Spreadsheet and Database

80% of the world business run on Excel. At least true for me. After years of developments, Excel has grown from one of minor opponents of Lotus 1-2-3 to the most popular business software in the world. Being challenged by online spreadsheet suites, the omnipotent software is absorbing more functionality into its backstage, aiming to provide more complex features for business data users. As a common Excel user, who only takes advantage of limited set of Excel power, I started to use it as a temporary substitute to Word, to make a curriculum table for new semester. The greatest finding was that Excel can generate arrays like Monday, Tuesday, …, Sunday through simple drag and drop. But I was confused that why Microsoft made such a big thing only to make tables: Word can do that, and maybe even better. If you only use it occasionally like I used to, building small datasets and performing little calculation and analysis work, you will never experience the true nature of a spreadsheet. Firstly, it’s visual. What you see is what you get. People don’t have to run any reports to see what’s going on inside the box. Secondly, it’s lightweight. My computer at work is a Dell laptop with 512M RAM running Windows 2000. The CPU could be Pentium III, loading Windows in more than 10 minutes every time. But Excel runs just as fast as Freecell (which could be another good name for spreadsheet application). It loads a 500K file in 5 seconds and executes almost all calculation in no time. Such a lightweight application can hold thousands of hundreds of data, up to 256 worksheets at the same time. And there is even a built-in script language (though it’s not so visual as how it’s named). It’s amazing, isn’t it? Thirdly, it’s extensible. This is a relatively high-end feature. The VBA framework behind the scene provides Excel good extensibility to all types of applications, even in the future. You can export and import data to and from Excel easily. This, along with Microsoft’s market share, makes Excel the in-fact data center of the world business.

I use Oracle 11 at work. What should I say? It’s big, ugly and slow. With Java as its interface language, the application runs at an incredibly low speed. I thought it might be safer. I was wrong. It’s full of errors, reasons not clear. The buggy, sluggish stuff cannot even export data in a considerably flexible way. People have to query manually and copy/paste one by one…

What people invent database for? It should be the center of data collection, storage, processing, analysis and reporting. Built on the steady base of network, database guarantees the security and proper accessibility of data. In my case, Oracle does this job well, although at the cost of performance. However, Oracle cannot do analysis. Users have to export the data to specific analytical software (for most cases, Excel). Unfortunately, the export format is not XLS but plain text. Excel expert wrote a macro to convert the plain text report into XLS format, which saves some time for analysts. Another tragedy happens here: the macro poorly designed, users have to check the converted report manually. If we all do it manually, why should we invest money on IT?

Spreadsheet should be also database, or database should take a well-designed spreadsheet UI. That’s what we expect, but might also be the beginning of mass unemployment of so-called analysts.