Ask HN: How do you avoid bullshit tech content in search? | Hacker News Reader

Ask HN: How do you avoid bullshit tech content in search?

27 pointshigerordermap15 comments

This happens when I want any deep technical treatment of a topic.

There are hundreds of SEO-optimized websites that basically post obvious shallow information and still achieve to trick search engine algorithms with great effort. I don't want those shallow information articles. There's wikipedia always for surface introduction of the topic.

Examples of such sites include GeeksForGeeks, Many medium blogs, DZone, ZDNet etc..

Are there any search tricks / filters / alternative search engines or proxying search engines that allow to avoid bullshit tech content?

View on HN

I typically only look for upvoted content. Prefix your searches with "site:reddit.com", "site:news.ycombinator.com" or "site:stackoverflow.com" - The social/human filter is quite a good one in my experience and it gets rid of all the Medium-like personal branding fluff.

So, I think it may be more what you’re searching for and less the SEO of it. At least from my experience.

1. Use more specifically targeting words. I.e. I needed a specific valve for a new gas line I put in a few months ago. The kind of thing that’s not sold at Home Depot, but at wholesale outfits specifically tailored for plumbers. Instead of search for “3/4 gas valve” I searched for “3/4 gas cock” as “cock” is an industry word for “valve” and the results of which would get me more precisely to the sources I was looking for.

The same can be used for technical stuff. An example, instead of searching for “how do I add a firewall rule to a router” I’ll search for “cisco allow udp 53 rule”.

2. Often I find that the same couple sites are the places that have the specific information I’ve looked for in the past. I do this for product reviews as an example: “cove security reddit”. For a more technical example, during the gas line, I searched: “3/4 gas 10” WC Reddit”. This got me into a few conversation threads where a master plumber was explaining the max BTU usage I could have on a similar line to mine, he then explained the math behind it with some very valuable knowledge from his experience on permitting and gas company nuances. Exactly what I was looking for.

These are two basic examples that show methods that work super well for me in finding what I need and skipping the bullshit out there. These combined with learning Googles operator syntax is the ticket.

You could go as far as messing with DNS or some blocking methods, but that seems likes a ton of work for little return. Just learn how to search a little better. :)

I personally don't know what the purpose of google is now. Any time I want to find something even a bit more niche or specific, it's completely useless. Try searching for some economic data, enterprise budget allocations or something similar, it's impossible. They seem to be optimising for recipes, questions like Katy Perry net worth and recent news.

> I want to find something even a bit more niche or specific, it's completely useless.

I have a bad habit of buying (cool) old hardware without checking if someone has archived the documentation. A few times I've been burned, having stuff for the which documentation seems to have been lost to time. This is typically business / enterprise hardware from the 80s and sometimes 90s, usually for a specific industry. Typically with google, a vast amount of the results you see when searching for info on these things are sites that seemingly scrape current / old eBay listings and make them appear as listings on their own site.

With that said, I recently made such a purchase. I was excited to start hacking around with something I bought, till I realized I couldn't find any documentation online. the original company, still in business, had 0 mention of the product on their site. The few pieces I could find were old marketing materials, sales crap that wasn't overtly useful to me in the long run. I tried so many things to optimize the query, but no matter what I did, I got very few relevant results. Often google would ignore chunks of my search (even ones wrapped with "") to provide completely irrelevant results for similarly named products (I got a lot of results for a car model). It seemed odd to me that this was lost to time, as it was reportedly used by some very high profile customers back in it's day (mid-90s - early 2000s). The fact that the documentation never uploaded somewhere was strange.

After feeling rather defeated, I gave one more try and found something interesting. One of the few relevant results I found for the product was an old PDF sales paper. IIRC it was around the third result down. I went back to it, and decided to navigate to the site it was hosted on. It was a Latvian equipment reseller with the site entirely in Latvian (the PDF was English). I poked around and with google translate and was able to find a link to a user manual right next to the link that had the aforementioned sales paper. I was completely floored. Why on earth, was google able to find that sales paper, but not that user manual on the exact same sites. They were both PDF's and both had select-able and so I assume index-able. For some reason though no combination of query terms, modifiers, etc. related to the company name product name and terms related to "manual" could I get google to return this manual.

While this was a shocking find, I'm still out of luck for now. The product came in two editions, with the one difference being very important. The user manual I found was for the one I do not have, and presumably more common. I emailed the company asking if they might still have a copy of what I need, although I expect I'll get ignored or told no, given that this was a product likely retired back in the mid-2000s and was much to expensive for a consumer to afford in it's heyday. In the meantime I've done a basic teardown and have managed to learn just enough to be able to proceed in absence of docs.

> A few times I've been burned, having stuff for the which documentation seems to have been lost to time.

My best advice would be Yandex. It has saved me on numerous occasions. It has the edge of getting much less(or not responding to) copyright claims and unique Russian language forums for sharing documentation. Of course it be nice if you actually could read it to get around easily.

Another place might be the various Chinese forums for that stuff, but my guess is they mostly have documents on the stuff that is made in Asia - e.g. iphone sheets, layouts.

Generally stay away from Google if it's not something overly mainstream.

1) Remove/Block Certain sites from Google/DDG. Use

https://iorate.github.io/ublacklist/

2) Using Host Files, i think that you can block domains of the sites that you are not interested.

3) Search Individual Sites .

Instead of Searching the whole web , there are some sites which contains good resource of certain field. If you are searching from that certain field, first search that site.

I save a txt file at the start of a search with a random number in the name (I let vscode decide that), then I can just save files with the random string and can sort them later.

I save all my search terms as I proceed (and record all word-stemming, plurals..), and indent hits, and then search for cites and indent again where hits, then when a sub-search dries up I start again with a new string and so on. I find it saves a lot of time, reduces saving junk and I can always go back. And I get a valid cite chain.

I search in epic-browser as it cuts down the clutter. Mostly on scholar (I turn off citations and patents and usually start with all years as some things are mis-dated, this also and issue with ResearchGare where their pub date May be decades after the paper date), sometimes the open web (altho' mainly on specific domains; .edu .ac ..). Some countries I exclude.

For me, I find search is ok for finding places to find places, if that makes sense. Like I've had a lot of trouble finding detailed information on a few technical niches, but search led me to forums that led me to private slack/discord channels on what I was looking for and active community members that had already filtered the good/bad content in their working memory.

We made a firefox plugin a few years ago meant for human-raking search results exactly for the same reason: too much noise. Then FireFox switch to WebExtensions APIs for programming addons and we did not had the resources to maintain it. It seems that the search issues get worse since.

Focus on the documentation. Experiment with new technologies by using the associated software. Usually a deep dive into the manual is enough. When this isn't provided, there's sources. When neither of those is present, I'm not interested.

Have you ever struggled with bullsit content in search results? Scroll down for the top 10 tips on how to avoid it.

Sounds like you need human recommendations for content - eg from IRC, colleagues or friends or maybe ask on Reddit or even HN. Always worth searching HN.algolia.com and then finding advice / links on HN from there.

nowadays I always add ycombinator after my search terms to see if it has been discussed here.

IRTFM.

Well, we live in the age of fake news, and fake celebrity pix, and also fake research. An ingenue to any field is easy to fool - an expert - not so much. In effect, the defence against fake anything is a comprehensive personal knowledge base on that subject, be it news, a celebrity or research. Ask yourself:- do you really think President Trump sold Alaska back to Russia for 10 times what was paid? Not so much. This SEO BS muddies the water and makes it hard to wade through the crap. Many give up and become flat earthers or antivaxers because they could not discriminate valid data from crap.

Reply on news.ycombinator.com