No Silver Bullet Fred Brooks

1
No Silver Bullet - Fred Brooks

• No single development in either technology or in
  management technique promises even one order of
  magnitude of improvement in productivity,
  reliability or simplicity
• Hard part of building software is the
  specification, design and testing of the
  conceptual constructs for a system, not the labor
  of representing it - essence/accidents

2
Four issues of difficulty

• Complexity - no two parts are alike very large
  number of states scaling up not merely a
  repetition of the same elements in larger sizes
  elements interact with each other in some
  non-linear fashion - complexity is an essential
  property.
• Descriptions of a software entity that abstract
  away its complexity often abstract away its
  essence

3

• Increased complexity causes increased
  communication problems. From the complexity
  comes the difficulty of enumerating, much less
  understanding, all the possible states of the
  program, and from that comes unreliability.
• From the complexity of function comes the
  difficulty of invoking function, which makes
  programs hard to use. From the complexity of
  structure comes the difficulty of extending
  programs to new functions without creating side
  effects.
• Management, communication and personnel turnover
  exacerbate all of the above.

4

• Conformity - to many human institutions and
  systems to which it interfaces conceived as most
  conformable
• Changeability - software changes very easily, but
  ramifications serious. Software is embedded in a
  cultural matrix of applications, users, laws and
  machine vehicles - all changing continually and
  their changes inexorably force change upon the
  software product.

5

• Invisibility - Software is invisible and
  unvisualizable. The reality of software is not
  inherently embedded in space. When we go to
  diagram we find several structures - flow of
  control, flow of data, patterns of dependency,
  time sequence and name - space relationships.
  Graphs are not planar, much less hierarchical
  (there is no good structure). Visualization is
  an important toolfor dealing with conceptual
  complexity.

6
Past breakthroughs solved accidents

• High level languages enhance representation
• time sharing - quick turnaround benefits
• unified programming environments - help us better
  manage conceptual constructs, but not figure out
  what they should be.

7
Hopes for the Silver

• Ada - encourages modern design and
  modularization, ADT, hierarchical structuring,
  etc
• OOP - helps with representation and other issues
• AI - the hard thing about building software is
  deciding what one wants to say, not saying it.

8

• Expert Systems - benefit is separation of
  application complexity from the program itself -
  where is the expert?
• Automatic programming - very domain dependent
• Graphical programming - when software is hard to
  visualize, how can this help?
• Program verification - works on code, but want to
  work on eliminating errors upstream.

9

• Environments and tools - hierarchical file
  systems, uniform file formats, generalized tools,
  SDE editors.
• LCAM - biggest gain yet to be realized from
  programming environments is the use of integrated
  database systems to keep track of the myriad of
  details that must be recalled accurately by the
  individual programmer and kept current for the
  group of collaborators on a single system.
• Workstations - speed compiling.

10
Promising Attacks

• Time of task S (frequency)i x (time)i
• attacks of the formulation of complex conceptual
  structures buy vs. build reqts refinement and
  rapid prototyping incremental development - grow
• conceptual constructs are too difficult to
  specify in advance and too complex to be built
  faultlessly
• great designers - NASA story cultivate