CB2500 Chap.7 Hardware, Software and Database Processing
Hardware, Software and Database Processing
Three layers with very different lifespans
Week four of the published schedule puts hardware and software beside database processing and database design, and the grouping rewards being read as one layered decision. Hardware and software choices are reversible inside a budget cycle. A data structure, once several programs depend on it, is the most expensive thing an organisation owns to change.
That asymmetry is why a business course spends time on database design at all, and why this chapter spends most of its attention on the bottom layer.
What a managed store replaces
Before shared databases each function kept its own file and each program carried its own description of that file. The arrangement is not impossible, it simply does not survive change.
The same customer exists in several files and the copies drift; a change to one field is a change to every program reading it; and the function owning a file controls who may see it. The third is the one business students should notice, because it turns every request for a fact into a negotiation and makes a company slow for reasons that have nothing to do with its technology.
Consolidation removes all three at once, which is the earlier repository argument applied to the organisation rather than to one process.
Design is deciding what each table is a table of
Database design here is not about syntax. The test is that a table should describe exactly one subject, and it can be applied in a closed book assessment.
When two subjects share a table, facts about one repeat once for every row of the other, and a value stored twice will eventually be stored differently.
The repair is a routine: name the subject of each table in one noun, choose the identifier before the attributes, move every attribute depending on something other than that identifier, and keep the identifier in both tables so they can be rejoined.
An identifier encodes a business rule
Choosing what picks out one row looks technical and is usually a policy question in disguise.
Deciding that a booking is identified by room and night decides that two people cannot hold the same room that night. Deciding to identify a patient by an identity document number decides how someone without one is treated. When a case describes a recurring data problem, that buried rule is frequently where it started.
What this chapter covers
- 01
Reversible hardware choices against durable data structures
- 02
Private files and the three faults they carry
- 03
Ownership of a file as an organisational constraint
- 04
One table, one subject
- 05
Repeated cells as the visible symptom
- 06
Choosing an identifier before the attributes
- 07
Keeping the link so tables can be rejoined
- 08
What an identifier decides about the business
Repair a table that a process has already corrupted
- 3Name the subjects and show the table carries more than one.
- 3Apply the identifier test to each subject.
- 4Give the split and the link, and say what it fixes.
Key terms
- Managed Store
- A single shared collection of records maintained for the whole organisation rather than privately by each function.
- Identifier
- The value that picks out exactly one row of a table, and which encodes what the organisation treats as one thing.
- Subject Of A Table
- The single kind of thing a table describes, expressible as one noun without the word and.
- Repeated Value
- The same fact stored on more than one row, which is where two versions of one truth begin.
- Link Attribute
- An identifier carried in a second table so that the two can be rejoined after a split.
- Program Dependence
- The condition in which a description of stored data lives inside each program, so a field change forces every program to change.
Hardware, Software and Database Processing FAQ
Why does an information management course teach table design?
Because the structure outlives every program that reads it and every manager who approved it. Choosing badly is a cost the organisation pays for years, and the choice is not technical: it encodes what the business treats as one customer, one booking or one patient. That decision belongs to people who understand the business.
How do I spot a badly shaped table in an assessment?
Look for cells whose value repeats down the page. A customer name appearing on every order line is a fact about the customer being stored once per order, and it is the visible form of two subjects sharing one table. Then apply the identifier test, because a table where no single value picks out one row is describing something that has not been defined yet.
What should the project data file look like?
One row per collected post or comment, carrying a post identifier, the date, the source page and the text. The identifier matters because the analysis has to be traceable back to the specific items that produced each finding, and without it a report cannot show where a topic came from. The submission accepts the common spreadsheet and delimited formats.
Exam move
Take any spreadsheet you already keep and find the column whose values repeat. Write the subject of the sheet as a single noun; if you need the word and, split it on paper and write the link attribute. Do this three times with different sheets, because the pattern recognition is what transfers to an unseen case.
Then write, for one identifier you chose, the business rule it silently imposes, which is the part of this topic examiners find easiest to test and students find hardest to see.
Working through Hardware, Software and Database Processing in CB2500? Sia is AskSia’s AI Information Technology tutor — ask any CB2500 Hardware, Software and Database Processing question and get a clear, step-by-step explanation grounded in how CB2500 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
