Post by kurochan in itch browse tag exclude comments

Viewing post in itch browse tag exclude comments

I’m not planning on reading that “never show me this again button” thread since that would likely tell me stuff I already realized about this topic, but if I had to give a possible user-friendly implementation of that type of button….

I’d have to say, the most user-friendly implementation of a “don’t show me __ tag(s)” type of filter I’ve ever seen in an online store platform is Steam’s “Exclude from my store” menu section’s “Tags to exclude” feature where the individual user can define up to 10 tags that they want the Steam store to globally exclude results that contain the tags selected.

since you mentioned it is known that the actual exclusion processing is happening on the server-side, all the itch.io devs would need to do is implement a quick validation that checks against the array of user-specified tags to exclude before serving the results to the front-end. That isn’t very difficult to pull off. I’m sure even a Junior Developer should have the skills to create this feature so there’s no valid reason why any of the veteran devs on the itch dev team couldn’t pull this off as well.

BTW, sorry if that sounds like a rant… that was me in programming geek mode writing with no filter (pun not intended lolz) and I really got excited about the idea of creating my own version of this feature just for fun… my bad! 😅

redonihunter17 days ago

It is not known why there is no multiple tag exclusion. A plausible explanation would be, that it would cause undue server load. If there is a ddos, complex tag browse queries are the first to not work. Like, browse for 2-3 tags, and it work. Browse for 4 or more tags, and the server does not respond. A hint is also, that the 1 tag exclusion only works for users that are logged in. They way Itch's system works under the hood probably cannot deal with the concept of exclusion. Not efficiently at least.

Steam has an ignore button for games and they do have a persitent exclude feature, yes. But they also have a maturity setting with like 4 different shades and only a hundred tags or so, and those tags are rather accurate.

On Itch there are probably 10000 tags. Some with only 1 game. And most tags are inaccurate or redunant. Excluding things on tag basis is just not a concept that works well on Itch, even if there were a ready tool. A user might not get what was expected.

Most people probably just want to declutter the front pages. And because there is so much horror and visual novels, they think that exluding tags is the solution. But the never show me this again button works wonders there and is more flexible too. It is basically a custom css that hides individual games and you can add games to that list by the press of a button.

You can hide the games you already know or are in your library with it. I really should rewrite the page and create a screenshot for it. But I have trouble showing things that are not there.

That hide button script is so much more powerful than mere tag exclusion - and if you really want to, you can simulate a multiple tag exclusion with it. Just browse positive and hide every result. You can scroll down and up again and then press and hold the hide button. Does like a hundred games a second or so. The script has basic import and export function for 2 lists.

kurochan17 days ago

sorry, I had trouble mentally parsing the first paragraph of your reply for a sec…

There is no way that it is possible for a single XHTTP GET request from a front end at a time would cause a DDOS attack on a back-end just because there are multiple URL params being sent to the back-end in a single GET request. The only way a DDOS attack is possible is if someone repeatedly fired off TONS of GET requests in succession without resting in between requests which would be pretty normal for the rests to occur if it was a human pressing a button to actually execute a search action. Its also physically impossible for normal human user traffic to cause a DDOS attack unless there was some crazy brand new ultra popular game that was released and everyone and their mother, their niece’s pet goldfish and the family’s pet dog all swarmed the site to download the sparkly new ultra popular game all at once (I’m pretty sure this type of thing only happened with games like Palworld and other major games that took the gaming community by storm). But under normal circumstances, a DDOS attack would be carried out by a hacker using a bot that repeatedly fires off nearly a thousand or more server requests at a time which overwhelms the server by pulling enough resources to crash the website.

I’m pretty confident the normal generic human usage of the site would never cause the symptoms of a DDOS attack on the itch.io servers just because the website devs were to add a multi-tag exclusion system to the website… I just can’t see that being technologically possible with normal human usage. That said, there are those freak incidents where a ultra mega popular game getting launched could cause a sudden rush of users flooding the site to download the game on release day… but even on shops like Steam which likely has a much more robust back-end infrastructure than what itch.io is rocking behind the scenes, most of the time it just leads to significant lag on the site and not a full-blown site crash. No clue what itch.io’s server-side infrastructure is setup like since I don’t work there, but I seriously doubt adding a multi-tag exclusion system would have any effect on the usability and stability of the site.

redonihunter17 days ago

There seems to be a misunderstanding.

The point I was arguing is: filtering the results by exlusion tags is hard work for the server, and might be part of the reason why there is no multiple tag exclusion.

I supported this claim with the observation that one can make, while Itch is under attack by a ddos. Before Itch introduced this cloudflare protection, there were many, many ddos attacks, where you could observe this.

And the observation was, that the first things that fail, are longer tag combinations. Combining two tags might work, asking for four tags might not.

This was only speculation about plausible reasons why this filtering feature is missing and only exists as an undocumented and incomplete feature. I do not think it is the only or even the main reason. But such a thing will certainly influence a cost/benefit calculation.