A great use for reddit is the ability to search posts and opinions about any niche topic. Will that be possible with Lemmy as it grows? Will I be able to Google “instant rice Lemmy” and get a comprehensive tier list of each brand?
I imagine search engines will have trouble with all the different instances(?). EDIT: Especially with instances that don’t have Lemmy in their name, I don’t think search engines would return them for Lemmy searches?
So I’ve been working on a solution for this.
As I see it Google and others are going to have a hard if not impossible time to incorporate the fediverse, and the fact that the same content can exist on multiple servers.
So I’m working on a search engine specifically build, for Lemmy at least. Where it’ll take you to whatever your preferred instance is when tapping on a search result.
I hope to have a MVP up and running in a few more days.
Can’t emphasize enough how important this is for the growth of Lemmy. Many people I know only access Reddit through google searches.
Yep and I’m one of them. Go look me up on Reddit and I think I have maybe 20 posts over the 14+ years I was on the site. …joined Lemmy and immediately got frustrated that I couldn’t find anything. So I figured I take a crack at it. Especially since I couldn’t see how Google would ever be able to link me to my instance. Let alone make it easy to search the entire fediverse without having to write out every possible site, with new ones popping up every day.
I wonder if it’s possible to have a sophisticated search engine similar to Google’s, with BERT and kNN or vice versa. It would be the closest thing to Google search but specifically for Lemmy posts.
Easier to find a Reddit post through Google than by Reddit search.
Please pop a reminder here. Commenting for a bump.
Search their name on GitHub and you’ll find it. Star it to follow.
Can the reminder bots be migrated too?
reminder: https://lemmy.world/post/963301
If this guy changes the internet include me in the screenshot.
I’ll invest in seed funding stage. 😂
Interesting. I hadn’t even thought about how the fact that instance1.[post] and instance2.[post@instance1] is essentially the same thing and how search engines would handle it. Interested in what you come up with!
Thanks. If you do some digging you can find the project on GitHub but note that it’s a work in progress still. The UI is lacking and it’s rough around the edges but it’s “working”. And I still need to do some optimizations on the crawler itself, etc…
It’s also going to be completely self-hostable just like Lemmy, etc…
Hey, can you dm me the git link, i would like to contribute if i can : )
Search their name on GitHub
Do you mean the ‘crawler’ repo?
probably, that’s pretty much what a search engine is. A crawler that saves it’s results into an optimized database, to make it easily searchable.
I can see this being helpful
(commenting so I can bookmark)
IDK, isn’t it the same for reddit? It also encourages crossposting, so the same content is on there several times. Maybe I don’t understand the fediverse well enough yet, so please correct me if I’m wrong.
deleted by creator
The mastodon crowd was verry anti on search engines, and killed projects like this.
But yea, do it! I think the lemmy/kbin crowd would mostly like itYa I only index Lemmy instances, for now. Mastodon and other ActivityPub servers may be in the future, but ActivityPub has some limitations so I’m stuck using the Lemmy-specific APIs.
we need a search engine for the entire fediverse, it would be a game changer
That sounds awesome. Can’t wait to see it.
This is a great project
That is great. Thanks for the initiative. Have you considered contacting the people at DuckDuckGo so that that search engine can access Lemmy/Kbin content?
deleted by creator
deleted by creator
Lmk once it’s up
This would be awesome!
Shit dude, that’d be a sweet little tool.
I am surprised noone mentioned https://fedi-search.com . It’s working pretty well. Full credit to Benjamin Pryor for this
Tried it, pretty cool, though seems still depend on search engines’ indexing. The instance that I’m on seems not indexed.
Also it’s interesting it uses
intextto identify whether the results are from fediverse.
You can use a search query to include only results with Lemmy’s footer, which is consistent across all Lemmy instances. I made a post about it here: https://lemmy.world/post/342365
EDIT: Fixed link
You formatted your link wrong. here is the correctly formatted link
The label field is for what we will see. And the URL field is the actual link.
The markdown syntax is [link label](actual link)
Also: You can always paste links without formatting https://lemmy.world/post/342365
Thanks! Updated my comment with a working link.
If Lemmy becomes a source of enough information like Reddit is, search engines will index it. SEO is a marketing thing, and a place like Lemmy doesn’t really need that. Google, DDG, etc. all put engineering effort into making sure sites with lots of information are indexed and available in their search results.
In the future they eventually might be, for some instances. Though definitely not for all of them, since some of the instances might disable indexing.
I’ve actually already seen a few Lemmy results (lemmy.ml) in Google searches, the trouble is it doesn’t link to individual posts, just the community so it’s not particularly useful. So it definitely is possible, just needs to be improved to be able to index posts.
It’s up to the individual instance owner and Lemmy the software itself enabling SEO. It’s just getting started now so it will be long time before that.
likely not THAT long. I’m sure things are already being crawled
Digg.com was the big thing with Reddit trailing. Digg began tweaking the experience toward a more profitable model. I had already come to Reddit when they went too far and there was a sudden enormous migration from Digg to Reddit. Digg went from being THE social media aggregator to being nothing in a matter of weeks.
Reddit is more deeply rooted, so I think it will stick around, I’m cool if Reddit keeps those who are happy with corporate model busy so we can do our thing here.
It’s certainly not going anywhere unless they end up selling it to someone who shuts it down and uses the posts and links as SEO boosting.
Well, Digg.com still exists. It’s just that no one cares.
when you just loaded their site to test you just doubled their monthly active users.
Most likely if it’s being sold for anything it’s for language model testing.
I think it is preferable to ask other search engines like DuckDuckGo to index Lemmy info. Google is full of garbage.
Brave Search would be better, they have a dedicated section on the results page for discussions.
Brave is an advertising company and should not be preferred.
By that standard, so is DuckDuckGo, Qwant or Ecosia. Until we have a good federated search service, we’ll have to deal with them and pray they don’t enshittify like Google.
Depends on Google. These tech companies don’t like new platforms, especially those competing with established ones like Reddit. You’ll see that Google often discriminates against Lemmy or Mastodon.
Respectfully: Fuck that.
If you want to find the best instant rice recommendations on Lemmy, Lemmy should have a functional post search function, rather than me relying on a malevolent corporate entity like google to index all the content.
Search has gone to shit as the Internet has embraced social media sites, an upside of this is that wikipedia+Lemmy+key word search, mayas accurate as asking Google Bard or bing, and they can be built on entirety open tech.
Cool rage but you dismissing search indexing is kinda hilarious. It’s not going away and it’s what makes the web. Would you rather have 3 big websites instead of indexed web?
Would you rather have 3 big websites instead of indexed web?
That’s what we already have, I’d you need to find stuff by doing site specific googles, both google & that site have failed.
The web is dead, it’s been dead for a while, now is the time to build something new in it’s wake that rather than depending on closed source algorithms, indexing 3 big websites, we could just search the 3 big websites directly.
That’s definitely not what we have.
I disagree. I’m not sure why you say that. I Google stuff as a job and it’s certainly not just the big 3 websites. I personally rely on selfhosted searxng.
What the world needs is a web indexer/search engine that operates similar to wikipedia. A non-profit that can focus on providing a useful service for the public good that isn’t driven by profit motives.
I disagree. I have a very successful technical blog and there’s a big community outside from the big websites that produce awesome content.
Though, you do have a point that it could be better but we’re all working on it - that’s why we’re here on Lemmy! :)
I came here to say something similar but you put it nicely.
@QuinicV Why would it not be possible? It depends on the software, if all text is open to be indexed. Kbin and Lemmy instances are basically open forum software and are indexed by search engines. You can test it in Google or other engines by forcing to search on the site only with
site:lemmy.world are posts indexed?, which would be an empty search result if they were locked down like discord content.But what if the post I’m searching for is not on lemmy.world? Say the instance doesn’t even have Lemmy in their name, like beehaw.org. How would a search engine index it? How would it know it’s part of Lemmy?
There will be links to everything somewhere. The same way you knew to get the cave in the same way you know to get to Lemmy. There are already links that have been posted to Reddit that are in archives that are easily followable. Google doesn’t just search one or two things they search all the links to the things and then the links from those things to other things. If Google can’t figure out how to get to it chances are you don’t know it’s there either.
This is a case for search engine optimization or SEO. Including keywords and tags in your robots.txt is an important part of making sure you exist to the rest of the internet if you want to be findable via search engine. “Lemmy” as a term will almost certainly be one of the standard tags applied to any instance that knows even slightly what they’re doing or else has been prepared a tutorial by someone who does. I also expect terms like “beehaw” and whatever other large servers prove themselves popular and robust enough to stand the test of time.
@QuinicV This was just an example how to prove that the content from Lemmy is indexed and searchable by Google. If you do a websearch without limiting to a specific domain, then it will search through all indexed Lemmy content that is known to Google too. At the moment there is no way to search Lemmy (or related) content only.
What we need is a search engine that only tracks ActivityPub content from Lemmy, Kbin and Mastodon (and others). Let’s call it ActivitySearch. Maybe SearX engine could be modified to do this.
Reddit did not start out as the thing to google, it’s 15+ years old, only in the last 5y I started prefixing my google searches with reddit.
I actually found Reddit by googling things. I had seen it 5 or 6 times over a few years, and eventually I just went to the main site. I might have even used Reddit in the search before I joined. Regardless, I had recognized that all the best answers for tricky problems that I had were coming from Reddit before I even joined 11 years ago.
Everyone’s experience on this will be different, but I personally started using reddit about 12 years or so ago largely because at that point a lot of my Google searches were already pointing me towards reddit. I wasn’t necessarily going to google specifically to find reddit results, but since that’s where I kept ending up i figured I might as well go straight to reddit. And since reddit’s search function is and has always been trash, i pretty much immediately started using Google to search reddit.
I realize Lemmy needs to get much bigger for that to happen. My question was more directed at how search engines would handle the fediverse. Though I see now that that wasn’t very clear.
Basically use
<query> site:lemmy.world OR site:lemmy.ml OR site:beehaw.org OR site:kbin.social(or whatever main instances you want to hit)You can also use this for custom browser search keys like the following https://duckduckgo.com/?q=%s+site%3Alemmy.world+OR+site%3Alemmyml+OR+site%3Abeehaw.org+OR+site%3Akbin.social
I imagine that would be quite inconvenient… Especially as Lemmy grows and has potentially many more instances.
I believe that DDG has a shorthand for site:Reddit (without the .com). If lemmy gets popular enough DDG may implement a similar shorthand that incorporates the fediverse without us having to use a massive string. Like if it gets big enough, we may not have to solve this problem because others will see the value in making it easy.
That’s my hope at least.
this isn’t very good, i cant even find basic things with it, and sometimes nothing at all
I would argue that eventually, yes, one will be able to google search Lemmy just like Reddit.
I wish there was a way to get an entire Reddit archive over here. Realistically I’m still going to have to search Reddit because it has 10+ years of answers to obscure questions.
Minds more intelligent than mine are probably already at work on these problems. I’ve seen multiple discussion of people saying they are designing and working on solutions. It may take some time to see results, though.
Only if we make sure the tech giants don’t kill this platform
How would they? It’s all decentralized
Google could prevent lemmy pages from showing up in the results for example.
Or they could adapt the protocol, make their own slightly tweaked version of it and let it die, which apparently often also kills the original protocol due to newly introduced compatibility issues, etc.
Not sure about the second part, I read about it here somewhere where they mentioned an example of that happening as well but I can’t find it anymore.
They could join the fediverse, attract a majority of users to their platform by adding attractive features, and then remove themselves from the fediverse effectively killing it
https://ploum.net/2023-06-23-how-to-kill-decentralised-networks.html
Good read, I had no idea about those past incidents. I always herd people say “embrace, extend, and extinguish” but I never knew the history of it. After finishing it I’m just worried now, ahah. I need to look into xmpp because it just sounds like matrix to me and if that’s the case why did matrix get created in the first place
Probably a legal attack/false flag op related to prohibited content and/or terrorism
Like xmpp
Make a giant instance, get all the content there by pumping users and making cool shit, slowly customize your instance and extend the protocol with features so that ours become incompatible in annoying ways
Add wikis, overhaul user profiles, achievements, posting to your own profile, games, whatever, then get tired of supporting the fediverse interface and shut off the API
You instantly can’t read 95% of your subscribed /f/s
I have seen at least one user claim they got a result from lemmy when searching a question on google. YMMV though. Lemmy is a fraction of the size of reddit, it will take time for posts to reach the level that google starts indexing them specifically.
I got one. The Google link brought me to the instance though and not the thread. I was able to find the thread though, so it kinda worked.
