Question

I have a webserver that hosts data for an application that I’m building on Android tablets. The data changes fairly frequently, but I realize that tablets and phones aren’t always connected to data, necessarily, and it would be good for users to already have access to data that they had viewed before.

I want an application to pull data from the webserver whenever possible, and when it’s not possible, to pull from the SQLite server (or whatever the proper method is) on the phone.

Is there an implementation or practice already set in place to do this? If so, what is the current method for implementing such a feature? And could this same method be applied to making changes to data and then pushing those changes whenever a data-connection came back?

Answer 1

There isn’t really a common framework for this in general as far as I know. There are quite a few posts here that deal with image caching which could be easily adaptable but nothing that I know of completely generalized or specialized for offline viewing. That said, most of the solutions just work with URLs and files, so it could work fairly easily if your data is very simple. It very much depends on what your data is and how you plan to display it though. If it consists of web pages to be displayed in a WebView, then it could get annoying because then you have to traverse links (including images, css, javascript, etc.) and cache those. However, if you are working with specific isolated URLs or structured data from which you can parse all related URLs easily, then it’s much easier.

Here’s one such image-oriented solution: https://github.com/ZaBlanc/WebImageView

The general approach is fairly straightforward. It might be something like this:

• Any time you try to “get” a URL, you always read it from your cache
• If your cache has it then you can use it happily and be on your way
• If your cache doesn’t have it, then you have to queue up a request to asynchronously download the file
• Once a file download has completed, send off some sort of notification to tell the application to handle it.

You also have to deal with stuff like cache expiration policies, and how to display the data, but might be a start.

If you’re thinking of two-way synchronization, then it’s going to get much harder. You have to think about locks and conflicts (can two people update the same item at the same time? if so, then what happens?), atomicity (do you have related items that only make sense if both are updated/added?), data validation (what if the user sends bad data) and a whole lot more. That’s probably why no one has come up with anything (that I know of) quite yet. It’d be quite useful, but it can get pretty complex and really involves both client and server work.