Technical debt is increased with C/C++, says research

Rsearch suggests that the public sector is significantly less able to switch applications than the private sector

Technical debt, which is the amount of money firms need to invest to bring their IT systems up to scratch, has long been a burden for IT departments, but with a decade of tight budgets behind us and a few further years of austerity ahead, the amount of debt residing in IT departments globally may rise to $1tn by 2015 from $500bn today, according to analyst firm Gartner.

Gartner's more specific definition of technical debt is as follows: it is the cost of clearing the backlog of maintenance that would be required to bring corporate applications portfolio to a fully supported current release state.

Gartner analysts said that this "IT debt" is a hidden risk for many organisations, and given continued tight economic conditions, this IT debt is growing, meaning the associated business risk is also growing.

This IT debt can also be a problem when companies or organisations want to switch from one software system to another, and a report called Worldwide Application Software Quality Study, from software analysis firm Cast, argues just this.

The research suggests that the public sector is significantly less able to switch applications than the private sector because of high levels of technical debt (the result of not maintaining or managing software systems adequately). The insurance sector is the second least able to change because of technical debt resulting from problems with legacy infrastructures.

The research from Cast delineates technical debt according to the type of coding language used within a company’s software.

The research states that the although the average technical debt per 1,000 lines of code is about $2,819, the violations present will obviously vary considerably according to the code used and therefore affect the amount that needs to be invested to bring it up to scratch.

Unsurprisingly, when ABAP, Cobol, .Net, J2EE and C/C++ were compared, SAP's proprietary code ABAP returned the fewest violations per lines of code at 20 per 1,000 lines. This compares with 438 for C or C++.

However, Jitendra Subramanyam, director of product strategy and research for the firm, said: “Companies need to work out whether they want to tie themselves into a proprietary programs coded on ABAP for example, which are not easy to modify and are pretty rigid in terms of what they have been programmed to do, and a more flexible programming language such as C/C++.”

In other areas of the study, Cast found that Cobol, a language used widely in the financial sector, scored significantly better in terms of security benefits than the other languages. While newer languages Java and .Net scored lowest for performance.