How to add live chat to your Parse-powered app with Hipmob | Hacker News Reader

How to add live chat to your Parse-powered app with Hipmob

engineering.hipmob.com

22 pointsfomojola8 comments

Hey guys - Ayo from Hipmob here. TL:DR; we've made it super easy to use Parse for push notifications, in an app with Hipmob installed. We're actively developing more stuff like this - making Hipmob easier to work with tools you already use, so if you have any ideas/integrations you want to see, we'd love to hear about it! Leave in the comments or email me at ayo at hipmob.com

EDIT: Also - iOS is coming soon :)

Doesn't seem like there's an easy way to do p2p chatting without hitting your server (often) with a new list of 'friends.' Any chance of supporting something simpler?

Our API lets you send individual friend list updates from your backend, and we're actually working on a web-hook based way of handling friend requests: if we need to verify that user A has permission to talk to a new user B we can POST to your web server and then update the friend list, or (if you use Parse) you can specify a Parse query and we'll look up the friends that way.

While Parse solves a huge problem, there is also a catch - You are locked into an ecosystem where you have no control over your codebase incase of an Emergency. For example, Parse's recent downtime made developers helpless as they had only two options - Re-write the entire app, or wait for the downtime to get over. So basically, its a trade-off between convenience and reliability. Choose wisely.

No, this is not meant to discourage you from using Parse. If it solves your needs, then go for it, but always have this catch in mind and preferably have some kind of a back up plan incase Parse goes down along with your app.

I chose to use it because I'm not a very good backend developer (my bad) and needed something up and running quickly. But I agree that it's probably not a good longterm solution.

Writing a full backend is serious work and it will cost you significantly in terms of time to market. For a green field app that doesn't require a companion web service it's still very attractive. Parse runs on AWS so if you build your backend on AWS, you still need to build a failover mechanism to handle that to be more reliable than Parse.

My main concern with the Parse SDK on the iOS is the lack of a clean model layer (and no, subclassing PFObject is not allowed) that would allow you to rip out Parse once you outgrow it. The way I understand it is the Parse SDK more or less encourages you to embed your models and business logic into the controllers as far as MVC goes.

It would be nicer to have a smoother transition path. A way to use CoreData as a local storage that seamlessly uses Parse would be a great option.

Minor gripes include the lack of a bulk delete api. Although there is a bulk save function so you could use that to soft delete.

However, with all that said, still seems like a great option even though the migration to a self-managed backend seems pretty rough (to maintain data integrity).

I'm working on that problem: http://github.com/jawngee/modelkit (iOS/Cocoa)

Have you looked at FTASync?

Reply on news.ycombinator.com