Urban Airship's AirMail takes push notifications to the next level | Hacker News Reader

Urban Airship's AirMail takes push notifications to the next level

tuaw.com

33 pointsthiele12 comments

Thanks for the submission!

I'm one of the cofounders of Urban Airship - I'd be more than happy to answer any questions if you have any.

Just want to say thanks for doing such a fantastic job with Urban Airship. We're using it for our app and it's proven to be extremely valuable and has saved us loads of time, we'll have no problem paying when we exceed our usage. :) Really looking forward to implementing AirMail too, can we sign up for early access anywhere?

One minor issue I've experienced is that push notifications sometimes never make it to the phone -- since UA/Apple doesn't return deliverability or system status, there's an inexplainable black hole and this is important given the nature of our app and potential consequences for failing to receive a notice.

I assume this type of integrity is related to why you've released AirMail? Is the deliverability failure largely an issue w/ Apple's push/queue system or UA's?

Awesome, thanks for using UA!

At the bottom of http://urbanairship.com/products/airmail/ , there's a signup form - we'll be emailing that list first with early access. Or, you can email me ( michael@urbanairship.com ) and I can get you on the list.

Apple's push service has a few quirks - in addition to connectivity issues (which can definitely be numerous - thanks AT&T!), there's also a problem when passing along a development device token to a production server. They'll silently drop the connection and we won't know until the TCP window size fills up that this happened - in other words, messages can be lost. There's a feature we've implemented called push debug mode that will watch for this, but it will slow down broadcasts or large sends. We recommend having this on during testing and are working on a more automated solution. So, you could try that.

But, in general, yes, issues like this were some of the initial inspiration behind AirMail - we tried as much as possible to talk to our customers and get feedback and the product naturally grew out of hearing what they wanted.

This looks very nice. I especially like it being a library so users are still taken to the relevant app. I have thought of adding a history of notifications to my app but was hoping apple would do this before long and didn't want to duplicate effort. I wonder if they will add this as it is a gaping hole in the functionality they provide.

I would love to add AirMail but as I send about 200,000 notifications a day with about $300/month of revenue there is no way I can afford your service.

Tighten the freemium model?

I only send notifications to the paid version of the app. But a one time $1.99 doesn't really map to the on going costs of sending the notifications. I have considered sending notifications to the free users as well as it seams to increase use of the app which should lead to purchases/affiliate commissions.

However a subscription model is probably the way to go of course.

Your website is fantastic. Who designed it?

Thanks!

Metalab ( http://metalabdesign.com/ ) did our initial site and identity, then we got our friends Refresh Media ( http://refreshmedia.com/ ) to revamp it to its current state. Cannot recommend both of them enough.

I thought about writing something like this for Android, but after thinking about it more, I don't see the point. My IMAP client can IDLE, so I get new messages (and a notification) instantly. Twidroid could (but doesn't) long-poll and get instant notifications too. GTalk is also push.

The only thing that I can't do on Android right now is get irc notifications without being connected to my irc proxy. This was why I was originally going to write something like this, but the problem is that no service is going to get private messages properly. Unless it is run on each server (doubt Freenode is going to let me bug their network), or the user runs their own client, it is simply not workable. Not workable == nobody paying for my service == don't really care.

So I don't really see the point of this sort of app.

No point for it on the android platform, that is.

So would this compete against the likes of notifo (the YC company)?

No - a Notifo or Prowl or other similar services could use us to deliver their notifications. They are enabling push on web apps and other items and creating a consumer app that serves as a central point for the notifications, whereas we provide a service for any app developer to integrate. Different targets and concepts :)

Reply on news.ycombinator.com