The other day, I was preparing a quickie assessment of the biggest office suppliers in the US, and I thought “Instead of trying Google or Wikipedia, I’ll use this newfangled Wolfram Alpha thingy that’s out now.” And after approaching that query, along with some other spur-of-the-moment questions, I’ve finally realized what Wolfram Alpha is.
Wolfram Alpha is a great way to find a fact and do some analysis about it. You might think that after 19 days of being available to the public that I would understand what it’s all about as a user. Just like Mathematica, I think this product is a tad ahead of its time. This quote from their site: “Wolfram|Alpha’s long-term goal is to make all systematic knowledge immediately computable and accessible to everyone” is now much more meaningful after having used the service: “immediately computable” means a whole world of analysis can be at your fingertips, and “systematic knowledge” is another phrase for data and facts.
Now, there are some recognized shortcomings but by-and-large, as a proof-of-concept, it’s fantastic! The above link should bring you to a query state that lists the following major US specialty retailers: Staples, Office Depot, and OfficeMax. Now what I was trying to do was to find information quickly about this particular sector of retailers, provide a basis to generate an estimate on the size of the market for office supplies, and see data points relative to each other. And of course, it did so beautifully.
There’s plenty more work to follow, however. I think there are two key high-priority enhancement projects Wolfram Alpha should consider for their next release cycle:
- a simplified and standardized (mathematical?) grammar structure, and
- a fact-drilling capability.
The first suggestion would greatly help users understand how to structure their input and reduce the incidence of the “don’t know how to handle the input” page appearing to users. After all, I can’t think of a single language that doesn’t have syntax with which to standardize expressions and context. And at the same time, I don’t think one system can understand every single kind of expression and context from a myraid of users without extensive effort into a (in my opinion) futile effort.
The second suggestion would be a means to offer transparency to the facts presented by the engine. Similar to Wikipedia and to the tradition of citations in written works, Wolfram Alpha should consider a way to cite every single fact in its database. After all, each one should be verifiable in its own way, and the source content would go a long ways to confirming that work. Good luck to the Wolfram Alpha team, I wish you the best of success!
Recommendations: Put in a grammar/syntax structure to help users understand how to input their queries, while keeping it at a minimum. And don’t forget the beauty of the Web being built on hyperlinks, so link to your facts where possible!