Actually, sadly, my experience is the other way around. Setting up mail server is easy, but taking care of spam and blacklist is never ending work, and you always have to wonder whether this time the mail will arrive, or whether some server on the way is going to have a bad day and drop your email without telling anyone. And of course, since it's your mail, everyone assume you're the one at fault.
At least, that's how it worked few years ago, but I don't think the situation got better. Some things only get worse.
Even for a lot of technical folks, the maintenance is a hassle. My anecdata: almost all of the technically capable folks I know who could maintain their own email server choose to outsource it. For the average Internet user it's outright impossible. Even if it only took "3-person days a year" for maintenance, the mountain of experience required to be capable of the setup just isn't there. Not to mention the stress and hassle of dealing with attackers, software upgrades, outages of various sorts, etc. And that pain tends to be multiplied by the number of services maintained.
In this light, perhaps tools like Docker are quite a bit more interesting. We might imagine a scenario where the maintenance itself is the product, not dissimilar to automatic updates on Apple's App Store. You buy a "cloud app", which is installed on your server (or cloud instances) in a container, receives auto-updates provided by the vendor. Purchase and configuration is handled by apps on mobile and/or web.
There a ton of potential wrinkes in this idea, and certainly some services (e.g. web search) wouldn't adapt into to this model... but it would be pretty neat to see even a partial shift to services-as-apps.
In the end, you have to make the tough choice between whether you want to make a controlled, voluntary disclosure of your private information (e-mail text and metadata) to Google, or risk an uncontrolled, involuntary disclosure of that private information to potentially malicious actors.
My hope is that we'll get some technological infrastructure solutions to this quandry through the next generation of privacy-preserving communications protocols that are currently being worked out (Tox, BitMessage, Flowingmail, etc - not endorsing any of these, just examples of the current crop). Even then, it'll be a while before we can securely protect both our data and metadata.
Eg., set up an alias that forwards to gmail and do your best to make sure spam doesn't get forwarded to gmail and put your entire server on their blacklist? Probably you didn't do good enough.
The only frustrating part is that I generally don't get to choose when those 3 days happen—most all time spent is putting out fires. But then again that's sysop life in general…
Wow, that's way worse than I would have guessed.
If your salary is $100,000 a year, and you work 40-hour weeks for roughly 50 weeks a year, that's the equivalent of $1,200.
I can't imagine spending $1,200 a year just to run my own mail server. The costs vastly outweigh the benefits. Especially considering that I'm far more likely to screw up all the rest of the security involved.
$20/year gets you decent email.
You can't escape by leaving on your own; you'd need to spur others to mass exodus.
Not my experience. It took ages to get a satisfactory setup (based on Debian's default config), but I've rarely had to touch it since. Admittedly I'm not exactly zero tolerance on spam.
http://www.google.com/apps/intl/en/terms/premier_terms.html "1.4 Ads. The default setting for the Services is one that does not allow Google to serve Ads. Customer may change this setting in the Admin Console which constitutes Customer's authorization for Google to serve Ads. If Customer enables the serving of Ads it may revert to the default setting at any time and Google will cease serving Ads."
I was tempted to pay for Office 365 and move away from Google but I'm not sure if it is truly a superior solution for my needs given Microsoft's habit of 0 support for Linux.
By why would they automagically scan your email and the like if they aren't serving ads?
Personally, I don't care how much data anyone has on me as long as it isn't someone capable of putting me in prison. :P
Choo Choo! Here comes the BULLSHIT TRAIN. All aboard!
For what it's worth I'm pretty sure the change was you, not Google. Google's business model hasn't changed. They are still an ad reseller, just like they always have been. They still don't sell your data, just like they never have.
Automatic account merging still doesn't happen, just like it never did. Google didn't link your friends accounts together, your friend did. And if you for some reason really, really want a different account for youtube and gmail, Google is perfectly happy with you doing exactly that. And there's somewhat reasonable multi-user support to help you do exactly that. And chrome's multi-profile support allows for even stronger walls between accounts. In the particular case of youtube they even let you use a different identity for youtube from the rest of Google, your "gangstakilla999" friend just didn't choose that option when asked when he was merging accounts.
It might be true in the general case, but Google did make a significant effort to require real names on Google+ (and by extension youtube / gmail). It didn't affect the large majority, but you can find some people it did affect. Google cared in some cases.
I have way more than 1 Google account, and I've had no issues with it at all. What evidence do you have to support your claim that Google is not happy with people having multiple accounts? Because the variety of ways Google helps you have multiple accounts is pretty strong evidence they are OK with this.
> Google did make a significant effort to require real names on Google+ (and by extension youtube / gmail).
Your "by extension" is false. YouTube never had the real names push like G+/gmail did. My account with a real name has always and continues to show a completely different pseudonym on YouTube, which was one of the options Google provided.
I'm yet to find an alternative that has such a good UI and sharing features.
My use case is sharing my family photos with my wider family. All photos are private. i'd like a self hosted option ideally.
Any suggestions?
It's so simple, just works, and new photos can be added and are automatically synced.
Anybody know what the author's referring to here?
If you requested Google blur your residence, Google just decided to invalidate your prior wishes, and you need to contact them again.
Google--some of us paid attention to the McCarthy Era in high school--hold a fun luncheon and go over that period in history with all your clever, brilliant employees?
I am waiting for DDG to get some real money and kick Google. And yes, you were Mila Kunis--now you're Kriss Jenner. "Your too controlling and I don't need that in my life now!" Yes, I finally saw Bruce stand up to her.
Still trying to grasp if this is serious or just a parody
There are other services that can’t even realistically choose between advertising and member-supported. Facebook is a great example: the utility of Facebook is directly correlated with how many people you know who are also using Facebook, and the only way to maximize that number is to make the service free, supported by advertising. Google is in a similar boat: the efficacy of search is in many ways tied to how many people are using search. Queries and clicks are the raw grist for Google improving its algorithm, and the more the better, which means making queries free.
I'd like to emphasize that latter point, "queries and clicks are the raw grist for Google." Approaches like PageRank and tf-idf[2] are only part of the story. A really fascinating and vital point is that the search user activity itself is an amazingly valuable source of relevance data. E.g. one can build a matrix over (search term, number of clicks for URL) as a search index. This supplements the other algorithmic relevance factors, is driven directly by human feedback.
Google employees have mentioned that they've worked on other aspects of relevance analysis based on user actions: it's possible to suss out situations such as: you clicked on a link, but clicked right back to the search results because that link sucked. Eventually you found a good result and it "stuck".
So user activity itself is very powerful source of search relevance and having more users just makes it work better. In fact, it's a bit startling to realize you can, in theory, build a search index without any source document data based solely on which links were clicked the most. In practice, we need to seed relevance with the purely algorithmic factors then improve it with the feedback data.
Given that, an incumbent search engine with a large user base has a big advantage over competitors: they own the base of user activity data. Google's advantage is not just the technology expertise that they've amassed over the years, but also the raw fact of its search market share leadership that any would-be competitor must overcome.
[1] http://stratechery.com/2014/privacy-dead/ [2] https://en.wikipedia.org/wiki/Tf%E2%80%93idf
They don't necessarily OWN the base of user activity, if others figure out how to sniff it out:
http://googleblog.blogspot.com/2011/02/microsofts-bing-uses-...
Once Chrome the browser became the seed of Chrome the OS, the writing was on the wall - Google didn't feel it needed to prove anything anymore, innovation turned to stuff it wanted, not what users wanted.
[1] Chrome, at least v25 and earlier, crashed parsing 10MB XML files - I have to test with much larger files all the time. At some point FF started to not suck, and recently it's got speed parity with Chrome for most actions on OSX. Not looking back.
