definition of "anti-feature"

[bill-auger]

the definition of "feature", added to the summary of the
anti-features' article, did not really clarify the issue raised
in the "Political messaging" thread - i presume the concept of a
software "feature", to be sufficiently well-understood by most
people - it is nearly synonymous with the common notion of
"behavior" - i believe that the "easter-egg" example is
difficult to capture, due to the definition of "anti-feature"

an anti-feature is not a feature to any user - "anti-feature"
should be defined strictly as: "any software functionality added
at the whim of the publisher, which has no obvious utility to
users of the software" - "anti-feature" carries no presumption
of malicious behavior, only that it is useless to the user (or
is generally not "what it says on the tin")

i believe that greg was suggesting, that an easter-egg is not
malicious and requires no warning - however any easter-egg is,
by its nature, an anti-feature, because it serves no purpose
related to any task, which the user is likely to be trying to
accomplish with that tool - there are many other common harmless
anti-features, such as "please donate" buttons, which people
generally tolerate

an easter-egg is a special case of anti-feature, which was
intended to be useless (eg: a joke) - if the program were a game
or other entertainment software, then an easter-egg could be
considered to be a legitimate feature (eg: if the user is
entertained by it)

per the wiki, that is why they are "undesirable from the user's
perspective" - more accurately: "all users' perspectives" - ie:
"this feature is plainly not applicable to the software's
primary problem domain"

if _any_ user of the software finds a legitimate use for some
functionality, within the primary task domain of the tool, then
it is not an "anti-feature" - that is a legitimate "feature",
which is not universally understood or appreciated
(controversial perhaps, but controversy notwithstanding)

the concept of "anti-feature" is also relevant to the FSDG, and
is defined loosely in that context also - i think it would
be most helpful to define "anti-feature" more explicitly and
uniformly for the FSD and FSDG, and to offer some guidance as to
which classes of anti-feature are acceptable (harmless
functionality), which should carry a warning (privacy-intrusive
functionality which some people may value), and which should be
removed entirely (malicious functionality)

i suggest a re-wording of the summary:

---

- The Institute of Electrical and Electronics Engineers defines
  the term feature in IEEE 829 as   "A distinguishing
  characteristic of a software item (e.g., performance,
  portability, or functionality)".
+ [DELETE]

---

- Antifeatures are flags applied to applications to warn of
  issues that may be undesirable from the user's perspective.
+ Anti-features are software behaviors, which were added at the
  whim of the publisher, with no obvious utility to users of the
  software, within the software's primary task domain.

---

- Frequently it is behavior that benefits the developer, but
  that the end user of the software would prefer not to be there.
+ The behavior is often invisible and/or unknown to the user.
  Frequently, it is behavior which benefits the developer, but is
  orthogonal to or inapplicable to the software's assumed or
  advertised task domain.
+ An "easter-egg" exemplifies a harmless anti-feature;
  while a virus exemplifies a malicious anti-feature.
+ "Spyware" is another common type, which lies somewhere along
  that spectrum.
+ Another: "Automatic geo-location", exemplifies the subjective
  cases, of functionality in which some users may value, others
  may not, and yet others may find objectionable or inappropriate
  for the task at hand.