Jim writes:
> I'm not sure the topic of default values is worth the time I'm about to
> devote to it.

Or the amount of time we'll spend arguing about it.

The problem today is that new Image features tend to default to being
disabled because of the fear of causing problems for people who don't
need the feature should it turn out to have bugs.  Unfortunately that
means that most people will *never* get the advantages these new features
provide, since many people don't pay attention to things like
communicator articles that tell them how great the new features are.

These features essentially get released as "experimental" features which
after a couple of releases become "certified" as probably being good for
all users to have enabled.

Today there is no difference between the default disabled condition of a
new feature and the user having explicitly said "I don't want this".  I
like Jeff's suggestion of the word "automatic" or something similar as
better terminology than "default" since it is less likely to cause this
kind of terminology based religious war.

An important question is who should be presumed to know more about
whether a particular option should be enabled or not.  The Image Lab who
wrote the database and understands the issues involved, or the customer
who probably doesn't understand the issue in the technical detail that
the lab does.  I would rather leave the Image Lab in charge until the
customer expresses a specific desire one way or the other.  Today the
lab is in charge of whether the feature is enabled or disabled by
default, but can't change their minds in the future.  Only the
customer can decide to enable something once the lab has decided to
disable it by default.

So the suggestion is that DBUTIL which now says:

  Use of Dependency SEMaphore(DSEM) is disabled.

Should say something like:

  Use of Dependency SEMaphore(DSEM) is unspecified and therefore disabled
  in this release of Image/SQL.

After a release or two, this would change to:

  Use of Dependency SEMaphore(DSEM) is unspecified and therefore enabled
  in this release of Image/SQL.

At any time the user could explicitly enable or disable the feature, which
would always override whatever the default for the particular version of
Image being used at the moment.  If the user doesn't specify a preference,
he gets what we think is good for him.  Once he specified an opinion on
the matter, that overrides everything else.

G.