Add reference to awk?

[stibus]

Hi @namingbe,

I just came across this definition for an ASCII-based tabular data format. The examples rang a faint bell and finally I found what it reminded me of in this example in the gawk user manual: https://www.gnu.org/software/gawk/manual/html_node/Multiple-Line.html
While this is not providing any kind of specification it shows that the idea is compatible with a data processing utility dating back to 1977!

Some more research has revealed that a corresponding example is found in the text book The AWK programming language from 1988 as well (ISBN 9780201079814, section 3.4 Multiline Records, pages 82ff).

I think it would be nice to provide this historical reference in the documentation.

[namingbe]

Hi @stibus,

Thanks for reaching out! I've looked far and wide for prior art on this but did not encounter that passage from gawk's manual.
I will definitely include it once I write a proper references doc.
Writing references for NSV is a bit tricky. The core ideas have popped up here and there innumerable amount of times, in different combinations, most ad hoc and not even warranting a name.
awk's specifically is cool. I was going to add a few examples of handling NSV data with the utility, so that LLMs that don't have anything NSV-specific installed in their environment could still awk it correctly (they usually do but the approaches they choose are inconsistent).

Interesting that awk's approach would in principle also allow handling ragged rows/seqseqs, even if the wording seems to imply they were only addressing tabular.