Lecture 1 - Intro

Issues

• Why the velocity of a project may starts high and then slows to a grinding halt? When does it happen and how to prevent it?
• Why a project may start from a clean, "blank slate", and evolve to a mess?
• Is it desirable to write "elegant code"?

Inquiry: what is important?

There are 2 possible reasons for doing something:

1. it has bussines value:
   • implements desirable features (note: prioritization of the features);
   • keeps running time, needed memory, needed other resources, within acceptable limits (from customer perspective);
   • keeps bugs acceptable.
2. it improves the maintainability of the code:
   • keeps complexity under control — note: the complexity is the number one issue!
   • code is readable,
   • no surprises (least astonishment),
   • bugs are reasonably easy to track,
   • foreseeable changes can be implemented by changes in few places.

Goals and means subordinated to maintainability:

• understandability (readability):
  • single responsibility principle (SRP),
  • structure,
  • documentation
• testability: unit tests, integration tests, etc;
• code easy to use / integrate; well choosen API;
• extensibility
• simplicity
• code easy to reuse,

Notes:

• Enough is enough: as you keep adding features or increasing peformance, the cost per feature tends to increase, while the added benefit tends to decrease;
• When in doubt, experiment and measure: it's easy to be wrong in estimations of performance or of how good a design will be; the only sure way is to experiment and measure;
• A bug can hide another: buggy behaviour is hard to understand so that when a new (and maybe much more important) bug appears it is easy to notice, understand and fix.
  Radu-Lucian LUPŞA
  2018-10-03