree is part of the project, in my
opinion.  This could be done via scripts or tied together with a
Makefile.

The big advantages I see over a recorded live session are as follows:

- no need to get it all right in one take
- prepared narration can result in much more effective communication of
  ideas
- easy to rebuild
- easy to translate
- no need for manual video editing to cut out irrelevant parts
- separate treatment of audio and video portions; audio portions can be
  recorded by native speakers.

I really think that we should make an effort to keep the sources of the
videos and make sure that we can generate them from source without too
much effort.  To me this means that we should avoid traditional plain
video recording or screencasting.

> What I'm next missing is a set of tasks that we want to be
> video-documented. Or should that be part of the internship? I think it
> is more on our side to define the requirements.

Yes, I think we should provide more guidance here.  We should have a
minimum set of concepts that we’d like to see covered.  This also makes
it easier for us to determine if the internship progresses as planned.

Beyond the minimum, I’d like to leave it up to the intern to come up
with more ideas for videos.  It’s possible, though, that the minimum
already takes up more time than we currently anticipate, so lets just
focus on the first three videos.

> How and on which platform should the final result be presented?
> This includes freedom, privacy and technical aspects and is a
> question of acceptence/broadness of people watching it. Depending on the
> choice, could that affect the preparation process (for example,
> subtitle format, video-container format, audio-channels/translations,
> automatic text2subtitle conversation)?

We can host the videos on http://audio-video.gnu.org/ and embed them on
the Guix website, but the sources should be added to the guix-artwork
repository, I think.

The videos could also be published on Mediagoblin instances, but I don’t
know if there’s an instance for GNU packages.  GNU Guix does not
currently have its own Mediagoblin instance.

--
Ricardo