Hacker News Reader

Comment by BenoitEssiambre | Hacker News Reader

I kinda wish that there was a timelapse for Google search. Search seems to be heavily biased towards stuff popular right now. It seems older things disappear fairly quickly.

I've been in a position where I wanted to search for a website that was popular in 2012 without having the exact name, and having to find an old forum post that linked to it to find it.

View on HN

https://twitter.com/searchliaison/status/1115706765088182272

Does this help?

foobar before:2012

It's not perfect but quick and easy for what I mostly do which is trying to find older news articles on stories that have changed

That helps actually!

Google Search does let you limit the results to old sites (though I realize this isn't quite equivalent).

A time-lapse of the search algorithm is what I would like to see

Given that the set of possible searches is unbounded, first instinct is that the whole system would have had to be designed around that use case from the bottom up for the problem to be remotely tractable

a git history of Google's code would suit me

At the very least, I wish they would update the UI for that. Every time I want to use it, it's like 4-5 clicks and very hard to use. I wish they could just give a nice timeline and let you slide across a given range, and maybe also see a graph of the activity (basically like Google Trends)

There's a good chance that the additional clicks are deliberate. Certain queries are a lot more resource-intensive than the average and it's not very difficult to turn Google Search into a money losing business, even with the ads.

A bit off topic: The Google's feature requires me to input date on MM/DD/YYYY format even though locale is ja-jp (We NEVER use it). When I just wrongly input "2021/04/16" to field and search, they even don't show format error, but show the result of 184/05/04. What??? I believe it never tested by locals. I've sent report but of course no response.

I miss the ability to get the cached version of a page easily, without having to jump into the advanced search settings - if it is even possible to do that anymore, they may have pulled it completely.

I hear you - it would probably need to be cross integrated with web.archive.org in order to actually vend most results given how rapidly we are seeing bit rot.

I think I read somewhere that behind-the-scenes the Google crawler is in-effect a super-archive.org, keeping a copy of every page it has crawled. This sounds outrageous, but I believe it's feasible with compression.

I believe this is actually a thing that Google implemented as part of a googling game. I don't have the link on me right now sorry.

Reply on news.ycombinator.com