I have a proposal in place to replace the single referrer with an array of all preceding referrers, to track through multiple redirects, etc. The header name would be "referer's"
So true. I'm occasionally confused on which spelling to use. My rule of thumb is using "referrer" for almost all place including variable names, except for the situation when I have to manipulate the headers directly.
i've finally got my brain wrapped around to believing that referer is the correct spelling within the context of HTTP. Any time i need to talk about a referer, it's spelled referer. if it's spelled referrer, it must mean something else.
Reminds me when I have to pronounce router the network thing vs router the woodworking tool. The former is pronounced the USA way, the latter the English way.
If I've learnt anything in programming, it's that inconsistent naming just adds useless complexity. At some point, you'll find yourself having to map from 'referrer' to 'referer'; that's entirely wasted effort. Wherever possible, name everything (variables, array indices, filenames, database fields, HTML input names) exactly the same.
Presumably any such fix would have to start with specifying both a referer and a referrer header, so that legacy clients wouldn't freak out. That said, it's humbling to think that because someone made a mistake in, what, 1991? We still have to go around spelling referrer wrong all the time, and even include the incorrect spelling in new specs.
So, all you standards proliferators out there: used the goddam spell checker, all right?
"[D]ocument co-author Roy Fielding has remarked that neither 'referrer' nor the misspelling 'referer' were recognized by the standard Unix spell checker of the period."
I would go a step further and say the IETF is dead ... so many of the new protocols are basically just after-the-fact corporate interests fronting up and publishing.
