Newsgroups: rec.games.int-fiction
Path: nntp.gmd.de!news.ruhr-uni-bochum.de!news.rhrz.uni-bonn.de!RRZ.Uni-Koeln.DE!news.duesseldorf.ecrc.net!news.ecrc.de!02-newsfeed.univie.ac.at!newsfeed.sunet.se!news00.sunet.se!sunic!news.sprintlink.net!news-stk-200.sprintlink.net!tank.news.pipex.net!pipex!nntp.primenet.com!uunet!in3.uu.net!xcs.contex.com!marvin.contex.com!liz
From: liz@marvin.contex.com (Elizabeth Cyr Jones)
Subject: Re: Playtesting (was Re: I-F, QUIPS AND BLOOPERS)
Message-ID: <Dv054B.6r4@contex.com>
Sender: usenet@contex.com (News Owner)
Nntp-Posting-Host: marvin.contex.com
Organization: Xyvision Design Systems, Wakefield, MA
References: <4s1hjd$72l@newsbf02.news.aol.com> <glt9_5O00WB5EPshwP@andrew.cmu.edu> <4s3ppd$c6t@news.multiverse.com> <4sded2$j4f@kodak.rdcs.kodak.com>
Date: Tue, 23 Jul 1996 15:16:55 GMT
Lines: 59

In article <4sded2$j4f@kodak.rdcs.kodak.com>,
Matthew Daly <daly@PPD.Kodak.COM> wrote:
[lots stuff snipped]
>
>I don't think that anyone outside Infocom uses the term gamma testing.
>I never took part in any of their gamma tests, but I presume that
>as a result of it they would have beefed up their vocabulary and
>remarks, and may even have changed puzzles around.  I don't see how
>any of this could happen AFTER beta-testing, or at least a second
>round of beta would have to come after gamma.  Unless they did
>alpha-gamma-beta, but why get the alphabet out of order?
>
>-Matthew
>--
>Matthew Daly             I don't buy everything I read ... I haven't
>daly@ppd.kodak.com       even read everything I've bought.
>
>My opinions are not necessarily those of my employer, of course.

The game development schedule at Infocom was determined based on the needs
of the games coming into testing:  some were more difficult, bigger geography,
more platforms, etc..  We used alpha, beta, and gamma, in that order, though
would vary the length of time required for each stage.  And, for some stickier
games, had more than one beta and one gamma.

Alpha happened pretty much when the Imp wrote the last lines of code, and the
game could play through without crashing.  We considered that sort of a
prototype of the game, and put it into testing where usually only one person 
would work it.  Alpha wrapped up when the game became stable enough not to 
crash even when fooling around with objects ('n stuff).  

Beta meant distributing the game to our outside testers.  Testing maintained a
database of a couple hundred people, and we usually used around 20 per round.
Beta and gamma testers were chosen based on the skill and hardware, as often
certain platforms were available before others.  We expected feedback from
beta testers regarding the difficulty of puzzles, and potential alternate
solutions. Beta wrapped up when the puzzles were solidly in place, and the
game "made sense".

Gamma was supposedly the cleanup round where we depended on people to add
vocabulary, syntax, little bells and whistles.  Certainly not major stuff as
though we liked to think that development worked outside of marketing/sales,
their needs helped define the schedule.  (And the doc guys were running around,
and the feelies were being picked out).

So, that was pretty much the schedule, except for the exceptions (aren't their
always?)  Bureaucracy got kicked out of beta and had to start over again,
months later, for example.  

Don't know how many other companies imitate this model, but the graphic design
company I work for now uses successive beta periods, and no gamma.  

My how I've gone on, oh well, my two cents.


Liz Jones
former Testing Mgr, Infocom


