Ideal, Simple, and Good Enough
September 1, 2009
I just read this Web page headline: “SAP NetWeaver Enterprise Search: Simple and Secure Access to Information”. Wow, simple and search pushed together like peanut butter and jelly, ham and eggs, and hammer and nail. The problem is the word “simple”. Who does not want simplicity? Life today is too complicated. Make it simple.
The three meta issues swirling around simple search and content processing have their roots in the fecund soil of user annoyance. Most users have zero clue about the more sophisticated features in any desktop or Web application. The evidence is not far to seek. Look at these three questions. How many can you answer without recourse to Google, your friendly power user, or digging through books in the ever smaller computer section of Barnes & Noble or Borders.
- How do you limit Google results to only those for US government and state information?
- How do you create a single, presentation quality graphic from Excel 2007?
- How do you delete unwanted colors in Framemaker 7.2 when you import a graphic format other than jpg?
The answer to the Google question is to navigate to Google.com, click on Advanced, scroll to the bottom of the page, and click on the Uncle Sam option.
The answer to the second question is to use a third party application from an outfit in France called GlobFX.
The answer to the Framemaker question is to open a version of the document with the correct color information. Go to File Import and select the option for importing a template. Make sure only the color information option is selected. Make the source the file with the “correct” color information and the target the file with the unwanted color information.
An even better example can be found in the usage of the advanced search functions for Web search systems. In general, users enter 2.3 words on average per query and fewer than five percent of search users access the advanced search functions.
Who cares?
I care a little bit, but not enough to give a talk about the way in which those creating systems make life almost unbearable for users. I am sufficiently motivated to define three terms and offer some comments.
Ideal
I find meetings in which requirements emerge from a group discussion. The focus jumps between a micro problem (“I can’t find my most recent version of this document”) to science fiction (“I need to see information from many sources on one screen so I don’t have to hunt or scan for the information I need”). Unless there is a method for capturing these requirements and assigning some meaningful tag for difficulty or cost to each, the exercise is interesting but often not super productive.
In my experience, folks like to talk about ideal features and functions. The chatter is similar to that I recall from my freshman class in Philosophy in 1962.
The problem is that when a vendor or a developer charts a course for the idea, the journey may be more expensive and time consuming than Odysseus’s return home from Troy.
When my team encounters a cost overrun and a system that is never completed, I think, “Ideal”.
Simple
An John Maeda’s The Laws of Simplicity (Simplicity: Design, Technology, Business, Life) marked for me the start of the “simple” trend. Consultants hopped on the bandwagon and began making complicated topics easy to understand. Consultants had to wait until a “real” scholar tackled the laws of simplicity. No consultant wants to follow in the footsteps of the Dummies series.
Learning math is a good example of one approach to complexity. For most people, each course is a building block. Math wizards don’t like this approach, but let’s focus on the majority of people in the group that uses only a handful of the features in Microsoft Word.
To be good at math requires that each course be taken, the concepts understood, and converted into principles that can be used like Lego blocks. Skip a step, and you ma have a tough time figuring out what is going on in the next course up the ladder.
Obviously, math when broken down and presented in well-shaped components can be “simple”. Confront a proof without the background, and you confront the complexity of mathematics.
The problem is that there are different types of simple. Making a concept “simple” requires quite a bit of intellectual work. Ignoring the rigor and the field testing means that yapping about simplicity confuses people. Here’s why: make a complex subject simple like enterprise search. Install a search system. Come back and survey the users and find out about the costs of the system. You may learn that the “simple solution” was a problem. Information access and knowledge centric systems are fairly complicated. The wrong type of “simple” creates issues. Smart consultants look less smart when the implications of simple are understood by the client.
Good Enough
The Google is the master of the good enough approach to online systems. Idealist and advocates of simplicity find themselves running aground on Google’s method. The “good enough” approach is quite powerful. For those seeking perfection, good enough can be maddening. I heard an expert on a Windows podcast make the point that Google Docs offered none of Word’s sophistication. Yep, that’s the point. “Good enough” can make use of quite sophisticated technology, but the application is designed to do some basic functions. Over time, more elaborate features can be added. “Good enough” permits fast cycle market probes. For those looking for simple solutions, “good enough” may be simple on the surface. But once one digs a bit deeper, the simplicity is a bit like stage make up. The actor can look quite different.
What about Search?
Enough of the philosophical meandering. How do these three notions relate to search. In my experience as an addled goose, my observations are:
First, make sure you know whether your chosen search and content processing vendor is in the simple category, the ideal, category, or the good enough category. You can eliminate costs, frustration, and needless engineering once you know what the vendor’s approach is.
Second, make sure you know what category you personally fit into. If you are into simple, do yourself a favor. License software that matches your view of what a search system should be. Buying Lego blocks when you need a search toaster is a guaranteed headache.
Finally, make sure you know what your user communities expect. If most users want one or two basic functions, you may find that “good enough” will deliver. A user who needs one type of information that is outside the function set of a simple solution can produce some fierce resistance in today’s business climate.
Do vendors know which category their systems occupy? Sure, but most vendors will use marketing razzle dazzle to get the deal. The buyer or licensee has to do the heavy lifting. Make sure you have the right concepts in your mental biceps before tackling the hard part of enterprise search.
Stephen Arnold, September 1, 2009
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
