Attention, internets: in English, we read left to right, top to bottom. This affects how we interpret static visual media generally. And this applies to emoticons in ways you do not appear to understand.
What I'm about to say may only apply to those of us with a Roman or Cyrillic-alphabet-based written first language, but for now, that's all I care about. So for instance, if you are reading this, chances are that if you see a sign with a single word written vertically, you probably don't start at the bottom and try to read to the top. Because even though the data cannot be interpreted left-to-right, reading bottom-to-top would violate the left to right, top to bottom template that becomes automatic within the first few months of learning to read. So you have no problem interpreting the Chicago Theater sign, even though there are no visual clues as to how it should be read--i.e., there is no capital letter signifying a starting point, etc. In short, when left-to-right does not apply, your brain switches to top-to-bottom to interpret the text.
It follows that if left-to-right seems to apply, but makes no sense, the English-reading mind's next logical step is to interpret left-to-right data as if it were top-to-bottom. So, if you are like me, and you are reading a line of left-to-right text, and you encounter a collection of marks that does not make sense left-to-right, e.g. the following emoticon: :-) your mind attempts to interpret it as top-to-bottom data (the reverse also works when top-to-bottom makes no sense, e.g. when a vertically-stored book has its title printed lengthwise along the binding). Anyway, for the above emoticon, you probably see a pair of eyes, a nose-like thing, and a smiley-face--as opposed to say, a baseball cap bill pulled down to the nose of a person with double upper-lip piercings. In other words, would this (-: make sense if we didn't already have :-) already cached in our pictographic lexicons? I say no, because it requires the English-language reader to stop, drop everything he knows about his written language, skip ahead a few characters to what he hopes is not an arbitrary point, read backwards for a bit, then skip ahead further and resume reading left-to-right.
And for precisely that reason, this symbol/emoticon is a problem: <3. People of the internet, because you are injecting <3 into English-language phrases, it must be read left-to-right/top-to-bottom. You are using it to signify 'heart' or, by extension, 'love', but it is not a heart! The heart shape has two rounded bumps on top and a point at the bottom. According to the the left-to-right, top-to-bottom template, <3 has a point at the top and two bumps on the bottom. So this symbol either signifies 'less than 3' or (interpreting it as if it were top-to-bottom data, because 'less than 3' does not make sense in context) represents a shape that is narrow at the top and widens into two rounded things at the bottom.
To restate, because of the way the English language--and the English-language-reading mind--work, <3 cannot possibly represent 'heart' or 'love'. Instead, it must signify either 'less than 3' or the nearest shape in everyday life that it resembles. And there aren't a great deal of things out there that look like that.
In other words, to put it bluntly, when you type 'x <3 y', if any English-language-reader reads it as anything other than 'x tea-bag(s) y', then something has gone wrong.
♥ = Heart. But <3 = Tea-bag. Sorry I had to be the one to break it to ya. Twitter/facebook accordingly.