3.6 Errors

If it helps, remember that verification (ve) means making the data equal whereas validation (va) means checking if the data is acceptable. 

It is quite an interesting exercise to implement this in a student dossier (well, it might be).

The patient smoked 200 cigarettes per day and was employed as an airline steward (not).

Self test
Which of the following ISBN's are correct?

0138993947

0201634569

020130564X

On this page: [ data entry errors | error detection | error recovery ]

Data entry
Data entry errors typically occur in data processing systems such as billing systems, stock control, banking and payroll. Errors can be made when entering data at the keyboard or when reading data input from other devices.

Typical accidental errors are transposition and transcription . A transposition error occurs when digits are entered the wrong way round, such errors are very common : eg entering 316 instead of 3.6.1. A transcription error occurs when one digit is written for another: eg 3.8.1 instead of 3.6.1.

Deliberate errors will be the result of a security breach or fraud - a student may break into the examination database and change their grade or a bank employee may credit $450 of a $500 deposit but keep $50 for themselves. Usually, fraudsters are slightly more sophisticated than this.

Detecting errors
There are methods of checking for both these types of error. It is important to distinguish between validation and verification . 

Verification means making sure that the data on the source documents is exactly the same as that input to the computer system.

Validation means attempting to make sure that the data input into the system makes sense.

Verification
There are two main methods in common use; Visual verification (or proof-reading) which means checking data on screen to see if it is the same as on the source document.

Double entry verification which involves having the data from the source document entered twice as a check. Typically the application will lock the keyboard and give an audible warning so that the data entry person can check carefully.

One of these methods is more reliable than the other. 

Validation
Data entered into the computer can be unreasonable, a data entry person in a medical clinic can easily mistype a person's weight, say or the number of cigarettes they smoke per day. This type of error is not caught by verification since the data may have been written down wrongly in the first place. 

Validation also checks for incomplete or inaccurate data. The main types of validation checks are range checks, type checks and format checks. 

Range checks see if numeric data is in the correct range (eg a percentage mark needs to be between 0 and 100).

Type checks check that data is of appropriate type (integer, real or character for example).

Format checks are used to see if the data has the correct length or other internal structure (eg ddmmyyyy).

Find out more about these types of check - see your book or a good website.

Back to top

Hash and Control Totals compared:

Consider the following set of invoices:

Check Digits
A check digit is a digit (single number) calculated from another number, added to that number and entered into the computer as a validation check. For example, an ISBN check digit can be calculated as follows:

0 1 9 8 3 2 7 6 3 3

Strip the last digit (the check digit):

0 1 9 8 3 2 7 6 3

multiply each of the remaining digits by a column weighting starting at 10 on the left down to 2 on the right:

Add up the resulting sums: 0 + 9 + 72 etc = 217. This total is divided by 11 and the

integer remainder subtracted from 11 to give the check digit:

217 mod 11 = 8 (remainder), subtracted from 11 leaves 3 - the check digit.

If the check digit is 10, then the X symbol is used. 

Recovery from errors

Re-input
If data is input incorrectly and caught by verification or is not acceptable and is caught by validation, it can simply be re-input in an online process. If it is a batch process, normally the errors will be reported (printed out) and the batch process will continue with the next input. The entries in error can then be included with the next batch.

Re-transmission
Transmission errors caught by parity and/or checksums can be automatically corrected in some circumstances but more commonly there will be a request for a re-transmission of the data.

Backing up
For organisations whose very existence depends upon computer-stored data (banks among many others) data is protected very carefully. Backups are taken every day and, indeed, every transaction is backed up somewhere, at an ATM or a local branch for example. Backup media are typically stored in a fireproof safe or at a different location to the computer installation in case of a fire, flood or other disaster. Full backups save a new copy of every file, incremental backups save only those files which are new or have changed since the last backup. Tape is a common backup media because it relatively inexpensive and, hopefully, the backup data won't need to be accessed directly.

Back to top

related: [ Core home | previous: analog & digital | next: utilities ]

Try IBID Press if you haven't got a book...

Questions or problems related to this web site should be addressed to Richard Jones who asserts his right to be identified as the author and owner of these materials - unless otherwise indicated. Please feel free to use the material presented here and to create links to it for non-commercial purposes; an acknowledgement of the source is required by the Creative Commons licence. Use of materials from this site is conditional upon your having read the additional terms of use on the about page and the Creative Commons Licence. View privacy policy.