I don't have very tidy handwriting, and for most things this is fine: there's enough redundancy in language that it will be clear from context what is a t and what is a +. When writing calculations, however, there's often a lot less redundancy and so it's more worth it to make sure distinctions aren't being lost. One way to handle this would be to have clean handwriting overall, but what I do instead is use glyphs that are more robust and remain distinct even if written poorly. [1]

Here are what I do with some tricky glyphs:

i j t + l 1 I 2 z 7 ;

It's nice if these are clear to other people, but the main goal is for them not to be confused with each other. Things to help this:

• i includes a right hook
• j is normal
• t includes a right hook
• + is normal
• l is loopy/cursive
• 1 is a straight line
• I has top and bottom lines
• 2 is normal
• z has a cross
• 7 has a cross

This lets me write thing like:

(zt + 2ij)/7Il

and still be able to read back what I wrote. This is also something I need when transcribing whiteboard coding interviews at work: [2]

for (int j = i; j < t + z; j++)

You might notice that I don't use this for the i or t in int. It's just for variables.

[1] I don't think this is original to me, but it's also something I picked up gradually as I realized that certain forms were more reliable.

[2] Yes, candidates write code on whiteboards, we transcribe it on paper, and then we type it into a computer for others to review. Someday candidates will write on laptops, I guess, but it was weird that we were doing it in 2010, and we're still doing it in 2018, so who knows.