Re: [Usability] How to test

[Stefan Kost <ensonic hora-obscura de>]

hi,
Matthew Nuzum schrieb:
> On Tue, Oct 28, 2008 at 8:49 AM, Calum Benson <Calum Benson sun com> wrote:
>> If somebody could write a *nix version of something like Silverback[4], that
>> would be a great tool to have in our armoury, too.  (Slight mixed metaphor
>> there...)
>> [4] <http://silverbackapp.com/>
> 
> I didn't write a program like this but I did work out a solution. I
> used desktop sharing to another computer that used vncrec to get the
> screen capture and a more traditional camera recording program (can't
> remember which I used now). To sync the two I said the word "click" as
> I clicked something on the screen so that i could line up the videos
> afterward. It took forever to export the vncrec file to another video
> format (4 hours for a 10m session iirc) which could then be combined
> together using a typical video tool (I used windows movie maker for
> lack of a better option). Maybe you could run vncrec on the same
> computer that's sharing it's desktop.

Such an app could be written using gstreamer:
* rfbsrc - Connects to a VNC server and decodes RFB stream.
* v4l2src (or in cvs autovideosrc) - captue video from camera
* videomixer (or a special purpose element) - mix the streams side by side

Stefan

> 
> In hindsight I feel that most of my time creating video recordings was
> wasted effort. It may be useful if you're in a situation where "powers
> that be" don't trust your evaluation. But making meaningful use of the
> videos is very very time consuming. I feel much better about just
> making notes on paper as I watch the user. I had to write fast and my
> notes were hard to read afterwards but I didn't get much extra from
> the videos.
> 
> Maybe my pencil and paper technique is better for early stage testing
> where there's tons of low-hanging fruit and the videos are better for
> final testing where there is likely to be more contention and debate
> over the possible changes.
>