Software Abstractions Lit Review

Rasheed
 Wihaib
 –
 Software
 Abstractions
 and
 Systems
 Integration.
 
Week
 1
 
 

 
1.

What
 do
 the
 authors
 mean
 by
 “concept”,
 “conceptual
 model”,
 and
 “conceptual
 integrity”?
 Give
 
examples
 illustrating
 these
 definitions
 from
 a
 system
 different
 from
 those
 discussed
 in
 the
 paper.
 


 
Concepts
 are
 explicitly
 referred
 to
 on
 the
 first
 line
 of
 the
 second
 section,
 they
 are
 
said
 to
 mean
 the
 “constructs
 and
 notions
 either
 that
 the
 system
 deals
 with,
 …,
 or
 
that
 are
 invented
 for
 the
 purpose
 of
 structuring
 the
 functions
 of
 the
 system”.
 
They
 are
 said
 to
 have
 “psychological
 content”,
 and
 each
 concept
 will
 have
 
corresponding
 cognitive
 processes
 from
 the
 user
 when
 using
 thinking
 about
 the
 
application.
 A
 concept
 is
 the
 “Timeline”
 on
 Twitter,
 this
 structures
 some
 of
 the
 
functions
 of
 system
 such
 as
 displaying
 posts
 and
 interacting
 (retweeting,
 
favouriting,
 sharing)
 with
 posts.
 

 

Conceptual
 model
 refers
 to
 a
 specification
 that
 focuses
 on
 concepts
 rather
 than
 
details
 of
 behavior.
 Twitter’s
 conceptual
 model
 has,
 at
 it’s
 core,
 a
 micro
 blogging
 
platform
 with
 constraints
 on
 post
 lengths.
 Then
 there
 are
 interlinked
 concepts
 
such
 as
 follow’s
 and
 retweets
 that
 incorporate
 a
 social
 element
 into
 the
 system.
 
These
 are
 all
 concepts
 of
 the
 system
 and
 do
 not
 specify
 any
 particular
 behavior
 
on
 the
 part
 of
 the
 system.
 
 

 

Conceptual
 integrity
 encapsulates
 three
 principles;
 orthogonality,
 propriety
 and
 
generality.
 One
 could
 argue
 that
 Twitter
 upholds
 conceptual
 integrity
 as
 
following
 another
 user
 is
 the
 only
 way
 one
 can
 receive
 updates
 from
 that
 user
 on
 
their
 own
 timeline,
 demonstrating
 orthogonality.
 It’s
 minimalist
 user
 interface
 
and
 restrictions
 from
 its’
 design
 have
 also
 allowed
 for
 propriety.
 Generality
 can
 
also
 be
 found
 in
 the
 system
 as
 favouriting
 a
 tweet
 does
 not
 only
 allow
 a
 user
 to
 
save
 a
 tweet,
 it
 also
 could
 place
 the
 tweet
 in
 the
 ‘discover’
 news
 feed
 of
 their
 
followers.
 
 
 
 

 

2.

What
 is
 their
 main
 criticism
 of
 Git’s
 conceptual
 model?
 


 

The
 authors
 claim
 that
 “Git
 is
 far
 more
 complicated
 that
 it
 needs
 to
 be”,
 claiming
 
that
 the
 level
 of
 complexity
 is
 greater
 than
 the
 “limited
 functionality”.
 They
 
attribute
 these
 flaws
 to
 its
 conceptual
 model.
 In
 their
 graphical
 representation
 of
 
the
 conceptual
 model
 (figure
 1),
 they
 show
 some
 key
 concepts
 of
 Git
 such
 as
 a
 
“Tracked
 File”
 and
 a
 “Branch”.
 
 Their
 illustration
 of
 the
 relation
 between
 a
 File
 
and
 a
 Version
 demonstrates
 their
 criticism
 very
 clearly;
 this
 reveals,
 amongst
 
other
 things,
 over
 complication
 of
 the
 concept
 of
 file
 versions.
 
 

 

 

3.

Do
 you
 agree
 with
 their
 analysis?
 

Their
 argument
 is
 indeed
 compelling
 and
 as
 a
 Git
 user,
 I
 have
 suffered
 the
 same
 
frustrations
 when
 using
 the
 system
 so
 can
 vouch
 for
 much
 of
 their
 analysis.
 Their
 
criticisms
 with
 regards
 to
 the
 conceptual
 integrity
 and
 the
 conceptual
 model
 do
 
hold
 but
 they
 are
 not
 handled
 independently
 of
 each
 other.
 The
 criticism
 of
 the
 
conceptual
 integrity
 of
 Git
 is
 done
 in
 the
 analysis
 of
 the
 conceptual
 model,
 and
 
while
 a
 conceptual
 model
 is
 specified
 to
 not
 focus
 on
 details
 of
 behavior,
 the
 
authors
 comprehensively
 analyse
 the
 model
 by
 its
 behavior.
 Despite
 this
 causing
 
some
 ambiguity
 in
 their
 arguments,
 I
 generally
 agree
 with
 their
 analysis
 of
 Git.
 I
 
did
 feel
 some
 concepts,
 namely
 rebasing,
 have
 not
 been
 considered
 and,
 in
 the
 
case
 of
 rebasing,
 removes
 the
 need
 for
 their
 “more
 general
 branching
 concept”.