Monday, March 25, 2013

[oawcxheq] Standard field and record separators

Create an alternative to stdin and stdout that allows both sides of a shell pipeline to unambiguously parse data separated into records and records into fields.  We need library support.

This is intended as a marginal improvement over the de facto but flawed method of lines and words separated by whitespace.

I guess CSV already exists, though is not standardized.  How to encode data with embedded field and record separators, i.e., a cell that contains a comma or a newline?

Avoid memory overflow on a record with an extremely large number of fields. This means that the number of fields in a record cannot easily be known, which is vaguely similar to buffer overflow with gets and fgets.

Many, many other alternatives, though most are much more heavyweight: http://en.wikipedia.org/wiki/Comparison_of_data_serialization_formats

No comments :