Artist Names & Song Titles - from UPPERCASE to Title Case (and the evil StUdlycAse exceptions)

3:40p.m., Sun 20 Apr 2008

We're getting a feed of data direct from the Xfm radio playout system. For some reason (that has been lost in the midsts of time) all artist names and song titles are ALL UPPERCASE. (Although it could be worse, in times gone by they were inputted FERNETICKALLY). I guess when the data was first input into the playout system it was never envisaged that anyone other than radio schedulers and DJs wuld ever see it.

We've now started displaying data from the playout system on public-facing websites - we're allowing users to see what tracks have been played on air over the last seven days. But the UPPERCASING is just plain ugly.

The root and branch solution is to clean the data in the playout system. Which, it turns out, is a manual job.

So the next solution is to transform the data at a point somewhere between the playout system and the website.

Transforming from UPPERCASE to Title Casing is not without its problems.

But here, in plain english is how I’m going about it:

• lowercase everything
• Uppercase the initial character
• Uppercase characters that are preceded by a space
• Uppercase characters that are preceded by a hyphen
• Uppercase characters that are preceded by a space and an apostrophe (or the start of the string and an apostrophe)
• Uppercase characters that are preceded by an open bracket
• Uppercase characters that follow O' (a uppercase O and an apostrophe)
• Uppercase characters that follow “Mc” (an uppercase M and a lowercase c)
• Uppercase characters that precede AND are followed by a fullstop (for acronyms)
• Uppercase characters that are preceded by a space AND are followed by a fullstop (for initials)
• Tackle special case MC – uppercasing “Mc “ (uppercase M lowercase C and a space (or if uppercase M and lowercase C are the final two chars in the string))
• Tackle special case plural MCs
• Tackle special cases “DJ” and “DJs” (in the same way MC is tackled).

Looking at playlogs for a week or so of Xfm Overnights and trawling lists of artists that exist for some stations turned up some real world examples of wonkiness that might occur (they’re include in a list at the end of this post).

A few of the real world errors that have resulted in a manual error in inputting the data in the first place (The band name “Lights.action!” was missing a space and the song title “D.I.L.E.M.M.A” was missing it’s final full stop).

Generally speaking, song and album titles are a lot less problematic. The Beach Boys “SMiLE” is one album that comes to mind.

In the main, there’s two types of act name that cause problems:

• Acts which have an acronym for a name. REM, NWA, Run DMC* are technically acronyms, but the name of the act seems to have transcended that and the dot notation is sometimes dropped in common usage. (* = for the record, Rapid Eye Movement, Niggaz With Attitude and er, I’m not entirely sure but it’s acronym M.C. Darryl McDaneils adopted)
• Acts with names where the case of letters has been deliberately manipulated: JJ72, k.d. lang (sometimes with the fullstops, sometimes without), JJ72, and (most annoyingly of all) t.A.T.u.

The first one, generally speaking, is a result of dirty, lazy data. REM use the dot notation on their official site and they’re referred to as R.E.M. on Wikipedia. Same with Run D.M.C. (annoyingly, N.W.A don’t officially include the final dot). With editors entering data at ALL UPPERCASE omitting the dots is an easy error to make.

True exceptions are rare and I’d prefer to take these errors on the chin. It’s a relatively small number to live with. You could make a list of these exceptions, but it could become unmanageable. And the problem of dirty data is still likely to exist and be a bigger problem anyway.

I've also ignored the world of pain that is roman numerals. You could do simple checks for low numbers, but testing for all possible combinations of m c x v i... It’s gonna get complex. So I conveniently ignored it... sorry.

The Specialized Capitalizations that I’ve found

It’s by no means an exhaustive list, fuelled by artist lists from Xfm and Choice FM (hip-hop and urban), but I include it here because it may be of interest.

• ¡Forward, Russia!
• ...And You Will Know Us By The Trail Of Dead
• a-ha
• AC/DC
• A C Acoustics
• AFI
• AntiHero
• AM60
• aPAtT
• ARE Weapons
• bis
• bob hund
• B*Witched
• Boyz II Men
• C+C Music Factory
• Carter USM
• cLOUDDEAD
• CNN (who later renamed X-CNN)
• CSS
• The deBretts
• dEUS
• D4L
• DC Yeager
• DJ GQ
• DJ MA1
• DMX
• EMF
• FC/Kahuna
• fIREHOSE
• FR3E
• Future Pilot AKA
• iLiKETRAiNS
• INXS
• JJ72
• k.d. lang
• KLF
• KRS-One
• KT Tunstall
• LCD Soundsystem
• LFO
• LL Cool J
• Loudon Wainwright III
• MC5
• McLusky (who officially spell it "Mclusky")
• MGMT
• mu-Ziq
• N*E*R*D
• *NSYNC
• NoFX
• OK Go
• OutKast
• NLT
• N.W.A
• PD Syndicate
• PJ Harvey
• Shy FX
• Soul II Soul
• stellarstarr*
• t.A.T.u.
• Tenacious D
• TJ Cases
• The DeBretts
• TLC
• TSOL
• TQ
• TV on the Radio
• UB40
• VHS Or Beta
• will.i.am
• XTC
• ZZ Top

And the super-common 'typos' that slip thru the net

(these are the correct spellings)

• M.J. Cole
• N.W.A
• P.M. Dawn
• R.E.M.
• Run D.M.C.