Pages

Friday, March 6, 2015

The Status of Genealogical Searching

CZmarlin — Christopher Ziemnowicz on Wikipedia 1959_Rambler_American_2dr-sedan_Blue-NJ.jpg
The core of the huge online genealogical database programs is their ability to find a particular name among billions of possible matches. Of course, part of the burden of finding those names depends on the skill of the person using the program. There are certain specific things that any knowledgeable researcher can do to increase their success rate. But just as a more skilled race car driver can win a close race, the driver needs to have a superior machine to win consistently. The researcher's skill can only go so far without a really good program to support his or her efforts.

Likewise, continuing the race car analogy, a poor driver cannot properly control a powerful race car. If there are those who dream about driving a car in an important international race, they would probably find out that they lacked the skills and judgment to survive in that rarified racing circuit. I say this knowing full well that here in Utah Valley, I am surrounded by frustrated wanna-be race car drivers.

Genealogists who approach the various search engines are confronted with a panoply of online genealogical databases each with its own "search engine." Here I can carry the car analogy one more step. What we see is a selection of search engines that vary in their speed, accuracy and utility as much as any selection of new cars. I remember when our family owned a Rambler American (sometimes known as a "Nash Rambler" as in the song). I drove it to school every day across Phoenix and then home again. The most remarkable thing about that car was the fact that with its manual transmission, its top speed in first gear was about 15 mph and top speed in third gear was a whopping 35 mph. I remember that it took almost two blocks of flooring the gas pedal to get up to the speed limit of 25 mph. There are very few steep hills in Phoenix and so I never did learn how fast this car could go downhill, but I can say, that we didn't own the car for very long. I fully realize that the car improved dramatically and much later, I was very happy to own an American Motors Javelin, which I drove while attending the University of Utah and for two full years in Panama, until, many years later after returning to Phoenix, the car almost disintegrated from use.

What has this got to do with genealogy? Well, a whole lot. Search engines in genealogy programs (and others also) are called engines because speed and performance are vital factors in the user experience. But speed and performance are not enough. The search engine needs to be accurate also. Developers and the engineers who work on these programs try to balance the three qualities to give a positive user experience. But just as with cars, some developers value economy over performance and some never do get their search engine out of the parking lot and onto the highway.

Now, it is not my intention to rank the current crop of genealogical search engines since any evaluation I would make would be based on my "race car" mentality and may not match your economy minded needs at all. What is interesting about cars is that I no longer have to buy a specialized, souped-up model to go really fast. My Subaru Outback does really well at the 80 mph legal speed limit on some of the Utah freeways. There is a point where a car is "good enough" even to satisfy a prima donna. There is also a point at which the genealogical search engines satisfy almost everyone who uses them.

Unfortunately, many genealogical search engines are living in the late 1950s as far as performance is concerned. I most cases, I am getting Rambler-like operation instead of what I would expect today. Why is that the case? Back to the cars. When I was driving that Rambler American, I had friends whose families drove Jaguar XKEs and Corvettes. I knew what fast was. Today, I am confronted with the same situation. We have the Jaguars and Corvettes of the genealogy search engine world and we have the Triumph Heralds. Oh, I didn't mention that my next car was a Triumph Herald. The top speed in first gear was just over 5 mph. It had a four speed manual transmission. It was a cute car, but entirely gutless. I remember going up steep hills in central Arizona and being passed by fully loaded semi-trucks.

Now a word about reality. Most genealogists are entirely frustrated when they search for Uncle Fred Jones in Missouri and get Fred Brown is South Africa. It is the perception of the user that the search engine is broken and does not work. From the engineers standpoint, they are being successful because the "search engine" has returned a "partial match" ranked according to their pre-set criteria. The problem is that the user has no idea what that pre-set criteria is. To mollify the user, the engineers then draw a line and say, we didn't find your match, but here is a lot of junk that might be what you are looking for.

But just like with cars, over the past few years I see glimmers of hope that the search engines will start to approach high performance. Presently, if they are allowed to do the driving, they can achieve high levels of performance and have finally entered the race circuit. Right now, we only have about three contestants (I'll let you guess who they are) but they all have moved far away from the slow car category. I am hoping that the rest of the industry takes notice and improves their own performance. Meanwhile, I am still looking for the Hennessey Venom GT of the genealogical community.

I will break my silence on specifics by mentioning the most improved genealogical search engine of the year award. That award goes to FamilySearch for their Record Hints and overall search engine performance. I laud Robert Kehrer and his team for their consistent and continued improvement in all categories. Well done.

1 comment: