Twitter has a great habit of introducing changes nobody wants (increasing the character limit to 280), and not acting on the things we’ve all been asking for for ages, such as an edit button, and maybe, just maybe … not allowing misogynistic, antisemitic and racist abuse to thrive?

It was a great development when Twitter introduced inline gifs and videos, but not so great when Jewish women routinely receive messages such as “go back to the gas chambers” and it somehow doesn’t violate their terms of service. Some 2.6 million antisemitic tweets were sent last year. One woman was even suspended for reporting a tweet that told her: “Welcome to Trump’s America. Welcome to the camps!”

Rose McGowan, however, the actor at the forefront of calling out Harvey Weinstein’s alleged decades of sexual harassment and abuse – just one of the women whose stories are only now coming out because of the aggressively patriarchal culture of fear that has rendered them silent – has been temporarily banned from the platform on which she had found her voice.

Good one, Twitter. Well done, lads. Speculation was that her crime might have been calling out Ben Affleck, tweeting, “Ben Affleck fuck off” (although not at him; not including his Twitter handle) and accusing him of dishonesty over what he knew about Weinstein’s alleged behaviour. (Affleck later tweeted an apology after it was pointed out he had groped the actor Hilarie Burton, and footage of him surfaced commenting on a television presenter’s breasts.)

I’d love to know how McGowan tweeting “Ben Affleck fuck off” is enough to get McGowan suspended for 12 hours almost immediately but, say, private messaging someone racist epithets or open neo-Nazi calls to war don’t contravene Twitter’s terms of service. Or – wait for it – threatening nuclear war. If McGowan suggesting that Affleck should fuck off sees her kicked off the platform, then why didn’t Donald Trump calling to obliterate an entire nation count as aggression? That is surely the actual apex of aggression.

Ditto, if McGowan’s accusing Affleck of lying is an issue, then how come Trump continuously calling people liars (usually female journalists in every other tweet, Hillary Clinton often) is fine? Trump on more than one occasion tweeting antisemitic memes is OK. Trump tweeting misogynistic slurs at women is OK. But someone calling Trump “a moron” is suspended. It is mind-boggling. (Twitter did offer a brief explanation of Trump’s nuclear war tweet-storm not violating its terms of service, which is that it was newsworthy. And a woman at the heart of the biggest Hollywood scandal perhaps, ever, isn’t newsworthy?)

It is possible that whichever specific tweet triggered the 12-hour suspension has been automatically removed (as is often the case) – and Twitter has since said it was down to the posting of a phone number – but it seems hugely over the top and counterproductive to take McGowan’s account offline at a critical time of debate. (It’s also possible that the initial decision to suspend was an algorithmic decision, but it’s hardly likely it will have passed Twitter by).

I haven’t spoken much about this, but for nine months a couple of years ago, my life was made a living hell by a group who objected to a report I wrote about an enterprise of theirs that was scamming people after the Charlie Hebdo terrorist attacks. I was hounded relentlessly in every conceivable way online. Some of the methods are more known about now, but back then they were not.

Fake news sites – populated with otherwise real content and respectable-looking domain names – were set up with hoax stories about me being corrupt, being mentally unstable, and even being an ex-porn star. The SEO was gamed so that they appeared at the top of Google searches for my name. Mass emails were sent out to journalists across the industry, warning people against me. Social media accounts were combed for personal information about my family and friends to spook me. But perhaps the most consistently disruptive aspect was this one: this group flooded my Twitter account with bots, and also bot-retweeted every tweet I sent. This made it impossible to use a medium that I relied on as part of my job. As well as this, my friends and colleagues were soon also targeted.

Now, Twitter didn’t do nothing about this. I was sent the link to a special form to speed up my harassment flags. However, I still believe that this was escalated only when a far higher-profile colleague who the group realised was an acquaintance of mine was targeted.

The abuse only stopped after months of sleuthing when I unmasked the culprit – some loser in Miami along with subcontracted Russian bot farms. This is all stuff that we are becoming more aware of now, but in 2014, it was an intensely lonely experience. If making someone’s life a misery in this way does not violate Twitter’s terms of service then clearly there is something not working about Twitter’s terms of service. Even though merely banning the main account of the group that I knew was responsible wouldn’t have stopped the harassment, because the abuse was so complex, it would have been a deterrent. Twitter didn’t do this, because no rules had been broken. That proves that the rules themselves are broken.

But actually, some of the rules as they are currently defined had been broken. Because this account also tweeted at a senior colleague: “fucking bitch”. Which I also reported. Nothing. She also reported the bot activity on her account (apparently an attempt to blame it on me and get me fired), and also: nothing.

There are plenty of other examples. It took Twitter days to kick Azealia Banks off the platform after she called Zayn Malik a racial epithet. And here’s a funny thing about that: I emailed Twitter about how, how!, none of this had broken its terms of services. I had a piece lined up about it. And then … I actually asked my editor not to run it. Because I knew a piece about Twitter abuse would garner a lot of Twitter abuse, and I just couldn’t handle it. That’s how much Twitter needs to get it together.

When months of abuse, when violent threats to female politicians, when the online version of hissing at Jewish people, or making monkey noises at people of colour, is apparently all acceptable but a victim of sexual abuse leading one of the most important and most open conversations on a horrific, systemic issue is silenced, even if only temporarily, then it almost makes me want to leave Twitter altogether.

It’s no wonder prospective buyers such as Disney backed off bidding for the platform because of its reputation on abuse. Former CEO Dick Costolo once said Twitter “sucks” at dealing with abuse, and despite some steps in the right direction (quality filters and mute buttons, for instance ), McGowan’s suspension shows that nothing’s really changed.

• Hannah Jane Parkinson is a Guardian columnist