Character conversion

This type of conversion is used when the systems involved do not share the same character set.
The most common incompatibility is between a system using ASCII (PC, Unix, and many others) where the other one uses EBCDIC (mainly IBM mainframes).
However, this does not imply that a 1:1 correspondence exists, not even for characters.

When converting EBCDIC to ASCII, you will need to pay attention to the following things :
(1) are the character fields alphabetic only (A..Z) or have national characters been used, and
(2) is there any other type of data other then characters?
The problem with national characters stems from the fact that IBM allowed only 3 national characters, which replace the characters $,@ and #.
In Denmark, these characters are replaced by .
Normally this would not cause any problems, until we got a prime minister with the surname Schlter. I cannot remember how I solved this problem on the IBM 360/40.
But even ASCII files have their own problems. In old times, one could have national characters in the three characters following Z : [ \ ] resp { | } for upper/lower case.
In Denmark we used these places for and , which fits the bill when we need to sort a file alphabetically.
This would not work for Germans, who used them as replacement for and . the problem being that these characters are to be sorted "together" with A U and O.

The other main problem is the existence of what IBM calls COMP-3 fields (packed decimal). These fields pack 2 decimal digits in one byte.
On top of that, the last half byte contains the sign.
So, a COMP-3 value of +0012345 would be represented by X'0012345C', the C saying that this is a positive value. In case of a negative value, the C would be replaced with a D.

In order to handle these non-character fields, they must be converted to character-only, e.g. through the Record Reformatter.

A very different problem, which annot be solved by the RecordReformatter, is the conversion of files originating from Word Processors, as there is hardly any correspondence between how they define attributes (Italic, Underlined, font size, etc.)
Conversion of this kind is a job for specialists, and even then no guarantee can be given that everything is converted correctly, as documentation for old formats is about just as scarce as hens teeth. However, the basic text can always be retrieved.
Front page