One letter variable names

From Red Blob Games's Blog

I admit I use one letter variable names in my code. But which names? Here's what I use in for loops:

# rg 'for \(let . ?=' 
#     | perl -pe 's/.*let //g' | perl -pe 's/=.*//g' 
#     | sort | uniq -c | sort -nr | head

509 i
132 x
122 y
118 r
 57 t
 53 q
 47 j
 43 s
 31 e
 14 k

My naming conventions:

• i, j, k: generic index
• x, y, z, w: coordinates
• q, r: col/row grid coordinates
• r: region number
• r: range or radius
• t: triangle number
• t: time index
• s, e: sides/edges of a polygon
• w, h: width/height

Outside of for loops, rg 'let . ?= | grep -v for' finds that I also use these conventions:

• a, b: endpoints of a range or line segment
• a, b, c: equation for a line, or quadratic formula
• v: generic value
• p: point
• t: time
• d: distance
• d: reference to svg <path> d= attribute
• g: reference to svg <g> element

If I were working in a team I'd follow their naming conventions, but for the most part I am writing the code for myself, so I'm ok with short variable names like these. The common objection is that I won't understand the code six months later, but I haven't found that to be an issue. These are all local variables inside a function, and I tend to use the same conventions across projects.

[2024-12-16] Also see A Dictionary of Single-Letter Variable Names for Haskell code.

Labels: code