This is a story about that path, one of many, which starts with prototype described in previous paper presented at CUC 2002.
Since then, we implemented completely free WebPAC under GPL license. This project didn't had strict plan about phases in beginning, it evolved with user needs and requirements. It also had life of it's own, directing us into different uncharted territories, from user-interface design to software architectures. Some of decisions where good, some where bad and we are here to tell the tail.
Different available Open Source tools where considered for different tasks. Some of them tested on small, spin-off projects which are also significant add-on to pool of free software. Best ideas where shared between projects, often resulting in improvements not envisioned at beginning. This enabled us to look beyond near goal, envisioning thing possible and enabling us to produce catalogue which is more that library catalog: it's a search engine for library data. Our users liked difference, while librarian wanted more strict searching. We tried to please them both.
Results where good even from commercial perspective. Using GPL licence for our work enabled us to improve WebPAC on different projects and share improvements between deployed installations. That way, this software project became service: changes where made based on requirements from users, and all installations received improvements.
Many software tools and components available under Open Source enabled us to do so. We are grateful for them, but finding right one for the task isn't always easy. Software design decision, like not using object oriented methods from beginning, also haunted us from time to time. We are now approaching limits of this code base, but have plans for next one. But the main question remains: where will WebPAC lead us next?