Push/sync-support for iPhone

[Isegrimm]

Hi,
For years we've been waiting for Fastmail to support pushmail and syncing of calender and contacts.

In my view they could offer those features in two ways:
- Exchange/Activesync support, thereby supporting both pushmail and calender/contact-sync
- CalDAV and CardDAV, this will just offer calender/contact-sync so for pushmail another solution is still needed, like Apple supporting imap-idle or releasing their API for their push-system so that Fastmail can make use of it.

Both methods have been available for years, but Fastmail appears to be in a status quo about it.

Could someone from Fastmail tell us which direction they want to go with this? I believe I saw someone from Fastmail say they're waiting for Apple and basically just do nothing about it now..
If that's the case, why isn't Exchange/Activesync an option? I realize it might cost money due to licensing costs, but you could make this available only for paid accounts for example.

[Kudbe]

Could you describe how you currently use your iPhone with FastMail?

[Isegrimm]

Hi Kudbe, right now I have Fastmail configured as an imap server on my iPhone. I let Fastmail forward a notification of every mail to a gmail account. Gmail then gives me push messages.
Calendar and contacts are synced with my gmail account.

[Adon]

Exchange ActiveSync would be a welcome addition to this already excellent service.
Right now I'm stuck with regular polling on my iPhone which really frustrates me, reminding me of the old days with POP.

AFAIK all iOS devices support Exchange ActiveSync and do not support IMAP IDLE. Since there are hundreds of millions of iOS devices in the wild, wouldn't it be appropriate to add support for push email for us iOS users?

[robn]

We're working towards CalDAV and CardDAV, since there's now an implementation available within Cyrus. We're not there yet. The largest stumbling block is that the Cyrus implementation is against Cyrus 2.4, while we run a 2.5 pre-release, which isn't compatible internally. So we need to forward port that. I don't have an ETA for you.

ActiveSync is interesting, and probably not too difficult to implement, but we would have to pay a license fee to Microsoft for every user that uses it, which complicates things. We'd prefer to avoid it.

As noted, we support IMAP IDLE for new mail notifications. As far as I can tell the main reason iOS doesn't support it is that it requires a persistent connection to the IMAP server, which isn't very friendly on the battery (IDLE comes from a pre-iPhone world).

I'm personally not a iOS user, so I don't know what alternatives are available. If there's a reasonably widely-used notification system that we interface with without too much effort then we might be able to do something. Remember though that notifications are only part of the problem. We could hook up a notification system like Pushover without too much effort, but then what do you do with the notification. Can we make it increase the red "unread message" icon? Can we make a tap on it open the mail app? Those are the kind of tricky little things that make us throw up our hands and defer to app writers.

Let me know if there's something obvious we could be doing to provide a better push experience on mobile devices.

[n5bb]

Google recently dropped free support of instant notification for iOS devices (using ActiveSync style services). Old users are grandfathered for now. You have to pay about $50 a year per user to get a new Google account which supports ActiveSync:
Google's push Gmail changes surprise new iPhone buyers, Apple store reps

I have both Fastmail and my corporate Exchange account set up with iOS Mail on an iPhone. The Exchange push and integration features (email, contacts, and calendar) do work very well. At this time, Microsoft Exchange seems to be the only real integrated full-featured solution for businesses which seamlessly works for both PC's and iOS battery powered devices. It's unfortunate that these features are now proprietary (using patented features in various mobile network standards, mobile phone architecture, mobile operating systems and applications, and server support).

Bill

[robn]

I've been looking into this a little bit this morning. I'm thinking about using Pushover as an alternate target for SMS forwarding. So you'd be able to get push notifications, but they wouldn't be integrated with your mail client, and at least for now, you would have to set up an appropriate sieve rule to enable it.

The thing I wonder though, which I'm curious to know your thoughts on, is whether this particular mode is useful? In a proper push setup you probably want to receive a notification for every message (or at least those delivered to the inbox), and you don't want to have to set up a sieve rule for it. I can make the argument that its really just a UI issue, and we could make it so that when you register your Pushover client key we automatically set up a rule on the inbox for you, so maybe its the same thing. We would have to consider cost in that case - Pushover is only free up to the first 7500 messages/month, system wide. More than that is not a massive cost, but we'd need to consider if we wanted to absorb the cost, or limit the feature to certain service levels, or pass it on, or what.

So talk to me about how you'd expect this work and I'll keep experimenting with the idea and see what I can come up with.

(obligatory disclaimer: this is all speculation at this point, no commitment to any future service enhancement).

[cptlo306]

Originally Posted by robn

ActiveSync is interesting, and probably not too difficult to implement, but we would have to pay a license fee to Microsoft for every user that uses it, which complicates things. We'd prefer to avoid it.

What about if using Z-Push? Would that require payment to Microsoft?

[robn]

Originally Posted by cptlo306

What about if using Z-Push? Would that require payment to Microsoft?

Its a patent license, so yes.

[soromak]

Originally Posted by robn

Its a patent license, so yes.

That's weird. Are you sure you need to pay a tribute for Z-push? You can have ActiveSync working with a free Horde webmail, there is no word about paying license to Microsoft and I am sure MS lawyers would already target Horde devs. if they were breaking the law in any way.

[robn]

Our best understanding from Microsoft's own documentation is that there are patents covering parts of the protocol. Its also instructive to note that other major service providers (Google, Apple, etc) have taken licenses in order to provide services. We're also a for-profit service, so it stands to reason that we too would have to do so.

I'm not a lawyer, so I don't know any of this for sure. I certainly can't speak for the situation around Z-push, Horde or other products, free or not.

Of course, if there was sufficient demand for ActiveSync support, it may very well be worth it to us to take a license and offer it to our customers. Right now we're not seeing enough interest to make it worth the development and administrative cost. If that changes, we'll revisit it.

[Berenburger]

Originally Posted by robn

Of course, if there was sufficient demand for ActiveSync support, it may very well be worth it to us to take a license and offer it to our customers. Right now we're not seeing enough interest to make it worth the development and administrative cost. If that changes, we'll revisit it.

I beg your pardon?
For years and years we talked here about our need of syncing (devices). A quick search gives me 482 hits. And for years and years you give us the same answer. I really can not understand how syncing is still not yet on the top of FM's priority list and time is screwed up to pushover, 2-factor authentication and dropbox, etc. I don't believe you any more. I'm done with FM.

[robn]

Originally Posted by Berenburger

I beg your pardon?
For years and years we talked here about our need of syncing (devices). A quick search gives me 482 hits. And for years and years you give us the same answer.

ActiveSync is a bit different to most features, because the licensing requirement means it has an ongoing cost. Most things we implement have the initial development cost but not much more after that. For ActiveSync we have to be very sure that the costs of both the licenses themselves and the time needed to manage any agreement with Microsoft can be recovered. That isn't at all clear at this time.

Its hard to know how many people would use the feature if offered, especially if its limited to only particular service levels, or was passed on directly. Its also not clear that ActiveSync is the obvious choice that it may have been in the past. Google are dropping support for it, and there are other technologies (IMAP NOTIFY, CardDAV, CalDAV) that are getting traction in some places, and that we're much better placed to implement.

And yes, this forum has discussed it in the past, but this forum is a small segment of our user base. Which isn't to say opinions voiced here aren't useful or important, but they do have to be weighed against other factors that you may not be aware of.

I really can not understand how syncing is still not yet on the top of FM's priority list and time is screwed up to pushover, 2-factor authentication and dropbox, etc.

Both the recent 2FA updates and Dropbox have been very well recieved. There are just as many people that would have been unhappy if these features hadn't materialised.

"Priority" also isn't an easy thing to define. There isn't a single giant list of stuff to do that we work on from the top down. Its very much dependent on the time required to implement, availability of necessary people, and so on. To use your examples, Dropbox was mainly the work of three people, one frontend dev, one backend dev and one tester. It did not consume a significant portion of their time. Google Authenticator was entirely my own doing (with some QA support) and was done in a day while I took a mental break from some systems work. The Pushover stuff I've been talking about in this thread is a side project that I'm playing with on my own time (that is, in the evenings while I sit on the couch with my wife).

I'm quite sure that someone will try to intepret something I've written above as "Fastmail doesn't care about users" or "what we say on this forum is irrelevant" or something. All I can say is that it isn't the case. At the most base and mercenary level, we have to care because its you lot that pay the bills! But its more than that. Fastmail is a service we care deeply about, and we want our users to be delighted with it. The way to do that isn't always easy or obvious, but we're trying as hard as we can to constantly improve the quality and feature set of the service. When I look at the list of stuff we've released even this year, I'm pretty happy that we're doing a good job.

(I'll admit that our reliability lately has been a bit concerning, but there's two people working to solve that, including being woken in the middle of the night most days. While its still not brilliant for users, I don't think our dedication can be faulted.)

[robn]

Back on topic. Does anyone have any good idea of how they expect push to work on iOS? Otherwise I'm inclined just to implement Pushover as an experimental alternative to SMS for Sieve notifies and see what the response to that is like.

[cptlo306]

Originally Posted by robn

ActiveSync is a bit different to most features, because the licensing requirement means it has an ongoing cost. Most things we implement have the initial development cost but not much more after that. For ActiveSync we have to be very sure that the costs of both the licenses themselves and the time needed to manage any agreement with Microsoft can be recovered. That isn't at all clear at this time.

Its hard to know how many people would use the feature if offered, especially if its limited to only particular service levels, or was passed on directly. Its also not clear that ActiveSync is the obvious choice that it may have been in the past. Google are dropping support for it, and there are other technologies (IMAP NOTIFY, CardDAV, CalDAV) that are getting traction in some places, and that we're much better placed to implement.

And yes, this forum has discussed it in the past, but this forum is a small segment of our user base. Which isn't to say opinions voiced here aren't useful or important, but they do have to be weighed against other factors that you may not be aware of.



Both the recent 2FA updates and Dropbox have been very well recieved. There are just as many people that would have been unhappy if these features hadn't materialised.

"Priority" also isn't an easy thing to define. There isn't a single giant list of stuff to do that we work on from the top down. Its very much dependent on the time required to implement, availability of necessary people, and so on. To use your examples, Dropbox was mainly the work of three people, one frontend dev, one backend dev and one tester. It did not consume a significant portion of their time. Google Authenticator was entirely my own doing (with some QA support) and was done in a day while I took a mental break from some systems work. The Pushover stuff I've been talking about in this thread is a side project that I'm playing with on my own time (that is, in the evenings while I sit on the couch with my wife).

I'm quite sure that someone will try to intepret something I've written above as "Fastmail doesn't care about users" or "what we say on this forum is irrelevant" or something. All I can say is that it isn't the case. At the most base and mercenary level, we have to care because its you lot that pay the bills! But its more than that. Fastmail is a service we care deeply about, and we want our users to be delighted with it. The way to do that isn't always easy or obvious, but we're trying as hard as we can to constantly improve the quality and feature set of the service. When I look at the list of stuff we've released even this year, I'm pretty happy that we're doing a good job.

(I'll admit that our reliability lately has been a bit concerning, but there's two people working to solve that, including being woken in the middle of the night most days. While its still not brilliant for users, I don't think our dedication can be faulted.)

Not true about Google dropping ActiveSync. They are dropping 'free' ActiveSync. ActiveSync is still available with paid accounts. I doubt anyone is expecting Fastmail to implement ActiveSync as a free service.

Also, Zoho and Rackspace both added ActiveSync support over the last couple years which indicates to me that they believe it is a viable technology for syncing.