This is pretty cool. As a feature optional language restriction would be nice. My first random tweet was in French, second in Russian, third in Hebrew, fourth in German, fifth was same as second (same person/same tweet), and sixth was in English.
I actually think that result aligns perfectly with the creator's intent: "This is intended as a live demo that most people on twitter are not like you."
I can't tell if they are like me or not because I can't read the tweet in another language. It might be my french doppelganger also complaining about his current lack of caffeine.
I think you're taking only half of a comment and changing the context. "We often think that people use social networks the same way that we use social networks. This is intended as a live demo that most people on twitter are not like you."
I understand that you're using "not like you" in the context of what language they speak, but it seems pretty clear that the author is referring to "the way that we use social networks".
Apart from the already mentioned language restriction I'd be nice if the "Another!" button would stay at the same position so you don't have to move your mouse every time you want to load another one.
A UI critique: the "Another" button appears after the tweet. Most tweets vary in height. The result is that the button moves up and down after each press, which interrupts my ability to look at the tweet and requires me to hunt for it.
Luckily, I could use Cmd+R to refresh the page, which lets me focus on the tweets without UI distraction. But you may want to do something about the button, which is unusable for me.
Gives you a tweet in the US for you to categorize as being related to veing sick or not. Kind of like crowd sourcing classification of sick tweets. It's pretty interesting just looking through the tweets..
Yes I think that was mentioned in the copy. It's also not really "random." It's one of the last 20 tweets globally. You probably won't ever see a tweet older than a few minutes.
It's actually because Twitter rate limits the oembed endpoint by IP, so the app caches the oembed codes and uses those if it gets limited. I've been meaning to flip that bit to the client but haven't gotten around to it yet.
