Software is never finished

A shop launches an app. It works well. Everyone is happy.

Six months later, a new law says every food product must show its allergy information. The app must change.

A month after that, a bad file damages part of the customer list. The app must be repaired.

Software needs work after it is released. That work is called maintenance.

Key words

Word	What it means	Example
Maintenance	Work on software after it is released, to keep it working and useful.	A fix, an update
Preventive maintenance	Change made before a problem happens, because we can see it coming.	Update for a new law
Corrective maintenance	Change made to repair an error that nobody expected.	Repair damaged data
Regulation	A rule from the government or an official body.	A law on allergy labels
Business process	The steps a company follows to do its work.	How an order is checked and sent
Zero-day	A weak place that attackers find before the maker knows about it.	A new fault in a popular tool
Targeted attack	An attack on one person or one company.	A thief attacks one bank only
Data corruption	Data that is damaged or wrong.	An age of -5

The cells students fill in are left blank in this preview.

Eight reasons for maintenance

Group	Reason to change	Example
Prevent: we can see it coming	A change to a regulation	A new law about data or labels. The software must follow it
Prevent	Compatibility with new products or technology	A new phone system arrives. The app must work on it
Prevent	A change in the business process	The company starts to send parcels from a second warehouse
Prevent	The release of a new product or service	The shop adds a gift card. The software must handle it
Correct: nobody expected it	A new weakness because another product changed (zero-day)	A tool the app uses has a new fault that attackers use at once
Correct	A targeted attack	A thief attacks one company and gets into its site
Correct	Data corruption	A bad update turns some ages into words
Correct	A system failure	The server stops and orders are lost

The cells students fill in are left blank in this preview.

Why "zero-day"?

data-callouts.important.label

The maker has had zero days to fix the weak place. Attackers know about it first. This is why these updates are urgent.

Try it: spot damaged data

Code: python

rows = [(1, 17), (2, "abc"), (3, -5), (4, 20)]

for customer, age in rows:
    good = isinstance(age, int) and 0 <= age <= 120
    print(customer, age, "good" if good else "DAMAGED")

# 0 <= age <= 120 means: age is from 0 to 120.
# Add a row (5, None). Is it good or damaged?