Monday, January 4, 2010

Architecture Study Notes - Heuristics

The following are notes from the book "The Art of Systems Architecting"
By Mark W.Maier and Eberhardt Rechtin
  1. Do not assume that the original statement of the problem is necessarily the best, or even the right one.
  2. In partitioning choose the elements so that they are independant as
    possible; that is, choose elements with low external complexity and
    high internal complexity.
  3. The eye is the finest architect. Believe it. Simplify. Simplify. Simplify.
  4. Build in and maintain options as long as possible in the design and implementation of complex systems. You will need them.
For a heuristic to be selected, the following must be true.
  1. The heuristic must make sense in the original problem domain, and shoud
    show a direct effect between cause and effect; it should reflect the
    success or failure of a system.
  2. The heuristic should be useful in solving more than the original problem.
  3. The heuristic should be easily rationalized in less than 5 minutes.

Architecture process:-

  1. Orientation (More Business , less technical).
  2. Orientation is purpose analysis , problem structuring , solution structuring, harmonizing and selection-abstraction.
  3. Purpose analysis is a broad based study of why the capability or system of interest has value?
  4. Value Model - an explicit model of the most important stakeholder's preferences.

Also discover new heuristics from technical journals, books, project reports, management treatises and conversations.

Generating heuristics:-
  1. Humor in heuristic.
  2. Use words that transmit the :thrill of insight" to the eye of the holder.
  3. For maximum effect try descriptive and prescriptive messages.
  4. Do not create a heuristic where it only has meaning to its creator.
  5. Rather than creating conditional heuristic, rather create one, where it analyzes the conditional statement.
Applying Heuristic:-

  1. If it works, then it is useful.
  2. Knowing when and how is as important as what and why.
  3. When applied early to the solution space it is best.
  4. Too much of a good thing make make it worse, strive for a balance.
  5. Practice. Practice. Practice.


They provide the successive transitions from
qualitative, provisional needs to descriptive and prescriptive
guidelines, and thence to rational approaches and methods.