For sub-project 1 (code-name: kickitch — I finally got sick of not having a simpler way to refer to it in these blog posts!), I am working with datasets that are represented in different formats, including XML and tabular ones. A big difference between XML and tables is that XML is sparse (missing or n/a data is not represented by a placeholder), whereas tabular formats are dense (missing or n/a data is represented by default or null values).

So, in the query syntax I'm developing, I was planning on using the asterisk (*) to mean "include all fields". But, for XML formats, I need to account for two variants: #1 - include all elements even if they are empty and not required, and, #2 - include all that are non-empty or not otherwise required.

As an example, <generator> is an optional field in the channel section of an RSS 2.0 XML file. I have a value of "blojsom" in this blog's current RSS 2.0 file.

If I didn't have a generator value, the typical RSS/XML way to express this would be to leave out the <generator> tag altogether. In contrast, I would store a default value in a tabular representation (and someone else might store a NULL were it a SQL table—not me, but it's one option).

Because what I'm doing, I hope, will be generating different contexts for using XML formats like (but, not exactly) RSS, I can imagine cases where one would want to include empty tags rather than leave them out entirely.

Anyway, I'm thinking of using ** for #1 (all, show empties) and * for #2 (all, don't show empties unless they're required). Tabular formats won't distinguish between * and **, but XML formats will.

Let me know if you think ** is bad for some reason. I couldn't remember other programming contexts where ** was used, but I'm sure there are some—hope my use won't cause confusion. I wanted to use a single symbol, but ** made more sense (visually / semantically) than others that I experimented with.