∴ikura

Moving Money Among Friends … with ikura.

feb 1

Lament for search

Looking for technical solutions via search-engines has never been harder. And while I can understand a type of Gresham’s Law is at play whereby bad answers can push out good answers to questions on Google, is it wrong of me to assume the richness in sheer volume of data ought make search increasingly better with every passing day, mitigating all kinds of the search-algorithm niggles?

It’s hard to put a finger on it, but search seems to be suffering from what I call a Norm Paradox. It’s a problem of perception, really, and one that search & suggestion algorithms/AI should well take account of.

Let me illustrate: one would think that two symmetrical choices, take that of Coke and Pepsi, are perfect substitutes for each other, so that when it’s time to correlate a weighted answer, Pepsi could stand in for Coke as yet another viable suggestion. But this is wrong. As Norm Macdonald once joked about it, if you love Coke, then your least favorite beverage is Pepsi.

Lost in the weeds

I started using Google when it was popular among the programming community as a way to bypass manuals and get a popular hack for a problem. The solutions that floated around were at best good leads, but more times than not, led nowhere. This was before the Google engineer Salar Kamangar and friends put ‘adwords’ in place (although it’s almost impossible to remember a time before several ads in my search result weren’t both prepended and appended to a given search yield).

Google search was hard to use then, and it’s hard to use now, but for different reasons. In plain speak, it seems like there should be no excuse for a shoddy search experience from the ‘titan of data.’

The Norm Paradox is fervent when the work-day is over and I turn to Youtube to waste some time. The suggestion-algorithms all suffer from thinking that Coke lovers would equally prefer Pepsi. Again, there couldn’t be anything I’d prefer less.

Local vs. online

Search only stinks online. Locally, if one needs to find a file, there’s find . | grep foo ; if I need to find text within some files, again, there’s grep. A man page has all the answers for a tool that’s new to me, and a book or white-paper is an even more penetrating resource. I seem to be clamoring to these antiquated methods more & more the ‘better’ technology gets. I find this also to be paradoxical, and perhaps a little sad.

When I think of online search, I get anxious. Trying out bad solutions after hours of perusing search pages has taken so much of my life. It’s become an out-of-body experience lately: to be ignorant of something, and then having to turn to Google. I watch as I reach for the mouse to open up Chrome. The anxiety is palpable. How long will it take to find the answer, this time? I ask. Will I swear out loud and scare my neighbors after my willpower reaches zero, like yesterday?

Long live search

Thankfully, search has gotten me out of some pickles in the past, so I can’t be too negative about it. I just think it’s retrograde at this point in spite of what I would imagine should be a search golden-period. But with eight ads per page, and search accuracy feeling like the year 1999, my preferences are turning increasingly to the methods used circa 1969 to get anything done any longer.

THIS END UPWhat’s ikura?