Outsourcing
Outsourcing is the process by which an organization hands over parts of the system development steps to another organization, though the source organization remains responsible for investigation, analysis and a few other steps. Reasons for outsourcing include: IT specialists lacking enough time or resources, the organization lacking the expertise, or it being cheaper to buy pre-written software than to build it.
The outsourcing process: problem definition and feasibility (always done by the source organization) → systems analysis, used as the foundation for a Request for Proposal (RFP) sent to possible vendors → evaluating the RFP returns and selecting a vendor → a legal contract stating the work, payments, timeframe and exit terms → testing and accepting the solution, retraining users and converting data → ongoing system support, which may be reassessed as part of the overall cost of the system.
Data Capture
Data capture refers to the process of getting data into a format that can be processed by a computer. In paper-based capture, people fill forms with information such as name, address or date of birth; in computerized data entry, answers are typed directly into the computer. Automated data capture devices include barcode readers, Magnetic Ink Character Readers (MICR), Optical Mark Readers (OMR) and Optical Character Recognition (OCR).
Types of Data Capture Errors
- Transcription errors: caused by misreading or mistyping data, e.g. confusing the number 5 with the letter S, or 0 with O.
- Transposition errors: occur when two digits or letters are swapped, e.g. ‘ot’ instead of ‘to’, or 1524 instead of 5214. About 70% of all errors are transposition errors.
Data Control
Data control refers to mechanisms implemented to ensure accurate and reliable data capture, through verification (checking that what is on the source document matches what has been entered, e.g. by double entry or proof reading) and validation (detecting data that is inaccurate, incomplete or unreasonable, which software can be set to reject). Common validation checks include:
| Check Type | Description |
|---|---|
| Character check | Makes sure the right type of characters have been entered. |
| Type check | Checks that the correct type of data has been entered in a field, e.g. number instead of text. |
| Check digits | A digit placed at the end of an original number, e.g. a bank account number. |
| Range check | Checks if a number lies within a specific range, e.g. ≥60. |
| Length check | Some data must always be of a certain length, e.g. an 8-character public service registration number. |
| Parity check | Makes sure data has not been corrupted during transmission (even number, even parity; odd number, odd parity). |
| Presence check | Checks whether a mandatory field has been completed – some fields are optional, others must contain data. |
