I’m continually thinking about ways to better find and curate content. Part of this is just a general desire to improve but part of it is because the social media is constantly shifting and tools are always coming and going. (RIP Listimonkey, Undrip, and Nuzzel.)

With or without ancillary tools, RSS feeds have been an essential part of my efforts for decades now (I started learning about them in early 2000.) Lately — maybe in the last couple years — I’ve started leaning on RSS even harder. There’s no point in following a Web site on traditional social media when its organic reach is about .00001% of its followers and you have about as much chance of seeing one of its posts as you do of getting hit by an asteroid. Further, garbage on Google (which ends up infesting Google Alerts) means that my keyword searches aren’t as productive as they used to be.

I’ve been trying to address this by finding RSS feeds for information-rich resources like institutions of higher education, government agencies, nonprofit news, etc. The snag there, of course, is that those feeds aren’t based on a keyword search and there’s going to be a lot of content I have no interest in. This is especially true of higher education institutions; announcements about student fees and registration dates are critical for students but will never be useful to me.

Simple solution: keyword-filter the RSS feeds! But I couldn’t find an existing solution that did everything I wanted, plus if I was able to keyword-filter the feeds as thoroughly as I liked I would be able to throw my entire set of feeds at it (roughly 7000) and I’m not sure how well that would work with a web-based solution.

I learned some about node.js and making a local server when I was playing with my Match Game machine vision thing, so I decided to use that again to make an RSS feed reader that can handle a lot of feeds and do intense filtering. And to explain to you how balsa wood my programming is, there is no database. The program runs daily, filters feed content to the previous day before keyword-filtering it, and saves feed items to a series of date-coded JSONs that I view with a local reader. I don’t generally go back and look at old RSS feed items, so I’m not sure I’m going to even keep them. A database seemed like overkill.

But it seems to work pretty good. The aggregator visits all the feeds every morning, filters each item by title and by description (separate arrays for each) and then saves to the aforementioned JSONs. Since the feed items are in one big glob, I don’t need a list of the feed sources and I can really spread the reader out:

The feeds load through a basic left-right up top, with a dropdown menu to choose one of the JSONs in the output directory. Feed item on the left along with a place to take notes (it counts characters in case I find something I want to post to Mastodon) with the right half of the page devoted to showing fetched content. I usually don’t need to do that because feed items tell me enough, but it’s a handy feature to have if I need it.

Since the data’s local I’d be able to go through it faster than a web app anyway, but with the this layout and nav I can fly through the content. I’m covering a lot more ground, with the bad news being that I’m ending up with more in my ResearchBuzz queue than I can use. Going to have to get more heartless about pruning it.

I’m just getting into seriously using this new feed reader as the holidays are starting up, which means the university feeds are going to be a lot lighter until January. I’m going to try to work on the filters to get my feed items as crud-free and information-rich as possible. Maybe some ideas for new features will present themselves, but I’m so delighted with how fast I can go through my feeds now that I can’t think of any improvements at the moment.