August 23: Happy Birthday, #hashtag!

Date

August 23, 2021

source

What would a philosophy of the hashtag look like? How do you feel about using # (pound) for groups. As in #barcamp [msg]? — Chris Messina, (“factoryjoe”), August 23, 2007 That’s the birth moment of the Twitter hashtag right there. . . .

What would a philosophy of the hashtag look like?

How do you feel about using # (pound) for groups. As in #barcamp [msg]? — Chris Messina, (“factoryjoe”), August 23, 2007

That’s the birth moment of the Twitter hashtag right there. The concept of tagging words with a hash sign was already common before. It was used in IRC to name channels, for example, and in HTML to jump to a particular place inside the same webpage. A nice history of the hashtag is here.

The hashtag is a curious thing. Its defining feature, the hash character, doesn’t even have a proper name. From England comes the name ‘pound sign,’ but the hash character is everything but. A pound sign looks like that: “£.” The only reason why someone called it ‘pound sign’ is that the British keyboard has its (proper) pound character at the same place (Shift-3) as the hash character. One of the inventors of the character apparently called it (God forbid) the ‘octothorpe’:

One of the first locations to be fully fitted with Bell’s new Touch Tone technology was the Mayo Clinic. When it was time to train the staff of the Mayo Clinic on using the new system, Bell’s Don McPherson coined the praise “octothorpe” to refer to the hashtag on the Touch Tone phones. It is said that McPherson came up with the name by combining the eight points “octo” and his favorite athlete, Olympian Jim Thorpe, to create a term that would be used throughout Bell Labs. – (Joe Fortunato)

Hash characters don’t have much of a meaning themselves. They modify the meaning of the word that comes after them. In this way, they are markup rather than content, similar to quote characters or the sequence “” in HTML. But, as opposed to those, their purpose is not to be displayed or to alter the way things are displayed. They are semantic markup, meant to group items of content into loose clusters.

Hash characters don’t have much of a meaning themselves. They modify the meaning of the word that comes after them. In this way, they are markup rather than content.

As markup, they are terrible to work with. First, what if you need a #multi #word #hashtag? The impossibility of escaping the space character is sure to infuriate every programmer who tries to use them efficiently.

Then, they are prone to small permutations that should not affect meaning. #Donald, #donald, #the #donald, #the-donald, #doland are probably all meant to refer to the same thing, but there’s quite a bit of algorithmic magic involved in finding out which similar-looking tags are actually referring to the same thing and which aren’t. Not to mention that completely different tags, like #trump and #potus, should also be understood to refer to the same real-world entity; except if we’re talking about #disney #donald #duck.

Since there’s only one level of markup and no hierarchical structure, it’s also impossible to distinguish #cloud #storage from #cloud #rain if the user only searches for #cloud.