2 ms·

I don't think it is unreasonable. I think we've fallen into a culture of low expectations regarding the ability of "normal users". Many people will (perfectly l

I don't think it is unreasonable. I think we've fallen into a culture of low expectations regarding the ability of "normal users". Many people will (perfectly legitimately) choose convenience over quality or reliability, but I think we should be pushing users to engage more seriously with the tools that they use. We can certainly do more as developers to make things easier, but I think we have a tendency to pander to an imaginary "AOL granny". We often present oversimplified advice, because we expect too little of users. To pick a random example, I think we're too eager to say "use Dropbox" rather than "use Dropbox, but buy yourself a NAS too because X, Y and Z". I think that there's a substantial latent desire among ordinary users to have more control over their technological lives. People are increasingly worried about the power of big tech companies, but they don't know what to do. We aren't doing a good enough job of informing users; those big companies have an obvious incentive to keep users uninformed. We're not communicating the risks and benefits well enough, we're not providing clear explanations of the alternatives to the Big Five. In the process of writing this comment, I've had several startup ideas. I hope that the people reading this are having ideas of their own. I think that there are substantial opportunities to start returning control to users.

xyzzy123 10y ago
For a whole variety of reasons, you need know what you're doing to properly host your own email now. From the risk of domain name takeover, to hosting provider takeover, to modern spam requirements (ip reputation, dkim, spf, dmarc), to just plain sysadmin ability and ideally run an HA system. Agree that there are probably some startup opportunities, but the the reality is that you'll just be transferring the user's trust from AmaGooFaceSoft to you. You might be more trustworthy... or you might go out of business. EDIT: To be fair, I'm probably missing a whole bunch of creative solutions that give the user control without requiring them to run a service.
- jdietrich 10y ago
  There are a whole spectrum of options between "running a mail server in your basement" and "having to reprint your business cards if your Google account is suspended". We can make big steps towards trustworthiness and control in SaaS, it's just often orthogonal to profitability. Lock-in is the obvious example. Lots of vendors expend huge efforts to simplify data ingress, but create arbitrary obstacles to data egress. We worry about SaaS providers going bust because we're inured to the idea that getting your data back is really difficult. The interests of the user do not align with the interests of the vendor. Predatory practices are so normal that we've half-forgotten that they're predatory.
- kuschku 10y ago
  Quite a few you're missing indeed. Many ISPs give out a domain name and email service for free with internet contracts — you can set it to redirect to gmail until you need it, then just switch it over. Tada, problem solved.
  - ucaetano 10y ago
    Until your ISP closes your account for whatever reason.
    - kuschku 10y ago
      That’s why you use a trustworthy ISP. My current ISP, for example, is actually a local company providing datacenter colocation and consulting services to the state’s government and local and national companies, but they also run a small ISP as side business.
      - ucaetano 10y ago
        >they also run a small ISP as side business Which means their small ISP business is something they wouldn't mind so much getting rid of, along with your account. The point is, you'll always depend on someone else. Your data will always flow through someone else's network. The cloud is someone's computer, and the internet is someone's cable.
        jdietrich 10y ago
        You always have to rely on someone, but you can structure your affairs so that you're only temporarily reliant on them to provide service today, not to provide service indefinitely. If your email address is "me@gmail.com", you're stuck with Google. If they suspend your account, you're in big trouble. If your email address is "me@myname.com", you're a free agent. You can use Gmail or Fastmail or some random hosting company or a mailserver on a Raspberry Pi. You can use DNS or forwarding to send your mail wherever you like, you can transfer your domain name to any registrar you like. Rule zero. If all your files are in Google Drive or Dropbox and nowhere else, suspension of your account could mean total data loss. If you have a local backup on a NAS or a mirror on Glacier or Backblaze, you're just mildly inconvenienced. Rule two. Eliminate single points of failure wherever possible. Avoid vendor lock-in. Don't rely on a service unless you have a contingency plan if that service fails. Treat everything as if it were hopelessly unreliable.

xyzzy123 10y ago

For a whole variety of reasons, you need know what you're doing to properly host your own email now. From the risk of domain name takeover, to hosting provider takeover, to modern spam requirements (ip reputation, dkim, spf, dmarc), to just plain sysadmin ability and ideally run an HA system. Agree that there are probably some startup opportunities, but the the reality is that you'll just be transferring the user's trust from AmaGooFaceSoft to you. You might be more trustworthy... or you might go out of business. EDIT: To be fair, I'm probably missing a whole bunch of creative solutions that give the user control without requiring them to run a service.

jdietrich 10y ago
There are a whole spectrum of options between "running a mail server in your basement" and "having to reprint your business cards if your Google account is suspended". We can make big steps towards trustworthiness and control in SaaS, it's just often orthogonal to profitability. Lock-in is the obvious example. Lots of vendors expend huge efforts to simplify data ingress, but create arbitrary obstacles to data egress. We worry about SaaS providers going bust because we're inured to the idea that getting your data back is really difficult. The interests of the user do not align with the interests of the vendor. Predatory practices are so normal that we've half-forgotten that they're predatory.
kuschku 10y ago
Quite a few you're missing indeed. Many ISPs give out a domain name and email service for free with internet contracts — you can set it to redirect to gmail until you need it, then just switch it over. Tada, problem solved.
- ucaetano 10y ago
  Until your ISP closes your account for whatever reason.
  - kuschku 10y ago
    That’s why you use a trustworthy ISP. My current ISP, for example, is actually a local company providing datacenter colocation and consulting services to the state’s government and local and national companies, but they also run a small ISP as side business.
    - ucaetano 10y ago
      >they also run a small ISP as side business Which means their small ISP business is something they wouldn't mind so much getting rid of, along with your account. The point is, you'll always depend on someone else. Your data will always flow through someone else's network. The cloud is someone's computer, and the internet is someone's cable.
      - jdietrich 10y ago
        You always have to rely on someone, but you can structure your affairs so that you're only temporarily reliant on them to provide service today, not to provide service indefinitely. If your email address is "me@gmail.com", you're stuck with Google. If they suspend your account, you're in big trouble. If your email address is "me@myname.com", you're a free agent. You can use Gmail or Fastmail or some random hosting company or a mailserver on a Raspberry Pi. You can use DNS or forwarding to send your mail wherever you like, you can transfer your domain name to any registrar you like. Rule zero. If all your files are in Google Drive or Dropbox and nowhere else, suspension of your account could mean total data loss. If you have a local backup on a NAS or a mirror on Glacier or Backblaze, you're just mildly inconvenienced. Rule two. Eliminate single points of failure wherever possible. Avoid vendor lock-in. Don't rely on a service unless you have a contingency plan if that service fails. Treat everything as if it were hopelessly unreliable.