Note that it is specifically not intended that the code be "complete", as in robust or figuring out every corner case, or whatever qualities you would expect in real code. I think this is some of the interest of the book – it gives the reader an outline that they can fully understand and then expand upon.
I have the sync operation working, and am at 275 lines. So I have some space to add more stuff. What?
- WebSocket support. I found this annoying and lacking in pedagogical value. But it seems more "real time" than polling XMLHttpRequests.
- Server-Sent Events might be easier, but seem more obscure: http://en.wikipedia.org/wiki/Server-sent_events
- File locking. I'm relying on a single-threaded server for safety. Feels a little boring. Probably not that many lines though.
- Error handling on the client. The client really only handles the server acting exactly like it is supposed to.
- Refine the HTTP protocol. I think Content-Range is probably a bad idea. (Maybe I should do this anyway, it might not mean any more lines of code.)
- Actually filling out something like a "database". This probably means watching objects, keeping shadow copies of the objects, and then syncing them. I think this as much as anything is what gets it closer to Firebase or Meteor.
- Similar to above but more minimal, automating more of the tracking of saved and pending objects.
- Authentication. It's all very open right now.
- Garbage collection. Right now the database expands indefinitely, including obsolete versions of objects.
- Pagination (maybe means range support starting from the end of the resource?)
- Protecting against corrupt clients.
- Adding operational transformations. Probably doable, highly algorithmic. I don't find OT very widely applicable though.
- A different direction, try to find an existing commodity and open source server/service that is workable for time-series data instead of the custom server. I like my algorithm though, I would feel sad to complicate it by supporting a suboptimal server.
- An IndexedDB backend so that it's a full persistence-with-sync library. IndexedDB is a freakin' pain. But localStorage feels like a freakin' hack. Right now the hooks presume that your app already handles client-side storage on its own.
- Integration with a reasonable existing backend. Not sure what the state of the art is in client persistence, most of what I see is a thin veneer on localStorage.
Thoughts? What would you like to see an example of in the form of a concise bit of code? Or what completes the promise of the library, such that you would be excited to engage with the code (as a reader or extending it)?
- ZeOmega Infotech Pvt LtdLead Sofware Developer, 2009 - present
- Peoplelex Services Pvt Ltd.Software Developer, 2007 - 2009
- Cambridge SolutionsSoftware Developer, 2006 - 2007
- JNT UniversityComputer Science Engineering, 2001 - 2005
PyPy Status Blog: Making coverage.py faster under PyPy
Technical details. So how'd we do it? Previously, using sys.settrace() (which coverage.py uses under the hood) disabled the JIT. Except it d
Learn about the atomic bombings of Hiroshima and Nagasaki on the Google ...
This August marks the 68th anniversary of the atomic bombings of Hiroshima and Nagasaki at the end of World War II. Working together with th
APSRTC Official Website for Online Bus Ticket Booking - APSRTConline.in
APSRTConline.in is a newly launched website for APSRTC Advance Online Booking/Reservation System. Book your tickets online at APSRTConline.i
(Via Cult of Android)Hacker Demonstrates Ability To Remotely Hijack An A...
You can get some pretty cool gadgets that can be remote controlled by your smartphones and tablets, such as helicopters, cars, and even tank
Engineering an Android App: How to Design, Build, and Test – XDA Develop...
Building an Android App is not hard. However, it is certainly not easy. We have covered How to Build an Android App in the past. We've showe
Apple’s Ridiculous Lightning Rules Force $140,000 Kickstarter Project To...