> Gmail does have IMAP, but it’s extremely unreliable and buggy.
That's never been my experience at all. I've used Gmail via Thunderbird (mostly for Enigmail) for years with very very few problems. It occasionally chokes when sending large attachments, but it always works on the second try.
I second this. I've used it through Thunderbird to Mail.app and now with Sparrow and haven't had any problems. Even during the web outage a year (or two?) ago, IMAP was still up while the web version was down.
I've also been using gmail since it was invite-only beta and haven't experienced any major problems through the 7 years I've used it, not sure what everyone is complaining about.
Worth noting: I think the original complaint/idea (I think from Gruber or someone similar, I don't remember exactly) was more along the lines of having Twitter offer premium accounts that would never failwhale, and similar equivalents for other web services these days that seem to be falling apart under load. That idea I could understand, just not with Gmail.
I think if the bugs affected everyone, no one would use the product. It is going to depend on your use habits and the path your email client takes through the API.
After 15 years of running sendmail/postfix/uwimap/dovecot for a growing company, and uncounted spam fighting tools, I use google for all my mail accounts now. On their scale, they can do a better job at spam fighting than I can with only access to my own mail streams. (But I can beat them for reliability.)
1) The 10 connection limit bites surprisingly early. Mail.app makes multiple connections, and if you leave a couple computers logged in, carry an iPhone, and maybe use an iPad you will find yourself randomly blocked from your email. (Their "help"[1] article says you can reach the limit with as few as two sessions open.)
Google's solution is for you to remember to disconnect your computer from their IMAP servers when you are away.
2) IMAP searching is a mystery. I routinely have to use the web interface to search for things that I know are in my mail, but I can't find from Mail.app. I don't know who is at fault here.
3) I have instances of attachments becoming corrupted. Again, I can't conclusively place blame.
4) The strange organization issues from the "tags" vs. "folders" issues can be confusing to people that don't understand the nuts and bolts.
5) They have rather limited logging available. I recently had two laptops stolen which subsequently did IMAP updates while in possession of the parties unknown. Google's low granularity logging information (day and IP but no time information) caused trouble getting a court order to compel the ISP to produce the subscriber data. The court wants a specific time.
So there you have it. There are warts. But I'm not going to leave because they are great spam killers.
[1] They call it a "help article", but it really just says "This breaks, we know why, we do it deliberately, and we can't be bothered to make it better. Suck it, meatbag."
For what it's worth, I can confirm that Gmail IMAP legitimately is unbearably slow for some of us (conversely, the web interface has always been snappy for me, though I know it's been terribly slow for some). To the point that I gave up on keeping my email in sync and used POP on my desktop for years.
Any technically inclined Gmail user should at least consider switching his email address to a domain he owns plus free Google Apps. It works just like regular Gmail, but you can change providers whenever you'd like. It may take many years, but someday Google is going to be technologically behind or an untenable steward of your data.
Changing email addresses is a pain -- better to do it now, while you can slowly transition with email forwarding. Use filters to auto-label mail that is sent to the old address, ask senders to change your address, and switch over any mailing lists and accounts as automatic emails come in.
In practical terms for those interested: in cPanel, for instance you can set forwarding for info@example.com to example.com@gmail.com. Then, when you want to switch providers, change the forwarding rule in cPanel.
This would still leave you with the problem of your history of emails being stuck within Gmail.
An MX-based setup does not use the cPanel provider's mail server, so it should be more reliable/secure/convenient in a dispute/etc. If you also use your @gmail.com account to send mail, it has the added benefit of allowing you to migrate off GMail easily, should it ever start to suck (just set up an account elsewhere and change the MX record - people should already have you@yourdomain in their address book).
Additionally, most spam filters take the sending host into account ("this host is known to have sent a million spam messages today - let's not bother talking to it"), but this information is lost (untrustworthy) if you forward via another mail server; it's likely that GMail has a better spam filter than your cPanel provider, so try to give it all the information you can.
There are some minor other issues - faster delivery, less wasted disk space on the cPanel provider's end, message sizes are limited only by GMail - but the above are the main ones.
Do note that you need GMail "for domains" instead of just a @gmail.com address. The "for domains" version is free, though (perhaps up to some upper limit - 50 mailboxes?).
If you use Outlook, an under-emphasized feature of the paid service is their Google Apps Sync tool. It provides excellent, fast two-way synchronization for mail/contacts/calendars in Outlook, using an external app that interacts with Outlook via MAPI. Basically, it brings Google Apps nearly on par with Exchange for Outlook users.
Note: If you do it this way, and use your own domain (I'd recommend registering it through someone other than Google), if/when Gmail goes lame, you can pack up and take your address with you.
(Meantime, keep your emails backed up.)
I also hope, a bit, that the paid version means Gmail/Apps may have some degree of independence/durability that is separate from the advertising revenue. I'm probably wrong, though, at least at this point.
You must own any data that’s irreplaceable to you.
An interesting additional point given the ongoing trend towards web apps and storing data "in the cloud". Note that "own" in this sense means you need to be able to export it in a standard way (e.g. access via IMAP). I don't see a lot of enthusiasm for this from the big sites out there. I can only hope that users start to realise the value in this and begin demanding more control of the data they are creating.
I would be extremely careful with that project: it relies on Javascript crypto extensively to project users from storage nodes. Javascript crypto is nearly always completely broken (crypto needs to take input-independent time for all operations, for instance), and their implementation hasn't been audited at all. (It has been lifted from some Stanford crypto group, IIRC, but that's hardly enough checking.)
I wasn't a LiveJournal user, but saw one of Brad Fitzpatrick's excellent scalability talks years ago. I think he had a special flag that could be set for paid requests that would give them priority in perlbal, the load balancer he wrote.
Obviously it doesn't do much when something big is broken, but Brad did have a way to make sure paying customers' requests were processed first, and had realtime stats on average request latency for paid vs. free accounts.
All online services have downtime, lose data, don't serve every user perfectly at every second -- paid or free. This post makes no sense. Not to mention Google makes a ton of money from serving ads and Gmail already runs at a 99.99% uptime rate (http://www.zdnet.com/blog/btl/google-aiming-for-9999-percent...)
And how did paying for service work out for Jott users who tied the whole thing into their voicemail and now have to untie it pretty quick?
He himself says he uses a small email host at $40/year, but paying for Google Apps is essentially the exact same thing as what he is doing.
He point of "its just IMAP" is mute. Gmail supports IMAP. Regardless of how reliable it is on and ongoing basis, if you look at it for the ability to export your data, its the same as what he has.
It sounds like his point is to use a smaller service. A smaller service would be less complex, but services that are bigger are likely popular for a reason.
A car that is a best seller is worth considering buying. You don't want to avoid a best seller just because its a best seller.
No, I certainly read the whole thing. With Google Apps you get all the support (SLA, uptime, phone contact) along with data ownership (by using an offline mail client for example).
I really like this post. A lot. The idea that just throwing money at something will in some way keep it alive is kind-of silly.
The closest you'll come to this if you have ONLY a paid option, which would eliminate the free users, and make the system far smaller and more manageable, in which case there WOULD be more up-time because people were paying.
But I agree, to to think a small subset of paying users will have any affect on the up-time of a major site is just a pipe dream.
>to think a small subset of paying users will have any affect on the up-time of a major site is just a pipe dream
You seem to rely on the assumption that every major site's budget is so lush that a small portion of the userbase paying would be unable to solve any of their technical or employee needs.
Actually, Reddit is the exact site I was thinking of when I made this post. They opened up the Reddit Gold, it helped a little, but lately the site seems to go down daily, or even more often. It's almost like it's getting worse.
Since 2007, I’ve used FastMail [..]. FastMail’s uptime has been incredibly good [..].
Funny, I clicked the link and saw this:
MessagingEngine.com Server Outage
[..] I'm sorry, the server your email is on is currently down. We apologise for the inconvenience; any email sent to you during this time is being queued by another server and will be delivered as soon as the server is working again.
While I had a FastMail account, I distinctly recall at least a dozen instances of it going offline. Though this was early 2000 or so for only a couple years.
This blog post has a commission link to the service he subscribes to (or is advertising). I've had less downtime with google apps than I had with any other solution and I don't have to spend time managing it. Google spam filters are effective enough that the service can be down for an hour a day and it would still be a good service for me because it saves me so much time.
I think that this is a fair assessment. It would be nice to be able to set up SLAs for a premium, but I honestly can't see any major ad-based providers offering this (breaking the SLA would be too costly).
I also appreciate the restatement of "You must own any data that’s irreplaceable to you." It's suprising how lost that simple fact is today.
I don't quite understand the marketing magic that the term "SLA" has over people. Are there SLAs that cover my actual losses due to downtime? Perhaps they exist, but I've never seen them
If my customers can't email me for a week, I assure you that a partial refund on the fee from my email provider is not going to make things right.
A good SLA makes downtime painful enough for the provider that they'll want to fix things. Most SLA's have no teeth though - marketing only. (Any "100%" SLA falls in this category.)
Google already gets paid for Gmail. It uses Gmail as a platform on which to serve very targeted advertisements on every page. Likely a greater overall value with their number of users than if they did away with the ads and charged for the service, instead.
Not to attack your post, but I feel as though the HN community is aware of Google's business model by now. I feel as though any time a story appears about Google (or any company that utilizes their model), there are at least a few comments about how ads are what is driving them. It just seems to be restating common knowledge, at this point. You aren't the only one to be doing it either, I just happened to reply to this post.
I don't see how it's any less relevant to the discussion of "they should charge so the service is smoother". If we're going to have that conversation, then it's important to acknowledge that they already have a business model that addresses this. Google isn't hurting for money nor are their projects and when something like gmail encounters a hitch, it has nothing to do with funding. Giving gmail more funding or funding it by direct service fees over their existing model isn't going to make it any more reliable, since it is still susceptible to "things just going bad" from time to time. Really, if my comment is inane, then so is the whole submission (which I would kind of agree with - no slight intended to the submitter).
It seems like there should be more startups out there trying to do a better job than Google is at email management. I hate to think that there's no more room for innovation or competition in such an important area.
Maybe but it's an expensive service to run. Google can make it work because they have the supporting infrastructure to harvest enough data to drive their ad business. It would be costly for a startup to spin this up on its own.
Eh, I don't buy it. Sure, searching on an storing a lot of mail is a tough problem, but there are plenty of startups working on what I'd consider to be much harder. With open source tools, AWS, etc. this isn't an impossible problem. Plus, you don't have to scale to millions of users right away.
I think people are staying away from this area for a reason. Email is of declining importance anyway. I know more and more people that, for better or worse, communicate mainly through Facebook.
Also, don't underestimate how much work it is to filer spam effectively.
more and more people using facebook sounds like a reason and an opportunity to fix email. why is it easier to send messages via facebook?
email has a few huge disadvantages: you need to keep your address book up to date and your inbox is full of email from machines - notifications and bulk.
make an email provider that has a twist - maybe automatically sync it with your facebook friends as the address book and limit lenght of your replies.
I don’t completely get what Marco is trying to say. IMAP is not exclusive to FastMail (which is is a great service). Gmail has it too. And POP3 of course.
Most of the time I use Gmail from its web interface. The few times I had problems with the webif, I launched my desktop client and used that.
As for using your own domain name, well, life is short. I can’t be bothered with all the small things.
I am not sure how you can pay them for the second part, but I found a plugin called Rapportive that replaces ads with more useful stuff: http://rapportive.com/
You should try it. It's really useful for business emails.
"I’ve seen many similar pleas recently whenever any popular, free web service has problems: “Please, let us pay you so there won’t be any problems!”
But it’s an impossible dream. If a web service is popular enough that you hear about it when it has downtime or major issues, it’s probably a large, very complex system. 100% uptime is effectively impossible."
True, but in fairness to bitskits, the author goes on to talk about paying to use FastMail as an alternative to Gmail. Yet, he doesn't provide any comparative uptime data to show that FastMail is objectively better.
As for being able to own your data and store local backups, you can do that with Gmail POP, or with Gmail IMAP and an email client that'll store messages locally.
(The author's argument that Gmail IMAP is buggy, is one anecdote; I can provide my own anecdote to the contrary. Again, we'd need some data to prove that FastMail is better.)
Yet, he doesn't provide any comparative uptime data to show that FastMail is objectively better.
This is true, but his solution is not one that must use FastMail. His address is his own, and he backs up his own email. So if FastMail lost everything of his, he would be able to switch over to something else. That he owns his own email address is the important factor, not that he happens to use FastMail.
For the record, I use Gmail for all of my personal email, and I don't bother keeping backups.
I was asking the question in general, mostly because I've seen a rash of articles saying "Let me pay to make twitter stable", etc. I should have formed my comment better, as it does appear to question this article rather than the idea.
I think it's based on the intuition that "you get what you pay for"; if a company has more revenue they should be able to spend more money on availability. Unfortunately, I don't think money is the bottleneck here.
That's never been my experience at all. I've used Gmail via Thunderbird (mostly for Enigmail) for years with very very few problems. It occasionally chokes when sending large attachments, but it always works on the second try.