A TurboCASH roadmap for Linux

A TurboCASH roadmap for Linux

11 September 2003

Mission

Do not lose sight of the fact that TurboCASH is a General ledger and not a front end workflow or ERP system. This applies no matter how big we get. In an enterprise world we would be a JD Edwards, not a SAP. It is our job to talk to all the workflow and ERP systems, but in the end we file transactions into accounts and report on them.

All the plugins and partnerships is what gets us the enterprise stuff, but that is totally variable. What always stays the same in every company is that they all have a ledger in the back room. So while we will design a generalised stock and invoicing system. This is going to fall short of everyones expectations and in fact if they require specialised or complex systems, these will best be written outside of TurboCASH.

Potholes

The biggest problem that we currently have is that: 1) the absolute vast majority of users in this end of the market use Windows. They are not going to take easily to moving their operating systems to Linux. That means that the vast majority of revenue (the juice that sustains developers) is going to have to come from installing TurboCASH and TurboCASH plugins on Windows systems; 2) TurboCASH on Linux is going to take years to get to the same quality level as the current Windows TurboCASH.

We need interim steps to make this sustainable.

WINE as an Interim

WINE may represent a means of quickly carrying the TurboCASH logic to Linux. Turbo CASH intentionally does not use any XP code, and may therefore make the carry over to WINE on Linux. At this stage we are uncertain of the outcome of this route. Success here will not bring us a final solution but should buy about 5 years, enough time to develop a native Linux product (that can also run on Windows). The key to this is whether WINE will support Paradox and the BDE. These two technologies which underly TurboCASH are essentially stunted and unable to make the move to Linux.

MySQL as Database

The decision to use MYSQL is really one of momentum; Windows/Linux interoperability and GPL licensing. I can debate the demerits of alternatives in another document.

Programming language

By staying with Delphi in the Windows environment we will be able to retain much of the Business logic of the current TurboCASH (20 man years of work).

Whatever we use on Linux, must hold the advantage of being able to be ported back to Windows. There are a number of options. We should explore firstly Kylix, who have a Open Source version (a reluctant Borland, who should not be trusted to support open source)

Java? Sorry guys we are Pascal programmers, I can move to PHP, but I just can't face the thought of those curly brackets. We will have to be able to remain open to J2EE and .Net processes. We will allow Java developers access to the project, but it is unlikely to be the core development language.

Route

STEP 1 is to make an MSQL version of the TurboCASH Database. This hsoudl be as a Delphi program or PHP Script.

Any takers?

STEP 2 Create a pump to move the current data from Paradox to MySQL, one way at first. This will allow TurboCASH user to publish their data to Linux. We can immeadiately have all the benefit of the TurboCASH Validation

STEP 3a Convert the current TurboCASH to read and Write directly to the MySQL database. This will get us to the situation quickly of having Windows Clients and an Optional Linux server.

STEP 3b Use PHP, Kylix or JAVA to give some Linus functionality. Allow other Linux systems to read our database

STEP 4 from the experience gained above set about developing a full Client for Linux.