I always believed Google was successful because it was pushed and advocated for by geeks, I feel that those days are coming to an end.
I use DDG and I think it's a pretty solid alternative, although it requires a little more effort when looking for certain things. Non-tech types have been curious about why I don't use Google, they seem to be getting the tracking is not good thing.
Those days ended the day Google went public, and quite possibly well before then.
It's also a self fulfilling prophecy though - as my threshold to scroll through results on DDG has dropped markedly from when I first started.
There may be a way to get these things from DDG, but I'm haven't had the time to look for them yet.
Still DDG is pretty good, and obviously will get better the more we use it.
!year girugamesh
vs. girugamesh
you definitely get two different sets of results, one general, and one discussing the term in retrospective (with urls denoting /2016/ for example). I think they don't highlight it well, but the filtering is definitely going on."it requires a little more effort when looking for certain things"
What type of things?
1. Older results. DDG's recent (<~5 year) results are pretty good. Going older than that for online content, not so much.
2. Books and journals. Google Books, Google Scholar, and Google Ngram Viewer are phenomenally useful tools, no doubt about it. There's no single resource which can substitute for them. Worldcat (library search) is good, if you've got access to various search databases (usually via an academic library) that's useful, and content access itself is possible (with questionable legality) via Sci-Hub and LibGen. The Internet Archive has less content, but greater legality.
3. Specific technical queries are generally less well answered by DDG, but easier to direct (via bang queries) to the appropriate source. StackExchange, Debian, Debian packages, Ubuntu, Microsoft KB, etc., are (I think) all covered. The ease of filtering from a general case to a specialised search is a real win.
4. Date-bound searches. I very frequently want to look for something between specific dates.
5. Less-prominent content. I use various sites (Ello, Reddit) as filing cabinets of a sort, and being able to search my own posts there is useful. Reddit's post search is good, Ello's not so much. DDG doesn't index (or retain) Ello results well.
Given there's no tracking (feels a bit odd saying that nowadays with any kind of seriousness), do they monitor which queries people use !g for, and try to optimize their engine for these?
One feature I do like though (I'm not sure if Google etc offer this nowadays), is they'll often provide the top-answered StackOverflow answer in response to a query above the search results. That has proven handy a fair few times, esp. for some of the shorter answers (e.g. arcane command lines, git invocations etc).
Google does something like this when it shows images or scientific articles on top of other results.
DDG could even predict that it needs to fallback to Google in last resort.
DDG's auto question-answer stuff is pretty good. Not quite as polished as Google, but IMO far less annoying. Google thinks it knows better than I what I want. DDG seems willing to allow me to provide reasonable prompts. "Weather" and "what is..." queries, for example.
When it comes to this, it's too late. The dollars have already won in the highest court.
You can't just educate everyone. The law need to change to make this sort of thing illegal. Personal data should be limited, and controlled by the person involved. You should not be allowed to sign it away.
They aren't complicit, they're ignorant. Capitalism breaks down if the customer has no idea what they're buying.