I am genuinely confused why search is so bad in the major email webapps/clients. Search is a well studied feature, and it seems like it's something that should just work but I can never find the thing I'm searching for in my email (especially O365). Knowing the date and then scrolling often seems to be the most accurate way of finding things...
I think search has been deprecated in general because it gives the user too much control over the output. Through search, people can quickly find what they are looking for, which is bad. The goal has instead become to feed people tiny scraps and hints of what they're looking for, while leading them on a long trip past any number of sponsors to where the thing they're looking for might be.
I have to assume that Outlook email searches have already been set up to have ads injected into them, when/if one day Microsoft decides to flip the switch. Actually, I'm so out of touch with Windows they might already be doing this.
> The goal has instead become to feed people tiny scraps and hints of what they're looking for, while leading them on a long trip past any number of sponsors to where the thing they're looking for might be
Search just seems bad in general in many applications. So many these days do not even support a verbatim (as in, find what I typed, exactly) search. They insist on ignoring certain characters, fuzzy matching, or treating everything as individual words and if it finds one it has done its job and earned a gold star.
I have a feeling it's based on tokenising the input rather than a string scan like we'd do in the old days. Harder to match a literal string if all you have is a tree of tokens or something, I guess.
Opengrok was the first time I ran into this years ago. We had a perl code base, perl syntax is well known as "an explosion in an ASCII factory", so it was a real pain trying to find exact text matches using it.
As I’m professionally working on a niche search engine, let me offer this: it’s a notoriously hard problem that seems simple at first, but requires catering to a bazillion different edge cases; every optimisation you do makes another case worse.
Having said all that: I also hate how shitty search almost everywhere is. It’s hard, but not that hard.
…which is the problem I was referring to: by optimising for that—your—use case, those of other people will invariably suffer.
We only have a single text field as the input; how are we supposed to guess whether you want to find an exact match of the phrase, a fuzzy match, at least one of the words provided, or any other possible variation? Also, are you interested in the content, the subject, the recipient, the sender address you used, a header field, an attachment, what have you? Do you want them ranked by the frequency of the word, or the position from the start of the text? Does it count those occurrences in quoted passages of previous mails downthread multiple times? What if it’s a stop word?
There are of course sensible ranking solutions and heuristics for these questions. I just want to highlight it’s not as trivial as it first sounds. Most mail clients probably don’t ship with a Lucene index—while they should.
I use Thunderbird and it's approximately 100x better at searching for emails than Excel. I just tell it if I'm looking in the subject, in the body, in the sender, whether it's fuzzy, etc, and then it pulls up the emails.
Whereas Excel doesn't ask shit and, in return, doesn't have a working search.
With mu4e (an Emacs package), you can have lightning fast searching across multiple mail accounts. And with a bit of work (https://stuff.sigvaldason.com/email.html) it will happily interoperate with Microsoft Exchange systems that require the OATH2 dance.
Haha, I love your reply because it brings up auth. You can't do the slightest search or action without first pinging a web server with an auth token. Yet another source of wasting milliseconds.
