KISS my Searchbox

“Give instructions only to those people who seek knowledge after they have discovered their ignorance.”
 — Confucius

Blind, unquestioning devotion (or opposition) to just about anything can be dangerous. Look no further than the atrocities committed under the “my country, right or wrong” mindset, or even this bizarre logic of career Republicans who allegedly oppose seemingly everything Donald Trump stands for yet endorse him nonetheless. Continual assessment of what everyone’s doing and why, as long as we’re not stubbornly spinning our wheels by objecting to embarrassing evidence, is a vital part of improving the human condition.

I like to look at things and ask, “if this didn’t exist, would we invent or enact it in a similar fashion?” One critique of various dogmas is a thought experiment involving the eradication of all religious doctrine and scientific discoveries. While it’s easy to believe that we would certainly come up with the same laws of physics we have today, the re-creation of more mystical ideologies in their present form is questionable. Gnostic disputes aside, in a work environment, I find it useful to consider our current services and workflow under this sort of lens: why are we continuing these practices?

Change is threatening. It challenges us to think, and also shines a light on our previously flawed systems. Not that it isn’t usually a trivial matter to cherry pick defects with new things either. Spell check and autocorrect can, at times, get things wrong. However, the overall efficiency of using these tools, compared to looking up each word in a dictionary, hardly means we should scrap them altogether. They may even work better and perform more reliably than their human-powered antecedents. Yet the problem some people see is that such labor-saving technologies are making us lazy, shallow, and stupid.

I am totally unskilled at driving a stagecoach. People in the future may be similarly inept at operating a manual transmission vehicle. So what? We shouldn’t feel compelled to teach long division and cursive and dead languages and manual citation formatting any more than it helps students accomplish present-day tasks. As a librarian, perhaps especially one with a library school on campus, I think to indoctrinate students with tried and true yet now historical research methods is detrimental to education.

Search engine technology has evolved considerably in recent years. One particular aspect of newer designs is the Googlization of interfaces by moving many if not all advanced functions from the initial screen to the results page. What was once handled by a slew of drop-down menus and even esoteric operators is now available in the form of post-search delimiters. Many of my colleagues absolutely loathe how this new means to and end is inexplicably both excessively dumbing down the search process while also making it so difficult that even skilled and sophisticated researchers purportedly cannot figure out how to get what they want.

I’m a fan of the simple, single search box, not just as a utopian concept or gateway, but something we already need to market and support.

  • It’s easier to use
    Mapping out search statements, although a tad simpler than navigating a subject card catalog or command line interface, was necessary only due to the constraints of how earlier systems operated. Research topics are not rigidly isolated terms anyway, so thinking of search parameters as algebraic variables is an unnecessary restriction. Beyond aesthetics, it’s hard to overstate the usability gains compared to earlier platforms.
  • It’s more efficient
    An unlimited search will work just fine in most cases, and maybe even in situations where people had been selecting provisions they didn’t really need to. Even if you are looking for a known item or something in a specific location or format, you can usually obtain it through an “Everything” search. Failing that, you can still further refine results if and only if necessary. It therefore saves the time of the user.
  • It’s more powerful
    You can readily do a lot more with multiple facet filters and other nested conditions and sortings than is possible with a text-only query, no matter how convoluted you learn to construct it.
  • You can even get better results
    Open-ended searches are a better way to fully explore a topic than immediately honing in on precise terms and parameters. Forcing or encouraging preemptive restrictions, even those hallowed Boolean operators, may override better mechanisms such as fuzzy pattern matching and other behind the scenes algorithms which provide more optimal results.

As for the “but not everything’s in there!” gripe, sure. However, which portion of the following pie chart should we be more focused on?

If we were to design a search engine today from scratch, would we find it necessary to adopt all of the historical precedents of how telnet catalogs functioned, or would it be better to start off with what can be accomplished via a modern web browser? That’s one of the reasons Google was such a success. It was not only easy to use, it delivered good results, as well as the convenience of instant full-text. It pioneered relevancy ranking by popularity (sorta), was the first major search engine to start automatically ANDing terms together, and continued to innovate with features such as autocomplete suggestions and faceted results. Libraries are continuing to play catch-up with Google as we begin offering more personalized results, recommendation engines, and linked data.

And by the way, you can’t even connect to the advanced search page from google.com anymore because it’s by and large a depreciated method for finding what you need. This is a clear case of how improving technology to make things more accessible is not so much dumbing down archaic but convoluted methodologies as it is enabling and empowering unmediated research. So it’s no wonder that any profession is openly hostile to inventions which make them a little more obsolete. The real threat here isn’t to anything but our traditional library instruction model.

Further Reading

Check out my other posts for related commentary.