NotFreeBeer

Created by: Lester Caine, Last modification: Fri 01 of Aug, 2008 (11:51 UTC)

There has been much discussion about free software and open source, so some pointers as to our views on this matter seem appropriate.

The major part of the software we supply is built on open source software and as a result there are no licensing restrictions. A significant  gap has been created as the text to speech engine that we used to use has been taken off the market following it's purchase by a competitor and this has become a prime example of the real advantage of open source!

Borland Windows Codebase

In the 1990's our systems were built using 'C++' in Builder 5/6 and migrated from dBase to SQL running on Interbase. Proprietary libraries provided additional functions, and programs had to be deployed to every machine that required access to the information. Some of the multi-screen display equipment originally designed under Windows98 can not now be recreated since the support is no longer available for later versions of Windows.
The first bombshell came in 1999 when Borland/Inprise announced the end of life for Interbase. Many major systems had been developed over 15+ years that relied on this relational database, and the backlash was such that Borland relieved some of the pressure by open sourcing the code base and the Firebird project was formed. Firebird has a large international following, and a vast range of third party support, yet remains seen as an underdog in relational databases. Something that needs to be addressed, but as there is no 'commercial' interests driving the current project, the project competes on performance. The project is supported instead by the Firebird Foundation which coordinates payment of the core developers.
Borland realised the mistake they had made over Interbase and tried to backtrack, pulling the Interbase name back in house, but many major players were disillusioned with Borland and continue to fund full time developers of Firebird today. Borland's development tools went through a similar 'sticky patch' with attempts to move the C++ developers tools to a different area, and this required another rethink.

Web powered development

With the increasing pressure on deployment of applications, a move to a web based approach for clients was obvious, and there was no discussion what platform that was going to be. Apache and PHP were the obvious choice in early 2000 and have remained a stable and ever expanding base ever since. There is no need to explain that this combination can be scaled to support any number of users far beyond anything we are likely to see in an intranet based system, and that the only cost is that of hardware  to support it.
With our development suite in a back water, a replacement came into view in the form of Eclipse. This provides a base to manage all of the old C based code, as well as PHP and all of the html related file formats in the one umbrella. It also manages project documentation, but this is something we are supporting ourselves in our web based framework. Commercial interests are prevalent in Eclipse, so the PHPEclipse plugin that we have been using for some years is not the basis for the 'official' PHP editor for Eclipse, but most Linux distributions supply PHPEclipse along with everything else we need to build a development machine.
Systems that are currently stuck using Windows98 for their multi-screen displays will be upgraded to use Linux and there will be no need to replace lower powered computers simply to run a later more powerful version of windows. In fact in most areas, there is no need to have powerful windows based computers when all the user is accessing is a web browser. How many counter positions need full video and interactive media capability when all one is doing is accessing a vast range of text based information.

Problems with commercial decisions

Open source is about cooperation. A lot of time is put in by those of use who use the software, and that time is not paid for directly, but when deploying applications, users can benefit from the work of others. Since the base on which the systems are built is open, USERS can contribute modules to the system rather than every separate site paying to have the same thing written. Because of the way the Benefit Agencies were broken up, our annual user group meetings were lost and this loss is something we would like to reinstate. Many of the expansions to the existing system came about as consolidated requests from these meeting, and the cost was then spread over all maintenance agreements. Users found it most advantageous since they could compare notes and often changed methods of working simply because of discussion with offices.
This brings us back to text to speech, which is currently unavailable, we use the original pre-recorded ticket numbers for announcements. Rhetorical had produced a very nice package that worked well with complex names and this was bought out by an American competitor, who no longer supplies it. The replacement is simply not as good, so either we will have to make do with lower quality speech at a higher price, or switch to one of the open source offerings which we can improve ourselves over time.

Moving Forward

On a more positive front, some of you may have already seen our mapping interface Isle of Man Mapping and we are involved with OpenStreetMap which is creating an open source set of mapping data covering the whole world. This data can be edited on-line locally and one can even add your own local material.
Lets get together and produce the systems that you the users want without the pressure of cost for every little change. Duplicated for every site that wants the same addition!