Preparation for the beta release has included:
- We have migrated T-PEN to a virtual server in the Network Services of Saint Louis University. T-PEN had always run in a virtual environment--in a linux box sitting on our Windows server. This started producing dropped connections, so much so that many of the javascript calls on the JSP pages were never completed. Now we are running on virtual linux server and the quality of service has increased significantly.
- We have automated our build procedure, and so regular builds will be pushed automatically (still allowing for manual builds to fix critical bugs) on a daily basis to the test-bed and then weekly to the live version.
- We've developed a user reporting system that will be accessible on each page of T-PEN. Users will be able to report a problem (bug), suggest a new feature, ask a question and/or suggest a new repository for T-PEN users. All these user-reported issues will be fed automatically into our bug-reporting and software management system.
- We have restructured our data store in a way that tightens the relationship between the image coordinates of a line and any transcription a user enters. This new data structure will provide greater flexibility in line parsing adjustments and will further ensure data is never lost when a user modifies how a manuscript page is parsed.
- We have reduced the number of individual pages that enabled T-PEN's functionality. We are nearly at the point that the Transcription UI is becoming an "application window." Manual line parsing and the line breaking of existing transcription text are now integrated into the Transcription UI. The also means the codebase executes faster, and there is a more streamlined way to update the data store.
This is all in addition to the continual acts of bug-discovery and bug-fixing!
Stay tuned for more updates as we head towards T-PEN 1.0.
