The old good email remains the most critical digital communication tool. What makes the venerable email so useful and sustainable over the long time is its openness and standardization. Email is radically different from the modern "apps" which integrate all pieces of technology--the server, the client, and the protocol--by a single monopolist provider. With email, we are free to choose the server (provider) and client with any combination. It provides enormous flexibility, added privacy and security. Indeed, the provider does not control my client and cannot add backdoors; there is no monoculture of client software with all the related security risks (any security vulnerability is global). Email is one of the few pieces of technology that is very resistant against internet censorship. Repressive state can easily block a web site and even force an app store to remove an app (as the Navalny's "Smart Voting"). Also, an app store can delete it for any other bizarre reason. But it is much more difficult to block a mailing list: it is easy to redeploy and recreate it on a different server (without the users even noticing anything). Furthermore, The user can easily create several different email-based identities (e.g. a separate one for politically sensitive activity) which adds anonymity. And anonymity means physical security in some countries.
It is not surprising that many internet services use the email address to register users, authenticate, restore password and other similar purposes. Open, standardized and decentralized email is one of the most critical technology everything else depends on. After all, the flexibility offered by the email technology--the freedom to choose all pieces (provider, client etc.) is just very very handy, at least for an advanced user (you can add new features on top of what the provider realized, even against the provider's will--isn't it convenient?).
The whole email technology is build around open protocols rather than a centralized platform. This facilitates competition, makes for better and fairer service and reduce possible impacts of malicious monopolists (Masnick, 2019).
Google's Gmail has long been one of the main pillars of email, millions used to rely upon every day. We should praise Google for popularising email as the basic mainstream technology among the masses. I started using Gmail many years ago when it was in its "beta" and available only by invitation. At that time Gmail openness and unrestricted nature was just blazing. The web interface was lightweight and not really cluttered with ugly banners, unlike other email providers. There were ads but they were small and unobtrusive. Gmail had long supported all the basic protocols (POP, IMAP, SMTP) that allowed to use any standard compliant client software, and that was available for free (some other providers were more greedy and allowed this only on paid plans). Google's POP, IMAP and SMTP implementations have been (and still remain!) quite idiosyncratic, incomplete and not really standard-compliant which caused various glitches (e.g. message deletion and default sorting are weird, I always hated Gmail's labels). But this was bearable.
The serious privacy problems and threats of Gmail, such as user email scanning for context-specific advertising (until 2017) or AI tool which could provide access to some pieces of data to third-party developers. That is nearly a disaster that cannot be fixed because spying on the user's data is at the heart of Google's business model. But who cares as long as it is free! I have long been using and promoting PGP encryption which could fix many of the privacy (and security) problems. Yes, PGP is crucial for individuals and businesses and yes, a motivated user can encrypt.
Gmail still remained free and relatively open while an alternative of deploying private email server is time-consuming and tedious (e.g. ensuring that emails from a tiny private server don't end up in spam folders of intended recipients). I used to pay with some of my privacy to get the usability and stability of Gmail.
But over time I became increasingly concerned about the clear trend taken by Google to make the open email more and more difficult to use outside of the Google monopolistic ecosystem. There are signs of the famous embrace, extend, and extinguish strategy. Gmail API is featureful and powerful... but only if you really need the complexity and like to play with the Google rules. If you don't like to see ads, for example, and for this use a standard IMAP mail client of your choice, your must suffer. If you need full PGP support on a mobile client, never offered by Google, you are out of luck and have to use an IMAP-based mobile app like Android K-9 Mail that requires sacrificing some usability.
Google tends to draw its users by all means into its browser, its own apps and APIs to get more user's private data and show ads. For that matter, Google's security usability has become just terrible. The intrusive access-blocks when a mobile user with an IMAP client moves across IP addresses can drive anyone crazy... Access can be blocked even if the user switches just to the next IP address within the same provider's IP pool.
I have to use VPN with fixed IP address to avoid these stupid blocks!
To help keep your account secure, Google will no longer support the use of third-party apps or devices which ask you to sign in to your Google Account using only your username and password. Instead, you’ll need to sign in using Sign in with Google.
The Google's insistence on rather complicated and heavyweight OAuth2 mechanism for basic email client access (remember, most email programs do not require you to enter your password every time, diminishing the risk of phishing) is understandable only as a means to limit all uncontrollable third-party clients. Yes, OAuth2 is logical for complex workflows of data access delegation across multiple web-based services with different login/password combinations (the "Auth" stands for authorization, not authentication). Whenever I need access to my own emails I need to authenticate my identity granting full access. But isn't OAuth2 client secret kept on the device just as the username/password combination? Yet, limiting the (power) users access to their own data provides just an illusion of security at a large cost to usability and compatibility.
The Google's move to OAuth2 authorization seem to point that the Gmail-hosted emails do not belong to me any more. My emails are now owned by Google, who just "authorizes" (delegates) me access to some of the data without trusting me. This is not what I need from my private communication. Does Google pretend to "zero-trust" any third-party apps? Maybe it doesn't trust its users (the owners of their data), assuming they are all idiots?
If you think your users are idiots, only idiots will use it [your service]. --- Linus Torvalds
And there is another side effect: as Google increasingly deployed more and more heavyweight frameworks and technologies, Gmail became very sluggish and bloated. It is cluttered and confusing, especially to those who don't use it often enough to remember all the idiosyncrasies. And it's still poorly adaptable to the user's needs. How can I get a fixed-width font for my plain text message? Where is my favourite basic (and very fast) HTML web interface?
Enough is enough. I now go away from Gmail, and primarily not because of big privacy concerns (which is quite expectable) but because of deteriorating usability and growing incompatibility. It looks like the people at Google have forgotten their old motto "Don't be evil." While I have been paying Google with my privacy currency in the past to get functionality and usability, the benefits of Gmail continuously went lower and now reached an unprofitable level.
Migadu is my choice
There are many hosted email providers, some are focused on privacy and security. For example, Protonmail is a fantastic project that makes it nearly trivial to use PGP even for an uninitiated. But its drawbacks are that it is non-standard and has too high publicity making it quite undesirable in certain authoritarian countries. Simply said, if you use Protonmail in some countries you may be suspected; Protonmail can be blocked by the authorities, and worse still, blocked in quite idiosyncratic way. Some services may also reject registration using this service.
What I have finally chosen is Migadu. It is not yet another standard email hosting provider. It is a domain-based service. Once you have got your own domain name (domains are now cheap), you can make your own email service for your domain. That simple. This makes it super useful for companies, families, groups and NGOs without large budgets. For a reasonable price you get nearly your own mail server with many configurable features (any custom mailboxes, aliases, forwarding, regexp, webmail, etc.) but without the need to maintain all this complex system.
If you have a web site, you necessarily get a domain name for it. Now it's easy to get your own email identity. True that some hosting providers also do host email. But if you decide to switch to a different hosting it will create a trouble: you need to move also email and this fact strongly limits your next choice. Having a completely indpendent email system for your existing domain avoids such hoster lock-in and makes life much easier.
By the way, the Migadu standard webmail interface is sleek and very simple. Looks modern but lightweight and quite fast. No bloat whatsoever, only the most crucial functionality. I am not big fan of web-based email, but use it from time to time. And there is even some very basic support for PGP! (But remember that web-based PGP is not a very secure solution.)
I found the mail server configuration (including more esoteric stuff like DNS setup and DKIM signatures) very easy. In my view you do not need an IT degree to configure your email server with full functionality. I like the admin panel, it is minimalist and easy to use, no stupid and distracting visual effects. And Migadu is advertised as fully open standard compliant service without proprietary glitches and limitations. So any standard (open source or closed source) software is very likely to be fully usable. This freedom is very important. And they are also clear and honest about the limitations and drawbacks.
Finally, goodbye Gmail.
- Masnick, M. 2019. Protocols, not platforms. A technological approach to free speech. Knight First Amendment Institute https://knightcolumbia.org/content/protocols-not-platforms-a-technological-approach-to-free-speech.
PS: Disclaimer: I have no links with Migadu.