TurboCASH debates moving to Linux

For small businesses, the move to a free desktop is hampered by the lack of viable accounting software. Whatever the benefits of switching away from Windows, most businesses can’t ignore the convenience, management information, and reduced accountants’ fees they enjoy by using the likes of Sage or QuickBooks. While free software fans are able to restrict their software choices in the pursuit of ideological purity, the typical small business owner has to be more pragmatic.

Today there is at least one top-tier open source accounting package. TurboCASH, from South Africa’s Pink Software, claims superior functionality to its closed source rivals, is licensed under the GPL, and has a user base of more than 30,000. It even integrates directly with the OSCommerce and Zen Cart ecommerce applications. However, as a Windows-only application, it hasn’t solved the free software desktop problem for small business — until now.

Since 2003, when TurboCASH’s project leader, Philip Copeman, made the decision to re-licence under the GPL, he has been a vocal proponent of the open source model. Until recently, though, he had been adamant that the only market for TurboCASH is with Windows users, saying, “While Linux has, say, a 5% desktop penetration, in terms of accountants this drops to near zero.”

However, following near daily requests for a Linux version of TurboCASH, Copeman now believes that Linux could be the key to his software’s future success.

Open source makes practical sense

Every decision that Copeman makes regarding TurboCASH is for practical reasons. He sees open source as a sensible development model, with any ideological considerations merely an added bonus. Realising that this make him less than popular with free software purists, he reconciles TurboCASH’s proprietary development platform — Delphi — and its Windows operating system with pragmatism, saying “Being pure open source is not as important to us as giving our users a great program.”

Lower-end alternatives
GnuCash‘s Quicken-like functionality is ideal for home finance, and basic double-entry accounting is available from SQL-Ledger. Neither project, though, has the same level of functionality found in their proprietary rivals.

However, that same commitment to expediency is now prompting Copeman to investigate free software alternatives. He realises that Delphi’s licence fees are a barrier to new developers and tie TurboCASH’s future to another company’s strategy. Similarly, he believes the persistent march of Linux offers the opportunity to double TurboCASH’s user base. “Right now, the Linux accounting space is fallow; it is inviting someone to put down five million South African rand (around $785,000) and take the position. We estimate that a million users would change to a Linux system if they had the choice.”

Re-engineering a Delphi application for Linux presents considerable challenges. Borland no longer supports Kylix, its Linux implementation of Delphi, and the closest open source alternative, FreePascal.org’s Lazarus, is not compatible with the Borland Database Engine, on which TurboCASH relies. Consequently, Copeman sees the new Linux project as an opportunity to start entirely from scratch. Major architectural choices are up for reconsideration:

• Development language and environment
• Database engine
• Thin versus fat client model

The debates surrounding these decisions sum up the major problem that the new TurboCASH faces: its 20- year legacy is both its greatest asset and its main liability. For example, it’s unlikely that a brand new project would have a debate over FreePascal/Lazarus or Python. Python is the language of the moment, particularly in GNOME circles. Pascal is, arguably, the great also-ran of programming languages. However, TurboCASH has to weigh the benefits of any new language against the potential for alienating its existing development team.

The mechanics of implementing accountancy software, though, are almost incidental to gathering data on the tax and accounting legislation in various territories, building trust that your software can be relied upon for such a business-critical role, and developing the paid-for support network essential to such an application’s success.

That’s why Copeman would be happy to outsource the initial development work, if he can find a suitable partner. Despite his faith in the potential of Linux, he realises that his small team of developers doesn’t have the capacity to maintain two entirely separate development trees. With two years’ experience of running a large GPLed project, neither does he suffer from the novice’s delusion that open source equals endless offers of free development time.

How many more TurboCASHes are there?

Perhaps TurboCASH’s move to Linux is of greater significance than bringing us a step closer to the much-heralded “year of the Linux desktop.”

The free software landscape has been developed by enthusiasts, who were then joined by large corporations looking to find a home in an otherwise monopoly marketplace. Now that small software houses, with no prior involvement in free software, such as TurboCASH’s Pink Software, are choosing open source development for the practical benefits it gives them, does it represent a groundswell of mainstream acceptance?

Philip Copeman now sees open source as the only way for his software to compete with his larger, aggressive competitors. As more companies join the open source community, for pragmatic rather than ideological reasons, it will be fascinating to see how we welcome them, and how they reinvigorate us.

Matthew Revell is a technical author from Wolverhampton in the UK.