A new dumb trend is threatening my patience: the all-inclusive search.
I’m probably just pissed because I’m old. Back in my day, search was used to find specific things. If you wanted variability, you included a wildcard, like “*” or “?”. People knew how to use these wildcards to focus searches on what they were looking for, and to filter out everything else.
Nowadays, the kids search with vague yearning. They hope to happen upon something that smells a little like a right answer. If something blows by that suggests relevance and tickles their informational appetite, their index finger is stimulated to twitch out a click.
Modern search is right there to serve that informational yearning and harvest those monetized clicky twitches.
Unfortunately, increasingly search is only there for that.
Any clear-minded soul searching for some specific needle in the infinite fuckshit stack is completely out of luck.
Here’s how it goes. If you enter some specific expression, because you want to search for precisely that verbatim thing, the search algorithm carefully picks the expression apart into isolated terms. Each term is blurred, fuzzed, inflated and broadened to include every possible form and tense of the word, every misspelling and homophone, every synonym, every vague association. The algorithm then searches for any individual instance of any of these maximally broadened terms and returns a galaxy-sized Boolean OR of chaos.
And of course, to keep things as simple as possible there is no way to override this default all-inclusivity. There is no way to narrow the search. Quotes do nothing.
Amazon’s past purchases search is one example of this trend. When I search for the two William James books I bought twenty years ago, I get 59 returns. Apparently, if a book includes the word “William” or the word “James” on any page, it is tossed onto the heap of returns. The two books I am looking for — the only two written by William James — are seven pages in. These are, in fact, the last two items in the return.
GoodNotes is another example. When I search, I am usually looking for a particular passage in a PDF. But my passages are disassembled into individual words, and each word is blurred to indistinction. Goodnotes tells me every single page where some wildly loose interpretation of any one of the words appears. Surprise! It could be anywhere! But at least I can feel secure that what I am looking for is somewhere in the results.
I’m sure that soon every product manager lemming will follow this shiny new best practice directly over the cliff.
They could save a lot of trouble just by providing an unfiltered all-inclusive list. They could even omit the text field and leave the search button. Click it and it searches for “*”. If the priority is avoiding any possible match, just including everything is a foolproof better-safe-than-sorry strategy. Best of all, it is quick and cheap to implement.
As far back as I can remember, Google search has had a “I’m Feeling Lucky” button that works exactly this way. Everyone took it as routine Silicon Valley whimsy, but now it is clear: “I’m Feeling Lucky” was the future of search.