[sldev] Performance measurement variable: possible immersion (was: More crash rate statistics)

[Boroondas Gupte]

Alex Dailey (Xan Linden) wrote:
> Here's a challenge to sldev. Viewer performance is complex, dependent on 
> bandwidth, hardware, the content on the region, other avatars, and so 
> on. What do you think a) is the one variable viewer performance should 
> ultimately be measured by and b) what are, say, three predictive 
> variables that influence the dependent variable.

Define "performance" and I'll tell you how to measure it. Or vice versa. ;-)

Using Allen's the non-technical definition cited on wikipedia
<http://en.wikipedia.org/w/index.php?title=Computer_performance&oldid=293171037#cite_ref-0>

    /The word/ performance /in computer performance means the same thing
    that performance means in other contexts, that is, it means "How
    well is the computer doing the work it is supposed to do?"
    /

the question remains what the viewer is supposed to do, which varies
wildly from use case to use case. I have the theory that most variables
that affect at least one of the use cases also affect the use case where
the user wants to maximize immersion. (By which I mean the type of
immersion one also has when reading a book, not the one achieved by HMDs
and cyber gloves. Note that the prior is still affected by technical
issues: If you read a book with missing pages, hard to read font or a
cover that smells funny (and the smell isn't related to the story at
all) your ability to immerse will undoubtedly suffer.)

Of course, potential immersion has the problem of being poorly defined,
too, and not directly measurable. You'd have to do surveys or something,
and if you did that you could just directly ask about user experience
instead. However I think it's an interesting variable for theoretic
considerations:

Note that the influence of most variables on immersion (and user
experiance in general) is highly nonlinear. Let's take update rates (FPS
if we look at graphics). Whether FPS are at 0.5 or 2, I don't mind. In
both cases I see single frames instead of continuous motion.Whether
they're at 100 or 1000 is also irrelevant, you can't tell the difference
with a 60Hz TFT. Though, for the interval between 3 and maybe 30, even
slight differences are noticeable enough to matter. This goes for server
issues, too: Whether 1% or 10% of group IMs get delivered ... who cares,
group IM wouldn't be a reliable medium in either case. The difference
between 81% and 90% or even between 90.9% and 99.9%, although nominally
also 9%, would be a big one, though.

Also, the influence of different variables is often interdependent. Some
problems might mask other problems, for example.

Ann Otoole wrote:
> User experience is tied more to the sim performance than viewer
> performance.
As someone using SL from a low end system I often hardly notice poor Sim
performance, because my viewer is struggling in crowded areas anyway.

Both, the masking effect and the nonlinearity of the influences, could
probably be used to optimize the experience by automatically balancing
them. The most simple variant would be to automatically decrease
graphics settings when the FPS hits a lower threshold (say, 5) and to
increase them again when hitting a higher threshold (say, the monitor's
refresh rate). (Having to manually counteract immersion breaking effects
is something that itself hurts immersion.)

This already got rather lengthy, so I'll stop for now.
cheers
Boroondas
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.secondlife.com/pipermail/sldev/attachments/20090808/dea75d09/attachment.htm