Web Search Training Wheels: A Play for Precision

August 10, 2017

I read “How to Instantly Boost the Accuracy of Search Results on Google and Bing.” i love the word “instantly”, particularly when coupled to “accuracy.” The write up describes an overlay called Advangle, which helps a person create a search with more than 2.6 words. Interesting neologism Advangle.

These services are what I call “training wheels.” The idea is that a person looking for information fills in a form, which helps the person create a query more sophisticated than “pizza.” Many systems in the last 50 years have tried these types of interfaces. In fact, one can find them in the whiz bang interfaces available to cyber OSINT software users. I won’t drag the old Dow Jones interface into this post, nor will I provide screenshots of Palantir Gotham interfaces. (Hey, you probably know about these already.)

The write up, however, does not explore the concept in too much detail. I noted this statement:

The Advantage interface makes it easier to string together targeted searches with the right syntax, and in half the time it would take to type it all out by hand.

Saving time, not prediction or recall, is the unique selling proposition.

It is useful to keep in mind that formal search operators are still available to users of Bing, Google, Yandex, and a number of other systems. The problem is that as Web search has massified, a tiny faction of the users of ad supported Web search systems bother with formal operators like filetype: or other oddities.

The real problems with search are far deeper than an interface overlay. Let me highlight several which I find consistently troublesome:

  1. Finding a way to impart the skills of well executed reference interview conducted by an expert in online search and retrieval. (Marydee Ojala, Ruth Patel, Anne Mintz, Ulla de Stricker, and Barbara Quint are individuals who can help a PhD formulate a statement of what information and data are needed, convert that desire into appropriate queries of appropriate databases, and deliver a filtered list of results.) Software, no matter how nifty the interface, at this time cannot replicate this expertise.
  2. Individuals who need information are more crippled than their counterparts from 30 years ago. Online systems have worked hard to let popularity and past user behavior provide a context for a query like “cyrus.” If you think you will get the pop star before a long dead historical figure, you are more sophisticated than the eager consumers of pop up ads on a Pixel phone
  3. Databases are governed by editorial policies. In the good old days of 1975, creators of databases figured out what and how to index. Today most users believe that Google has “all” the world’s information. Nothing could be more wrong headed. Indexes, particularly free ones, include what creates traffic. If the content gets a little too frisky, censorship, filtering, and smart / predictive software steps in and delivers “better” information.

I suggest you give the Advantage service a try. You may find that it is better than a room stuffed with Quints and Ojalas and others of this ilk.

My approach is simple: Know what one wants. Formulate a suitable query. Pass the query across the sources/databases likely to have indexed the information. Review the results. Think about the information gaps. Repeat the process.

Pretty crazy today, right?

Who has time to figure out what companies are in the cyber OSINT business or what Dark Web sites continue to offer contraband in the wake of AlphaBay and Hansa.

Research via digital resources, unlike checking Facebook, is a bit of a mental workout.

On the other hand, why not let the ad supported search engines deliver exactly what they think you need. Better yet, let these outfits provide that information before you know you need it.

A system that actually delivered precise, on point, timely, and authoritative results would be great. It would be nice to be able to live forever and travel to the stars.

Reality is a tad different. UX is not yet a replacement for knowing how to research in a way that moves beyond finding Game of Thrones.

Stephen E Arnold, August 10, 2017


One Response to “Web Search Training Wheels: A Play for Precision”

  1. Marydee Ojala on August 13th, 2017 6:36 am

    Well said, Steve, thanks. One slight correction, it’s Ruth Pagell not Ruth Patel.

  • Archives

  • Recent Posts

  • Meta