Consider improvements to Tablo syntax

[brendanberg]

Open Questions

The current Tablo specification seems to work well for simple data sets, but there are a handful of extended features that would suggest additional syntax.

• Cell Range syntax: should a standalone column identifier indicate the entire column as a range, eliminating redundant syntax, or should all ranges require a colon separated pair? (i.e. A vs. A:A)

• Formula syntax: a formula should be enclosed in braces (i.e. { Math.sum(A1:A12) }) but what about formulas that get applied to each item in a range? Can we eliminate the need for relative cell references?

• Table Metadata: the Tablo format does not support (arbitrary, unstructured) comments, per se, but a syntax could be developed for annotating cells, rows, columns, or ranges.

• Inline Formatting: can individual cells have format declarations? [No.]

[brendanberg]

Syntax options for adding key-value annotations to format syntax:

[A0] { bold, "comment": "could you explain this?", "author": "Anne F."}

[A0] format { bold }
[A0] comment { "text": "could you explain this?", "author": "Anne F.", "user_id": "1059235", "timestamp": #2022-06-22T03:34:03Z }

(Format entries are a separate line from objects that are attached to cells)

[brendanberg]

Cell specifiers don't need brackets:

A:A {bold, red}

This paves the way for conditional formatting expressions where cell ranges are terms in an expression, which tests each element in the range and applies the format if the expression evaluates truthfully.

E:E > 500 {bold, red}
500 >= E:E > 400 {orange}

The property section is a list of keyed tags. Tags are text labels ([A-Za-z][A-Za-z0-9_-]*) and may either appear as a bare keyword or be immediately followed by an associated value. Values are standard tablo types, plus a dictionary type (curly braces enclosing a list of tags / tag-value pairs).

A0 {comment {
    text "could you expand on this?",
    author "Anne F.",
    user_id "1059235",
    timestamp #2022-06-22T03:34:03Z 
}}