The whole category in one page.
IDENTIFY THE LAYER BEFORE DIAGNOSING
Hardware, firmware, operating system, application, network. Where a symptom sits directs everything that follows.
Where the boot process stops tells you which layer failed.
THE QUESTIONS THAT ELIMINATE MOST POSSIBILITIES
What changed? One user or many? One device or many? One application or all? Wired as well as wireless?
Most faults follow a change, and testing with a different user account separates profile problems from application problems.
WHAT PEOPLE MISDIAGNOSE MOST
Slowness with several programs open is memory, not processor. Good performance that degrades is thermal. Unexplained restarts under load are the power supply. Software faults imitate hardware faults convincingly.
COPY THE DATA OFF A FAILING DRIVE BEFORE ANY FURTHER DIAGNOSIS
IN SUPPORT, SILENCE CAUSES MORE DISSATISFACTION THAN DELAY
Update even with no progress. Acknowledge impact before explaining. Never blame the user — it guarantees the next problem is concealed.
Log everything, including what you resolve immediately, or recurring problems stay invisible.
A RECURRING INCIDENT IS A PROBLEM TO FIX, NOT A TASK TO REPEAT
RECORD WHAT YOU HAVE, AND UPDATE IT AT THE TIME OF THE CHANGE
You cannot secure, maintain, budget for or recover what was never recorded — and records updated later are not updated.
