Found in Stowe Boyd’s blog:
Hashtags (Twitter tags) were proposed by Chris Messina, and in use by Chris, me, and others before tools existed to do much with them, aside from search. In similar fashion, I am proposing a new sort of microstructure, just a little bit ahead of tools to support it.
The idea is similar to tags: use a distinctive character to set off some microstructured metadata, although in this case, the metadata is location, and the character is ‘/’, the slash.
He gives as examples:
Just landed at /JFK
hanging at /Starbucks, 93 Greenwich Ave, NYC/
What it means: the same way that hashtags (#) are helping to structure and organize concepts on Twitter and other social networks, Boyd proposes to create a new “standard” around the use of / in short-form messages to help organize the information around “location”. As he discloses, “I write this post with a ulterior motives, since I am involved in the development of a Twitter appliance, called Thweres, that will exploit location information of the form proposed here” So, I’m really curious to see what Boyd is up to and if this proposal takes off.
Seb Provencher's Blog 
Intéressant – je vais commencer à l’appliquer mais je me demande s’il n’y a pas un impact, justement, dans search.twitter.com
Si j’indique /JFK/ est-ce que Search va quand même me sortir ce Tweet si ma recherche se fait uniquement par JFK?
En d’autres mots, est-ce que je vais devoir faire deux recherches ( JFK et /JFK/ pour voit tout ce qui se dit sur JFK?
Et ça serait /JFK ? ou /JFK/
comme quoi j’ai déjà plein de questions 😉
Présentement, si tu cherches dans search.twitter.com pour /jfjk ou pour jfk, le moteur de recherche enlève le /. Donc, les deux résultats de recherche sont identiques. On peut penser que, si ce standard décolle, Twitter va créer un local search engine qui va permettre de retrouver spécifiquement cette info.
I think the idea is great – a location microformat for Twitter (et al). I think using the slash is a bad idea, though. The slash is used for too many other reasons to be trusted as a marker of a location. Many multi-author Twitter accounts, for example,use a slash and initials to denote the author.
A better character would be ^ since it exists on every standard keyboard but is not in common use.
^Cambridge, ON^
I agree with @Colin, the slash in used a lot already in URLs.
http://microformats.org/wiki/picoformats has some intersting suggestions, the obvious @ is used by convention now to refer to people.
The most sensible choice seems to be “L:” as per http://microformats.org/wiki/microblogging-nanoformats but it’s not “natural” and breaks the flow a bit for humans. Maybe using braces could do [placename] or {placename}.
Another easy option is this one: I am at *Praized HQ. I know it’s harder to parse for developers, since we don’t know explicitely where the string ends, but that’s not so hard to do.
Just as a side-note, this is our rough plan for adding location information to Laconica version 0.9.x:
http://mail.laconi.ca/pipermail/laconica-dev/2009-April/001225.html
Ah, but the * is often used for emphasis/bold… I’m sticking with ^location^ … 😉
Wouldn’t something like “>” or “->” (i mean, like an arrow…) be nice?
@Guillaume, I really like the >
We can even use the reverse if we want a closing marker.
I am at >Praized HQ
I am at >Zenya Sushi on Ste-Catherine
I’m in >Montreal< today and in >Barcelona< next week
I like it too, except that it wouldn’t work outside of a plain-text-only environment like Twitter… anywhere that HTML is allowed (blog comments, for example) it could get confused with HTML tags (links, in particular).
We’ve already seen the @ convention migrate out of Twitter and into blogs, and we would expect (and hope) to see the same with this, no?
woeid? yahoo placemaker
I agree that a slash isn’t ideal and the L: standard would work better. However, since location-based updates would typically be made via mobile phone and not a computer, I’d see L: as a temporary fix for people without geo-location-enabled phones. Ideally that information would be meta-data associated with the message, and not take up characters within the message itself.
I think we need to make it easy and natural for people to add the information in a way that doesn’t break the flow. As such, I think a special character like *, ! or ^ is better than L: — which just seems a bit TOO much like an annotation. The problem is that no single character (other than @) really has any natural association with place.
And I’ve decided I don’t like the > character because I think it’s meaning is ambiguous. It feels more like a sentence constructor or separator as in “thing a > thing b” where > tends to be read as ‘thing a leads to thing b’. It is already used by some people in this way.
So, I think [place] is the most natural, human and machine readable way to express it. It is immediately obvious how it is being used – which I think is important for adoption.
“Just landed at [JFK]”
“Grabbing a coffee with @friend at [Starbucks on Whyte].”
I dont think the slash is the best idea as it would get confused with other uses for a slash. yes/no/maybe? = my location is no. Forcing us to be mindful of how we ALREADY use, or misuse, a slash so as not to get mistaken for location tags will not work.
Other special characters are problematic too if they are not common on cell phone keypads.