[Beowulf] User support systems

[Jim Lux]

At 11:54 AM 4/9/2008, Chris Dagdigian wrote:

>I'll second Joe's recommendation for Wiki's -- they have worked great
>both for conveying Admin-centric information as well as User-specific
>usage and application integration information.
>
>I'm still trying to weigh the utility-vs-effort of screen recordings
>that show users video of exactly how something is working or exactly
>how to do something. In general though I've found Camtasia 
>(http://www.techsmith.com/camtasia.asp ) is a great (if windows only 
>tool) that makes recording computer
>usage trivial. I can point Camtasia at my SSH terminal window and it
>will follow the mouse focus and even "smart zoom" in on the text I'm
>typing into the terminal. One click later and I have a nice flash
>video that I can embed in a wiki or email to someone. I've been
>running it on a Mac under Parallels and have had no issues.
>
>Video is great for some things -- think of illustrating exactly how
>the pxeboot -> DHCP -> tftp process all tightly play together in a
>network boot or server provisioning process. A recording of a serial
>console window while this happens is a good way to show someone at a
>basic level what is going on behind the scenes.



While video captures are nice, and fairly low in effort, I think that 
for true value, nothing beats a decent 
animation/demonstration/lecture.  It takes an inordinate amount of 
work to do this well, but the value is in the fact that things that 
are important can be shown, while things that aren't, aren't.

Taking your example of the boot process... imagine you have logs 
running on the DHCP server, the file server, and the client (or 
two).  Sure, you could display 4 windows (on a sufficiently large 
screen) in time synchrony (which, I might add, is non trivial to 
actually do), but you're also going to see a lot of extraneous 
detail, and the speed will vary.

Instead, if a lecturer took those screens and edited them, and used 
the ability  to slow/speed/freeze frame as the process unfolds, it 
has a lot more pedagogical value.




>The cut-off point on the effort-vs-reward is whenever I'm thinking
>about adding text callouts or titles to the videos. If that happens I
>know from experience I'm better off writing up a wiki page and
>inserting static screenshots.

Exactly... the live demo approach is great in terms of pedagogical 
value, but requires a LOT of work (especially with today's tools) to 
execute. At least by a subject matter expert.. a professional 
animator/video editor can zap this stuff out really fast from 
cues/ideas on the proverbial back of a napkin, after all, doing this 
is "what they do for a living".  So maybe the answer is to recruit 
some visual artists?



>  Video screen recordings are useful to me
>only when I have to put next-to-no effort into producing it. As soon
>things get complicated its best to fire up a good text editor.


That's because you're facile with the editor AND it's a medium that 
you find useful.  However, considered from the media consumer's 
viewpoint, the production methodology is immaterial (except as in 
inevitably affects the availability or not of the material in the first place).