Hi, recently I thought about lemmy and came into the conclusion, that it lacks one thing to be a true competition for reddit and stackoverflow and that is the search engine and user visibility.
Let’s say I want to fix some issue on my computer. I type the question into the browser and what?.. first 3 links lead to reddit, stackoverflow some random website or eventually quora. If there was answer to my question on any lemmy instance, I’d be lucky if it was on any of the 3 first browser pages. Also the fact, that the links of lemmy instances look well… not so standard, doesen’t help.
I think the solution would be to create one “central” instance, which wouldn’t have any users, but instead would aggregate posts from technical forums of other isntances. It would greatly improve search engine visibility and provide centralized access to content without dropping the federation philosophy. It would also help creating a brand, as everyone instead of searching for answers on reddit etc. would just go to lemmyhelp.xyz and look there, knowing it’s somewhat trusted and official source. Also moderation on such site would be MUCH easier, than on reddit or SO.
What do you think?
Lemmy URLs suck, tbh. It’s just an instance-specific post id. They should definitely offer speaking URLs at the very least, ideally cross-instance
I don’t know what a speaking URL is, but I generally agree with you. It’s a problem with the fediverse on the whole. It seems like it’s a problem for the clients to solve, with maybe a useful additional URL resolver service that clients can easily use (supplementary services in the fediverse may be an area of future maturity for the fediverse IMO).
One wringle the fediverse creates, though, is that there is absolutely no guarantee that any link has a counterpart on your instance. There’s every chance that the underlying post, user or community is just not seen or subscribed to by your instance, in which case it doesn’t exist on your instance and there is nothing to point to other than the original URL.
A speaking url is human readable. How it is structured can vary (eg. /year/month/date/my-title-is-here).
Currently it is a post I’d (eg. /post/17659).
Human readable is better for SEO.
What … a search engine can’t scrape the contents behind a URL?!
It can, but afaik the url is one of the many parameters. And with an “id” syntax you “score 0” for that parameter.
Yea … just checked and both Twitter and Reddit are using IDs, though Reddit also has a title in the URL for a post, which I’m guessing means you can’t edit the title of a Reddit post (which is actually shit if true).
TBH, SEO seems to be silly big corporate bootlicking or knee bending at this point. Especially being so worried trying to please the search engines to this extent. Search engines are shit today, and they’re fully capable of indexing lemmy if they wanted to. I’m with Devs (see their AMA), SEO isn’t a priority and shouldn’t be.
You are correct: you can’t edit the title of Reddit posts.
Simple unique IDs all the way then.