These are what we call the 'YMMV items.' Things that some people find in this work. We call them 'your mileage might vary' because not everyone sees these things in the same way. This starts discussions in the trope lists, a thing we don't want. Please use the discussion page if you'd like to discuss any of these items.
Tween, a Japanese-developed Twitter client, is popular in its home country and is one of the top downloads on SourceForge.JP, but outside of Japan, it's obscure at best and dismissed as user-unfriendly at worst.
The upgrade to API 1.1 and the "Display Requirements" that came with it has not made some people, especially developers, happy.
Subtweeting, or replying without mentioning the user you're replying to. While some find it amusing, others find it exceptionally annoying, especially if one isn't following all participants of the conversation.
Dude, Not Funny!: Every so often, some users like to spread a rumor on Twitter that some celebrity has died when they're still very much alive.
Or even worse, joking about city council budget cuts.
After the Newtown shooting, one troll flat out crossed the line by pretending to be the shooter on Twitter (@Blastinkids) and saying he enjoyed killing them. His account is currently suspended.
Ensemble Darkhorse: Pretty much everyone is following one or two Horoscopes (usually and obviously, their own).
Fan Hater: There was once a "Ban One Direction Fans From Twitter" trend. Yes, because a particular fandom is totally a higher priority than issues such as harassment or bigotry.
Notably, "CALA BOCA GALVAO", a complaint against an obnoxious sports announcer that somehow became a campaign for saving endangered birds.
Japan is quite fond of Twitter as well; many users are big enough fans that they take features such as real-time tweet streaming very seriously. It also helps that the Japanese language allows squeezing in the same amount of information in less characters than in most Western languages, allowing one to practically write a short essay in 140 characters.
Spambots, who once in a while will randomly reply to one of your tweets with anywhere from offers for free electronics leading to a fradulent link to nothing but a link. They're usually suspended within a few minutes, not that it matters because new spam accounts are always pouring in. This is due to the fact that Twitter, owing to its roots as an SMS service, has no human verification measures during its registration process; you can start using your account even before verifying your email address.
Reddit favorite bots. They take the form of accounts with the username "Reddit<topic>", have their bio as "<topic> of the topic in Reddit", and will randomly favorite tweets with matching keywords. Most if not all users agree they're a nuisance and just report them for spam on sight, much like spambots.
Hype Aversion: Even people who have never used this website have had their view of the number/pound sign # changed forever.
Hype Backlash: As with any website with social networking elements, it gets quite a bit of this.
Twitter revised their blocking mechanism to allow someone a user has blocked to follow that user. This came under heavy fire, with millions of users complaining about the change in their tweets, to the point where Twitter reverted the change. However, it should be noted that although you can't follow blocked users, you can still see the tweets of public users even if they've blocked you by logging out.
The token limit on third-party clients has been a source of ire for many users. The popular Windows Twitter client Janetter, for example, can no longer accept new users because of this.
A lot of people are livid over the abuse report form outright saying that unless you're the one being harassed or you have legal right over that person, the issue won't be looked at. So essentially Bystander Syndrome is enforced.
The hashtag, popularized by Twitter, is often used even outside of Twitter.
sameExplanation Used when agreeing with something the user replies to or retweets. Sometimes it's used in reply to things that wouldn't make sense to agree with, on purpose.
SubtweetingExplanation Replying to a tweet, but not attaching the @-username of the tweet. This has not been well-received by some users, as it often creates needless clutter especially if some of the participants of the subtweet conversation is not being followed by the users seeing the subtweets.
Version 1.1 of its API. Among other things, it only allows clients to perform 15 API calls per 15 minutes (as opposed to the 150 or 360 calls per hour in 1.0), and all clients under 1.1 must conform to a series of "display requirements" that were not part of 1.0. This gives developers less freedom in how to design their clients, and users, particularly heavy users on mobile devices, will experience problems adapting to 1.1. Worse, every client is bound to a limit of 100,000 authentication tokens (or 100,000 users, but some people have more than one account) before requests must be made to Twitter staff for more tokens, effectively forcing a cap on the maximum number of users each client is permitted to serve.
As of July 2013, Twitter has backpedaled on some of the Display Requirements. Some of these changes were to reflect the existence of mobile Twitter clients; not everyone can fit a user's display name and @-username on their mobile device.
Putting your account into protected mode, which only allows users already following you and users whose follow requests you approve to read your tweets, is a useful way to preserve your privacy. However, it will not allow you to interact with users not following back, unless they send you follow requests or have a non-Twitter means of contacting them.
For the longest time, protected users' tweets would not show up in searches, rendering the ever-popular hashtag function near-useless to such users. However, an August 2013 update finally allows protected users' tweets to show up in their own searches and their followers' searches.
Direct messages can only be sent to people following you. While understandable in that it prevents users' DM inboxes from being filled with spam, it also means you can't have a private conversation with someone not following you (back) unless you have an external means to do so (e.g. instant messaging). However, an udpate in October 2013 allows users to, if they wish, enable DMs from anyone following them, though this comes with the risk of receiving spam messages.
Another October 2013 update brings inline image previews to web Twitter and official Twitter smartphone apps. This wouldn't be a problem if previews were optional; while a later update to the official Twitter apps allowed toggling them on or off, users of the Twitter website (rather than third-party clients) are stuck with the previews. What makes this feature a concern is that this feature can be abused to spam pornographic images. While some users are courteous enough to mark their accounts as "containing sensitive content", not everyone who posts such material either knows or cares.
Twitter's 100,000 token policy effectively killed off third-party clients for good, which was probably what it was intended to do.
They Changed It, Now It Sucks: The new retweeting system has been unpopular with the userbase, mostly since the retweets can no longer be edited to add personal comments. Of course, the old RT system still works just fine.