Why isn't RAISERROR spelled RAISEERROR? Where is the second E? I could understand if it were some ancient keyword length constraint, but I wouldn't expect it to be a nine-character limit.
Is RAIS or RROR a technical word such that "raise-error" is just a mis-reading? Are its (immediate) origins in a different language?
I've searched Google but not finding much on the subject.
Interesting - I've honestly never even noticed that before.
I would suspect that it's an early oversight that was just never corrected, though I'm a little surprised that support wasn't later added for RAISEERROR, with the mis-spelling left alone for compatibility.
Update: Aparently, there's even some internal confusion about what it ought to be - Check out this connect request, though MSFT didn't respond to it.
why is there still a TIMESTAMP data type? support for old versions. Someone Sybase person long ago decided to make it RAISERROR and not RAISEERROR, when Microsoft made a deal with Sybase they continued the trend.
I believe it's intentional in that it is actually easier to read RAISERROR vs RAISEERROR. (see also sp_helprotect for another example), especially when there's no distinction based on case.
This stuff goes way back into Sybase's history.
This is just a coding standard used when TSQL was created. Any API needs a coding standard to ensure all developers who are writing items for the API do so in a constance manner.
At some point somebody made the choice if a word ends and the next word begins with the same letter to have it only once.
In the end it is code which spelling does not technicality matter. Be happy they didn't call it RE.