The practice of indicating type when one names something is an old practice, and still a great idea. Some say that it’s not necessary if one uses a modern IDE that shows type on hover, etc. But have you seen any IDE that does that for interpreted languages like JS? And what about when you have to look at code outside of an IDE, such as in an email or in a diff?

I don’t know of any industry-wide consensus on what characters to use for common types, but my (ever evolving) set is:

x – unknown type
i – int
o – object
b – bool
s – string
f – function (especially helpful when passing function pointers and refs)
sb – StringBuilder
e – DOM element
as – array of strings
ao – array of objects
hs – string of html
zs, zhs – “sanitized” string; stripped of possible cross-site scripting attacks
ht – hashtable (used in JS when an obj is used just to collect properties)

Similarly, it’s very helpful to know what kind of result a function returns. I indicate the return type after the “f”, like this:

fbSubmitCompose – returns a bool
fsbSubmitButtonHTML – returns a StringBuilder of markup
fPreparePagesDropdown – doesn’t return a value

btw, perhaps fhsbSubmitButton would be better.

If it’s a member var, prepend with “m_” and if it’s a static member, with “s_”. This is very important, because if you ever work with a web framework that creates just one object per class (like Java servlets), you want to be very mindful of what’s stored in member and static vars, since some uses lead to loss of thread safety. It also helps when reading a long function to know which of the objects are defined outside.