Re:Show HN: A search engine based on RSS feed
by toa697 on Thursday September 15, 2022 @10:06AM
(#32852219)
Love the idea behind this.
Generally any new way of attempting to find signal among the noise of the internet is good, and I think RSS is niche enough to automatically remove a ton of noise from the dataset, and the fact this will mostly target articles and blogs gives it a distinct flavor.
Gonna put this with the marginalia astrolabe on my "small web search" bookmark folder.
Re:Show HN: A search engine based on RSS feed
by skilled on Thursday September 15, 2022 @04:56PM
(#32858342)
I like this idea a lot but this specific project is done pretty badly. There's a lot of spam and noise in results, and no real filtering system in place.
It would be awesome to create something like this from a massive archive of personal blogs in spaces like tech, development, design, etc. Basically, a massive RSS reader curated by the community itself.
Re:Show HN: A search engine based on RSS feed
by kbyatnal on Thursday September 15, 2022 @11:57AM
(#32853988)
I was very excited by this and it's promising, but it needs a bit more work to be usable. I was hoping the RSS technique would help avoid the SEO and blogspam that Google is now filled with. Unfortunately, I still get quite a bit of that (my query was "espresso machine" and I got a bunch of listicles, etc).
I think applying a layer of curation on top would go a long way to fixing this.
Re:Show HN: A search engine based on RSS feed
by gaius_baltar on Thursday September 15, 2022 @04:27PM
(#32858014)
Great! Also, good insight for basing a search engine on RSS. It is naturally spam resistant: SEO people and other spammers don't like feeds because they cannot bring them clicks and ad views, so no incentive for spam.
Re:Show HN: A search engine based on RSS feed
by aendruk on Saturday September 24, 2022 @11:19AM
(#32963478)
This is indeed the dominant view so the observation is true, but the history here baffles me; web feeds aren’t required to include the full content. Ad-funded publishers could easily just include links and descriptions—the same as they’re content to do for Open Graph previews—and continue to benefit from web feeds instead of apparently writing off the technology entirely.
Re:Show HN: A search engine based on RSS feed
by 0x9b on Thursday September 15, 2022 @08:18PM
(#32860297)
Super cool project!
I tried solving the search quality problem (for technical content like engineering blogs) a while back by filtering using heuristics based on websites/urls. At one point I was experimenting with RSS, but found that many websites only show the past X number of entries and gave up that direction since it excluded too much content.
Re:Show HN: A search engine based on RSS feed
by pasttense01 on Friday September 16, 2022 @10:44AM
(#32866710)
The major problem is the selection of the RSS feeds. If you look at the list of feeds you see a lot of traditional news feeds like CNN or New York Post. These are well covered in Google/Google News and are fairly low quality.
Generally any new way of attempting to find signal among the noise of the internet is good, and I think RSS is niche enough to automatically remove a ton of noise from the dataset, and the fact this will mostly target articles and blogs gives it a distinct flavor.
Gonna put this with the marginalia astrolabe on my "small web search" bookmark folder.
Though now that there's two I'm probably gonna pick up a few more from here that I like: https://seirdy.one/posts/2021/03/10/search-engines-with-own-...
It would be awesome to create something like this from a massive archive of personal blogs in spaces like tech, development, design, etc. Basically, a massive RSS reader curated by the community itself.
I think applying a layer of curation on top would go a long way to fixing this.
I tried solving the search quality problem (for technical content like engineering blogs) a while back by filtering using heuristics based on websites/urls. At one point I was experimenting with RSS, but found that many websites only show the past X number of entries and gave up that direction since it excluded too much content.
Are you seeing a similar problem now?
What are the long-term plans for this? Building a crawler? Or just manual curation?