Show HN: Search engine for software eng blogs and technical content

by 0x9b • 4 years ago

13 7 4 years ago

Sodha is a micro search engine focused on software engineering with an index hand-curated(ish) with several hundred blogs and resources.

The motivation behind it is that there is a lot of great content written by engineers in personal/company blogs, niche websites, and even by big tech in some of their more obscure doc/devrel pages. And I wanted a way to quickly access that by topic.

Vaguely inspired by ArxivSanity and HackerNews :)

I've only been working on this for a couple of weeks, but would love to get your thoughts HN! Some ideas I have (beyond indexing more pages) are: adding similarity/recommendations, fully automate the crawler, and allowing voting/ranking or some similar mechanism.

Content wise: there's a lot that should be on here, but it's a modest start so far. If you have suggestions please let me know here: https://a.sodha.dev/suggest

annagrigoryan2• 4 years ago

actually neat idea!

Google is worse and worse for finding useful information and I think that community curation can be a better way to find knowledge.

actually here's an article i like on that topic https://future.a16z.com/the-future-of-search-is-boutique/

I'd say you don't need ranking in the sense that it is done in Google now. You need to able to make clusters of information based on searches and let people dig in.

like "react keynotes" or "routing in react".

0x9b3• 4 years ago

Thanks for sharing the article and the feedback!

Community curation is a good concept. I'm a little hesitant because community contributions can degrade a platform (e.g. people writing marketing content on Wikipedia). But the biggest hurdle had been building a reasonable content index, so despite that risk, the reward will seem to be worth it.

Re ranking: what you're saying makes a lot of sense in the context of topic discovery. But is less useful for q/a or reference style searches. Although it will still probably be good enough.

msarrel• 4 years ago

I tried your site out and it is fast and found good results for what I searched for. Thanks?

knowmad• 4 years ago

This is awesome! Great work.

The only thing that's stopping me from bookmarking the site to use for all my technical searches, is the lack of a time filter. The first search I did resulted in a blog post from 2017. Sadly in the world of software I usually have to use the "past year" filter to get useful results.

This would be amazing if it had a time filter with: - past week - past month - past 6 months - past year

0x9b3• 4 years ago

Super cool to hear you were considering using the site for all your technical searches!

Time filter is a great idea. In an earlier iteration, I tried ingesting creation time in the crawler. Unfortunately, not all sites would have a timestamp included and they way they incorporate it was a mess (metadata vs url vs a random <span>), so it didn't make the release. But this is definitely something to revisit soon.

jeauxlb• 4 years ago

I visited the link to suggest some missing sites but received a modal I can't get rid of without accepting cookie tracking:

Approve third-party cookies

In development and staging instances, Clerk uses third-party Javascript cookies to manage user sessions. The browser you're using requires us to request explicit permission to use third-party cookies.

0x9b3• 4 years ago

Thanks for contributing (or trying to)!

I goofed, but it's fixed now! I use the endpoint where the public suggest page is hosted for private admin access as well. Those admin pages are authed with Clerk, but I was never using the production Clerk instance in production. Whoops.

Show HN: Search engine for software eng blogs and technical content

Related Stories