I've been running siftrss for about five years now. It lets you enter an RSS feed, add a filter, and get a new RSS feed excluding the stuff you specify. I made it to scratch my own itch.
Originally it wasn't going to be public but I thought, "eh what the heck, it's only a bit more effort to put a simple interface on it." Since then more than 100k feeds have been created and donations have paid for the minimal hosting costs.
I do get feature requests from time to time,and I would like to fulfill them eventually, but for the most part I rarely work on it. I'd love to have richer, more powerful filters with boolean logic along with feed combining, rewriting of tags, proxies, all sorts of things... but I've been a software engineer long enough to know that the greater complexity would mean more people emailing me their demands and blaming me for their (mis)use of the tool. It just wouldn't be fun anymore. It'd be another chore.
I've thought about trying to monetize it but 1) it seems unlikely that it'd ever amount to anything substantial, 2) probably wouldn't be worth the effort, and 3) kind of feels against the spirit of RSS.
I guess in some sense it has succeeded, but in reality it succeeded on day one when I was able to use it myself.
Do you know https://linklonk.com from @lonk11? It feels like RSS is slowly regaining traction with all those small projects that add missing functionalities.
Indeed, RSS is a great minimal API and we need more tools that can create and consume feeds. For example, you can create a feed in siftrss.com and submit it to LinkLonk so it could rank content from feeds you follow based on their signal-to-noise ratio.
Regarding the OP, I'm building LinkLonk to test out a different mechanism for cultivating trust among strangers and for discovering interesting stuff along the way. I wouldn't say I don't care whether it gets attention. That's the whole point of the experiment - to find out if it is a viable alternative to the AI driven feed. I am prepared though to maintain it for years and not looking to "fail fast".
Ooh, this is something I have contemplated making for a while. I think a really good way to get sophisticated filtering is to allow filters to be specified in snippets of JS (or LUA) that are then executed by a sandboxed interpreter that presents the filter with a very restricted “DOM” representing the post+feed metadata. Include things like author and title and body and links, and let the user do whatever filtering logic they want. And if you really wanted to make it user friendly, instead of having them edit the code, give them a blockly UI for the filter editor. You have to limit the CPU time given to a filter to avoid runaways but that isn’t hard.
I've been running siftrss for about five years now. It lets you enter an RSS feed, add a filter, and get a new RSS feed excluding the stuff you specify. I made it to scratch my own itch.
Originally it wasn't going to be public but I thought, "eh what the heck, it's only a bit more effort to put a simple interface on it." Since then more than 100k feeds have been created and donations have paid for the minimal hosting costs.
I do get feature requests from time to time,and I would like to fulfill them eventually, but for the most part I rarely work on it. I'd love to have richer, more powerful filters with boolean logic along with feed combining, rewriting of tags, proxies, all sorts of things... but I've been a software engineer long enough to know that the greater complexity would mean more people emailing me their demands and blaming me for their (mis)use of the tool. It just wouldn't be fun anymore. It'd be another chore.
I've thought about trying to monetize it but 1) it seems unlikely that it'd ever amount to anything substantial, 2) probably wouldn't be worth the effort, and 3) kind of feels against the spirit of RSS.
I guess in some sense it has succeeded, but in reality it succeeded on day one when I was able to use it myself.