Anyway, here are a few features that I think Google could add, which would make it a more attractive platform.
* Some form of cron-job-like functionality; that is, the option to have a function executed at set intervals. This is essential for all sorts of applications, and could be relatively easily implemented in the current model.
* Allow users to easily create widgets to fit into iGoogle, Google Apps for Domains and the like. This would be particularly handy for the commercial sale of applications for corporate domains thing that I mentioned earlier.
* Allow application end-users to authorise applications to access Google APIs in their names without further action. Google already kind of has something like this for their GData APIs, but the user would, in this case, probably end in having to enter their password twice per session, once to access the App Engine app, and once to authorise it to talk to Blogger or Calendar or whatever. It would seem simpler to just allow them to permanently authorise the app, given that it's using Google's authentication system anyway.
* Somehow provide support for Comet; that is, webapp push to client. This would allow people to write chat apps, real-time multiplayer games etc. Google already has a high-capacity instant messaging system, Talk, which could serve as a backend; the user application would simply call an API function to send messages to users who were listening with a standard Comet technique. A Comet-like thing is already used by the Google Talk client in GMail, and is, by all accounts, one of the more stable and client-compatible ones around.
* For large sites which use more than the allocated free resources, allow the owner to ask or require the user to pay for resource use through an integrated micropayment system. Micropayment has never worked all that well before, but in this scenario, a user may be accessing many different App Engine sites through a Google account which could be linked with a Google Checkout thing. Allow owner to offer additional features, ad-free pages, etc. for micropayments.
Just a few ideas, but I'd be surprised if at least 1 and 2 don't turn up at some point. 3 might present privacy concerns, and 4 is dubious because widespread push-to-client is still a pretty new thing, even though it should certainly be technically feasible. 5 is radical, but I really do think that if App Engine apps take off, micropayments could really work there.
Subscribe to:
Post Comments (Atom)

No comments:
Post a Comment