Why is setting up your own mail server a bad idea? I've been running my own for 13 years now (plain old postfix and dovecot on whatever linux distro I favor at the time). It works great.
I used to own and manage my own mail server when I had to do it for my business back in the day. I had to be up to date in all involved email server management and its perks anyway, so it wasn't a lot of extra work. Now it would be. For a normal email user, it would be a nightmare.
I'm the kind of person who's very disrupted by having a ton of small tasks in the background all the time. Maintaining your own email server adds a bunch of them, even if you are already knowledgeable (keeping your domain(s), storage & redundancy, having to maintain a server with good uptime and with a lot of security concerns - it's online and it broadcasts its IP in headers, it's immediately spotted as running an email server and targeted to be made a spam-relay or worse).
If you're not even knowledgeable about it, the amount of stuff you need to learn and be familiar with is ridiculous. Maybe they don't even occur to you off the top of your head now, but the amount of little things one learns over the years about server maintenance is massive and a lot of it is absolutely necessary to run an email server with guarantees. Having to "insource" all that shit work is something I've been trying to avoid but I'm afraid I will have to do. I rent, this means sometimes I have to move and keeping servers 365/24/7 is a problem. Typical home connections are rather shitty for an email server in terms of uptime - you'd have the occasional email silently not arriving (depending on sender retry config) and also the occasional bounce (server coming back up but not properly - happens) and that doesn't look good for serious communications this day an age. And like that, a large number of concerns both particular and common to each email user.
Having backups (also involves shit work but not as much) mitigates the problem but for some of us, simply to stop receiving email at a certain address for a couple of days can cause a lot of trouble.
2) You value your time. Some people don't, it's not really worth arguing this point. But it is a reason running a mail server is a bad idea for most people.
And #2 is just wrong, sorry. I spend minutes a week doing anything at all related to maintenance on that box (I use it far more regularly for productive purposes, though). If you can handle running a linux box from a console, you can learn to do it too. Or don't, it's up to you. But telling me I don't value my time is just out of line.
>> I've been running my own for 13 years now (plain old postfix and dovecot on whatever linux distro I favor at the time).
For those of us without the experience of 13 years running postfix and dovecot (and spamassassin and writing perl filters), there will certainly be at least some time investment. That's what I was talking about: the price in hours to go from zero to competent. You may be too competent by now at email hosting to realize that it would not be a minutes per week affair for most people to do well.
Obviously if it works for you, great. Interesting to note that you started running your own long before GMail; the calculus of starting to self-host is different now.
Re #1, you should lend your spam filtering tools to Yahoo! In all seriousness, a handful of unwanted messages per day would be a dramatic improvement to my Yahoo! inbox. Whatever they are doing over there is not as good as what you're running.
I sometimes wonder if GMail lets these through for certain people, and relies on the "mark this as spam" button as some sort of mechanical turk...
And MX records don't support an alternative port to my knowledge.
Unless you shell out for the "real internet" (the business package) it's like fighting an uphill battle to run your own servers anymore.
Ahhh the Internet... I remember that.
What gets me is that email was one of the first peer-to-peer networks, and 90% of people, including myself, on residential links, are excluded from using it as designed. It seems more wrong to pay to solve a man-made problem. Free webmail is "good enough" if I'm just going to pull my messages offline and use it as a relay...
I remember there being "more internet" on my 28kbps modem... Port 25 and 80 worked from home, SMTP servers didn't reject mail from anyone on a residential ISP. I actually wrote a letter to my ISPs when they started blocking SMTP (yeah, I'm THAT guy)... Their argument was "but spammers," and they wouldn't make an exception for 1 out of a thousand. I even wound up switching providers over it. A year later everyone was doing it and there was no stand left to make. Spam is our "airport security" scarecrow (among others... copyright, porn, etc)... We'll undo the whole thing if we have to.
And so Gmail it is, until Email 2.0 comes around, and is new enough not to be intentionally broken, or I decide to shell out and license the real internet from my oversubscribed ISP that throttles uploads so noone can offer new and interesting services using their networks that might compete with them.
Don't get me started on QoS -- the neutrality killer... (We moan when people throttle BitTorrent, but when it's called QoS, it's "Smart"!)
Behold, de-evolution.
claps the disappointed clap... of the disappointed :o)
If you run and maintain your own SMTP server and it goes down, you won't be able to send email.
Everyone else will still be able to send email to you, and it will be delivered to you, and you'll be able to read it.
Now, if you break Postfix or something on your server and start bouncing emails then you can be in trouble, or if you mess up the DNS somehow.
But generally running your own mail server is a set it up and leave it alone type of affair. Any junior level hacker can cobble one together with guides online and have it up and running with no problems in a few hours.
Generally, running an SMTP server is quite a responsibility: not losing mail, not being exploited to send spam. What matters is not so much that you can set up in a few hours, but whether you want to take on that responsibility.