Columns

Perspective on XML: Steady steps spell success with Google

• By Uche Ogbuji
• July 1, 2004

Google has always been a marvel to those of us whose living revolves around high technology. When the beta emerged, we quickly passed our colleagues the tip about this new search engine with the mercifully simple front end and the uncanny ability to return the most relevant results for most searches. As time went on, and Google became an indispensable tool in our trade, we began to clue in our less-technical acquaintances. Google is a technology even our grandmothers can use. Like an oracle, you express what you seek in simple terms, never minding the underlying magic; unlike an oracle, however, you don’t have to bring an offering before you’re seen, and the answers are generally informative and not couched in puzzles.

The media eventually picked up on the buzz as journalists discovered the power of Google, as well as irresistible stories of people doing gumshoe work on potential dates using the omniscient Google. The name “google” became a generic verb, the surest sign of zeitgeist possession, and has become a top brand with very little traditional marketing.

Feeding on all this interest and admiration, Wall Street came along to work its unique specialty: generating true mania. Googlemania is now the center of hope for an end to the high-technology slowdown. As with all manias, this one has a hint of the irrational about it; but among developers, it is the very appreciation of the underlying magic that offers hope that the mania is somewhat less irrational than some of its predecessors.

Indeed, there are important lessons developers can glean from the success of Google. It so happens that I am often called upon to offer solemn advice to my fellow developers. Since I’m human, such sermonizing usually emphasizes those things I think I do well and that perhaps others do not. What stands out for me about lessons from Google is that in most of them I know that I, myself, often fall short.

* User interfaces don’t have to be elaborate. Lately it seems the conventional wisdom is that simple HTML Web pages aren’t good enough and that slick, animated confections are the future. There is certainly true in the case of information dashboards, where reams of information are arrayed on a screen. To be fair, simplicity is advocated even among devotees of the dashboard (and I am too often one of them), but this is too often lip service through ingenious dashboard layout.

When Google emerged, established search engines were trying to expand into the promised land of portals, pushing search engine interfaces into the corner of pages festooned with ads, tickers, articles, icons and lists that were all ostensibly “personalized.” In comparison, the Google user interface was a refreshing change. Rather than having editors and algorithms push volumes of data to the browser in the hopes that the user would serendipitously extract valuable data, Google is completely given over to the idea that users pull just what they need at the time, with minimum clutter and distraction. It uses plenty of white space, search results are largely undecorated and ads are discreet text boxes. Even as Google has added features, from Usenet and image searching through e-commerce comparison, the emphasis on the simple interface remains and is one of Google’s most admired features.

It is extremely difficult to channel computer power through a simple front end, and sometimes elaboration is inevitable, but Google reinforces the lesson that even the most powerful technology can sport a minimalist interface.

* Build on legacy rather than on writing it off. I have long been an advocate of technologies -- from XML through the Semantic Web -- that would make it easier to search and process information by more clearly expressing its structure and context. The problem is that creating a critical mass of such material would require a tremendous evolution in tools and discipline -- certainly an ambitious vision. Google realizes a respectable cross-section of the promise of the XML Web generation by merely finding creative ways of harnessing the mountain of legacy from the original Web.

* Deployed technology should be transparent. Despite our famous advocacy of openness, many developers do not practice what they preach. Whether because of intellectual property or security concerns, or just out of vain secretiveness, we often seek to mask the heart of our creations behind sophisticated front ends and APIs that sometimes make it very hard for other developers to build on what we’ve done. Google has built an important culture around developer empowerment, and the popularity of “google hacks” has been another important factor in the search engine’s ascendancy. Certainly Google is so popular among developers that few things generate as much excitement as a mere hint that the company might be hiring. Google shows that when a technology creates a commons of innovation, any loss in control is far offset by gains in value. Many have stated this lesson. Few actually apply it. Google provides an important reminder.

One of the many remarkable things about Google is that it is more an example of sensible problem solving than miraculous innovation. Google took the time to create text search techniques, and created an optimizer based on carefully chosen meta data. Most truly important breakthroughs come from just such modest but practical increments. Google gives us an idea of how much of an effect we can have when we maintain disciplined focus in applying the best principles of our profession to clearly defined problems.