Ok, it is just one of those things that I can’t understand. Why there is no common, widely used solution that is used to connect mobile devices to web services? I am talking about J2ME, to be specific. For the last two years, I have written three of four J2ME apps that uses web services to connect to some .NET or Java enterprise application. After quite a bit of a search, ksoap 2 seemed like the best option. It has a licence that allows me to use it for whatever I want, and it has some documents (not much) that helps me out.
However, the interesting thing is: there is not development in this front. JSR 172 is the standard by Sun that vendors would probably be following for J2ME implementation, but for some reason I can’t find a usable implementation of it. Yes Sun provides a reference implementation , but due to some security issues it can’t be used in many phones, at least it did not work on four or five different ones I have tried.
Apache had a project named Mirea, but it is apparently dead now. Wingfoot has a soap library that is free, but none of the free options seems quite active. By quite active I mean a community of people who use it and discuss it. For me, no matter how impressive a piece of software technology is; if there is not a user base around it, I’d not consider it as a good choice. What I want to see around a software technology is users, since they are the ones forcing the technology to go forward.
Even though I can still use current frameworks now (ksoap2 or wingfoot soap library) I am not fine with it. Why using web services over j2me did not took of is an interesting question. Unfortunately the answer does not fix my problem. So I did another batch of searches, and REST based approach for web services seemed like a good approach. REST uses current standards but it does not seem to have the the complexity of SOAP based approaches. Soap based web services has come to a point where it can cover quite complex requirements, but like every system power brings complexity, and for small clients, this complexity decreases productivity. In my case, I have the advantage of having control over both sides of the communication, meaning both the web service and J2ME client, so I’ll give REST based approach a try. I’ll be posting the results here.
Oh, and all this is for an experimental project in healthcare. It is built on mostly young technologies like Ruby on Rails and if REST approach works, that’ll be another unusual step for me, since I have been using .NET, JAVA and large scale tools for most of the time. Maybe some agility is really possible after all?