I recently spent the day with one of our largest customers to discuss the options for their next generation mobile architecture. We have worked with them over the last ten years and they are now in their third generation architecture and looking to the future.
1st Generation:Thin-Client (DOS, VAX)
Initially they developed thin client applications (DOS terminals running STEP) using COBOL on VAX servers. Although this was fifteenth it was a pretty good architecture. With this architecture the mobile applications were really an extension of their other store applications, and updating the applications was juust updating the server code. The big down side was writing the applications were difficult since they had to manage all of the client state in a multi-client server.
2nd Generation:Thin-Client (DOS, NT)
In the next generation they migrated from VAX to Windows NT and began developing with Symbol WaveWorks.This was a big step forward in ease of development since WaveWorks eliminated the need for developers to manage all of the terminals themselves. The business objects became stateless COM objects and the WaveWorks DisplayServer and Script Engine managed the terminal state.This continued to work very well and with StepOne’s STEP-CE client it was able to run on modern terminals.
3rd Generation: Thick-Client (CE, NT)
Next came the migration to the StepOne snapRETAIL .NET architecture which brought a modern UI and Web Services to access the business objects.Overall this has been a big improvement in the presentation layer as well as allowing WS integration. In addition this has added improvements such as Intelligent Alerts, Popup Messages and Improved UI controls that are now easier to implement in a thick client.
Overall this is a big improvement in functionality and performance, but now has the challenge of deploying the apps to thousands of terminals for any change. There are plenty of tools for managing this including AirBEAM, Avalanche, Afaria, SMS etc. but still not as easy as the big thin client switch which has its own set of challenges.
The Future: Are Mobile Web Apps Next?
Now the question is what should we be looking at next? Is it time to start moving towards web applications or is it too bleeding edge to commit their enterprise apps.