First of all, it is no surprise that computer scientists would consider DNA as an excellent information store.  Treated correctly, DNA can be frozen and kept for potentially hundreds of years if I believe the magazine article which I scanned very quickly today. Also, there are 2 billion base pairs in the entire chromosomal content of one cell nucleus which are responsible for the encoding of roughly 21,000 genes to make a human being IIRC.

However, here is the bit where I need help. The article claimed that 9 base oligonucleotides could be used to code for short instructions and that 13 of these 9 base oligonucleotides could encode 13 trillion "words" of code (in the same way that sentences could be broken into words). 

I am assuming that I read this correctly.

Not knowing coding, can someone clarify:

a) What it means by words of code and

b) What technique would allow this code to be read quickly enough for practical purposes?

In the meantime, I will hunt for the original article...

https://en.m.wikipedia.org/wiki/Word_(computer_architecture)

Guessing based on the article linked below, the words are the 150 base long sequences

Quote

In the end, they had a DNA library with over 13 million unique DNA 150-base-long sequences. Each snippet starts with a coded address that shows its location in the file. And snippets that belong to the same file are flanked with the same “primer target,” a short DNA strand that is a kickoff point for the polymerase chain reaction.

https://spectrum.ieee.org/the-human-os/biomedical/devices/dna-data-storage-gets-random-access

Not sure about reading it back. Sounds like that is still being worked on.