It is known to all that the use of Spreadsheets, such as the current Excel, since the beginning when VisiCalc appeared in 1979, have allowed many users without programming knowledge to be able to develop many and varied applications.

In particular, the world of economics has been empowered by Spreadsheets. An accountant can directly carry out all or almost all of his work using them, since from the beginning they were conceived with a varied communication with other data processing systems.

At this time, using Excel, data can be taken or linked from practically any source, including the cloud, and very varied results can be passed to other applications where they continue to be processed or are presented on screen or paper.

Now, the basic programming model of Spreadsheets is anonymous, that is, it is not "named" since letters (columns) and numbers (rows) are used to locate any box and there are many integrated procedures that even using automatic wizards using this convention. Take for example the sum calculated from a number of rows above the result "sum".

All Spreadsheets use this convention of "naming" with letters and numbers, which is its great advantage but at the same time its Achilles heel, we'll see why.

Another characteristic that distances Spreadsheets from the relational model of Codd, so widely used today for databases, is that in the relational model each column has a unique name and can be of a single data type, while in the Spreadsheets, even each cell can have data of various types and even results of formulas or access to external data, which is superior.

The latter makes most programmers think that "it is not serious" to use Spreadsheets for data, that you have to use Databases as they are currently developed, instead of thinking that what The only thing to do is to develop a new database model, even a mathematical one, that reflects the wide spectrum of Spreadsheets.

Also, most programmers think that "it is not serious" to program in Spreadsheets, that what should be used are program development systems and use languages like Java or C# to give two examples.

But what happens? Well, it takes more time and effort to develop systems, for example, for the entire economic management of a company, than to do it directly in Excel. Even if a company has internally developed a system for such purposes or uses a purchased one, in any case life leads to developing many things in Excel that are not contemplated in the already implemented system, especially views and reports.

All this means that many of the economic results used in decision-making and even for the payment of payroll, products, and services, are made or supplemented in the Spreadsheets, and the risks that this implies is widely documented. Take for example the publications of the European Spreadsheet Risk Interest Group (EuSpRIG).

And why does something so useful cause so much risk? Well, for the simple reason that it is not totally necessary and essential to "name" data and results, which prevents even the effective use of "Open Source" in application development.

For example, Javascript, which does not have a data "type" for the use of variables and variables can take any value, has achieved that through the "Open Source" collaboration of thousands and thousands of programmers there are totally effective systems, made all of the internet, and there are no "collisions" of using variable names for different purposes while being used for various intercommunicating libraries.

It seems simple but in reality it is very complex, so complex to obtain something similar without "Open Source" that even a powerful firm like Microsoft is adopting this form of application development, even very large ones.

What to do then?. Well, basic and applied research that leads to the development of Spreadsheets "naming" what is accessible and what has been accessed, such as the relational one developed by Codd, and from there, without ceasing to use the advantages of the visual and varied programming that currently provides Spreadsheets, develop new Spreadsheets even fully integrated into the world of data so connected that already exists.

Thank you for reading,

Octavio Báez Hidalgo.

No thoughts on “Spreadsheets are limited because ... and serious errors can occur.”

Leave your comment

In reply to Some User