Question about web-based search options? (Take II)

Subject: Question about web-based search options? (Take II)
From: Geoff Hart <ghart -at- videotron -dot- ca>
To: "TECHWR-L" <techwr-l -at- lists -dot- techwr-l -dot- com>
Date: Wed, 01 Dec 2004 12:05:50 -0500

Steven Brown picked up on the earlier thread about searches vs. indexes and tables of contents (hierarchies): <<I think the popularity and pervasiveness of the Yahoos and Googles of the world will almost require technical writers to rely more on search engines and less on a traditional, book-like structure and index.>>

I'll agree with you that future search engines may eventually become superior to traditional methods. However, modern search technologies have several crucial flaws that make them "better than nothing... sort of":

- They are not contextual, and are thus incapable of parsing the semantics of the content: even when you add half a dozen keywords to narrow down your search, a frequent primary keyword can generate thousands of unrelated hits.

- Most search engines cannot do a "sounds like" or "synonym" search, and very few can do stemming (looking for root words such as "print" when the user types "printing" or vice versa). This makes them useless to someone who doesn't know the precise term to search for--arguably true of most naive users, who are the majority of any audience.

- Search results provide little or no indication of context; it's rarely clear whether a "hit" is in the troubleshooting, reference, or overview section of a document, among other things. Without that context, you need to explore each hit to see whether the context is appropriate for what you're seeking.

- Search engines require extensive knowledge and training to use effectively. Few users understand Boolean searches (and, not, or, xor), few recogize that advanced options are available, and few know how to combine and exclude keywords to progressively refine their search.

These flaws explain why an index is always better than a concordance if you're seeking a specific topic and why a table of contents is always better than a search engine when it comes to understanding the organization of information and browsing to find an answer.

<<It's easy for us to rely on a TOC and index, because that is what we know and probably prefer to use ourselves.>>

No, we use them because they're more effective than any current alternative. Both have been used for centuries, if not millennia, and represent the result of centuries of evolution. They're not perfect, but they're about as good as it gets. Give search engines another decade or two and they may become as well developed, but right now they're still "fish with legs crawling out of scummy ponds and trying to conquer the Earth". <g> As the saying goes, the important point is not that the bear dances badly, but that it dances at all.

<<At the end of the day, I think most users will tend to use a search engine to find information before they'll refer to a TOC or use an index.>>

On the contrary, research by Jared Spool and others have shown that search engines can frustrate users to the point that they abandon the search engine--and sometimes the entire Web site--in favor of random browsing or a call to tech support. Many users start with the search engine, because they understand that if they get lucky, they avoid a long and possibly fruitless search through masses of disorganized information. But success rates are often much higher with well-designed indexes.

You noted that indexes are difficult and expensive to do well. The same logic applies to technical writing, but I doubt you'd suggest we should dispense with technical writers and leave the writing to the engineers. Quality is always expensive. Information is inherently difficult to create and organize--that's why we have professionals to do the job. We can create that quality for less than it would cost for others to do the job.

<<The last time I sat through usability sessions, most of our users couldn't even find the online help much less navigate the TOC or index. It was downright painful to watch!>>

I'm glad you reported this: it's another data point to support my recommendation that "how to use the help system" should be part of any modern software manual. I've found this in my own work, but more data points are always welcome.

--Geoff Hart ghart -at- videotron -dot- ca
(try geoffhart -at- mac -dot- com if you don't get a reply)



RoboHelp X5 is a giant leap forward in Help authoring technology, featuring all new Word 2003 support, Content Management, Multi-Author support, PDF and XML support and much more! View an online demo:

You are currently subscribed to techwr-l as:
archiver -at- techwr-l -dot- com
To unsubscribe send a blank email to leave-techwr-l-obscured -at- lists -dot- techwr-l -dot- com
Send administrative questions to lisa -at- techwr-l -dot- com -dot- Visit for more resources and info.

Re: Question about web-based search options?: From: Steven Brown

Previous by Author: Re: Help! fonts in WinXP, Acro6
Next by Author: Usability study?
Previous by Thread: Re: Question about web-based search options?
Next by Thread: "Printer-friendly" web pages

What this post helpful? Share it with friends and colleagues:

Sponsored Ads