Hacker News Reader

Comment by Ixio | Hacker News Reader

> There’s absolutely no use case out there, where I search for something, and then I say, hey, I believe my search result will be item #3175 in the current sort order.

I actually have a use case. When the search filtering functionality is lacking or overly complicated to use. For example with Gmail if I can't be bothered to look up how date filtering works, since I receive emails at a fairly constant rate I can sort of guess that item #3175 might be around the date I'm looking for.

I strongly disagree with anyone that thinks Facebook "got it right" with their timeline. As far as my experience goes it's very easy to see something interesting on the Facebook timeline only for it to refresh and lose it forever. It can be very frustrating not to be able to get a consistent timeline.

View on HN

Also from the article: "when was the last time you googled for SQL and then went to page #18375 to find that particular blog post that you were looking for?"

The key difference in use cases here is that one is searching my stuff, and one is searching everywhere. Nobody wants page #18375 of everything. People do occasionally want page #3175 of their own stuff.

Even when I'm searching other people's stuff, I don't always just want to browse through it linearly. Sometimes I want to skip around. Meaningful keys such as timestamps would be better than just page groupings, but page groupings, so long as they are stable, are still useful. Sometimes I do want to skip to page #18375 because I know that I've already browsed pages #1 through #18374 on a previous visit and I want to start where I left off. You can't typically do that with infinite scroll.

Unfortunately, the "so long as they are stable" constraint is incompatible with idea of mutable database backend. SQL pays massive performance price for OFFSET and it's results are still neither fast nor stable.

Which is why i browse twitter chronologically. I hate when sites don’t provide a way to edit the offset e.g. by editing the url

I've got plenty of use-cases of this to know that GMail's number of pages is not accurate.

I think it isn’t meant to be accurate. Not even sure if it is stable on page refresh

Then why do they include it at all?

what matters is that there might be more search results, not their precise number

I couldn't even believe I read that, the facebook timeline is easily the worst web UI out there.

The best part is when you follow a link from your timeline, then press the back button. With any normal website, you'd be back at the same position in the page where you were before. With Facebook, not only do I not get that, I get what seems to be a random position in the timeline.

Facebook having a timeline which is not stable/deterministic is a peculiarity of their implementation. OP refers to a simple timeline which would not show such undeterministic behavior

Reply on news.ycombinator.com