[Cryptography] Interesting discussion of Web 3.0 ...

[Jerry Leichter]

> I don't think the claim is that technological innovation favors
> centralization but that *users* do. You run your own SMTP server and so
> do I ... but the vast, vast majority of the *techies* I know gave that up years ago when Gmail's spam filtering got good and they show no interest in returning.
> 
> All this stuff is insanely complex and getting more so, and every bit of
> added complexity means more people will want someone to do the hard
> stuff for them. And that, inevitably, brings centralization.
In fact, this is now moving to the next level:  Most *businesses* no longer want to run their own servers.  Mail?  Use gMail or Exchange - not on premise, but letting Microsoft manage it in Azure.  Web servers?  They're all out on the Cloud somewhere. Even development machines are increasingly out in the cloud.

Now, there are two ways of looking at this.  The first is that it's an inevitable dynamic - there are things about the technology that lead it to be easier and better to be run centrally in a few places.  There was a time when factories had their own power plants.  There are perhaps examples still out there, but they are pretty rare - better to get your power from a company whose expertise is generating and distributing power.

The second explanation is that we just have not built the technology to make it feasible for people, or even businesses, to run their own servers.  It's just too complicated - and perhaps it doesn't need to be.  Think about the complexity of editing photographs.  Before digital photography, you needed special equipment and a great deal of experience to accomplish even fairly simple stuff.  Then when photography became digital, much more became practical - but initially it took great expertise.  It still does, if you want to do really complex stuff - but there are consumer-grade programs that are easy to use and can easily accomplish things that only a handful of people could have done not so many years ago.

Mail servers no longer require you to understand a multi-hundred-page book of arcana for getting sendmail to work - but if you consider the effort required to configure even a simple mail server today versus the effort to apply some very sophisticated algorithms to optimize a photo you took ... well, we as an industry haven't really been trying very hard, have we?

What would be needed is perhaps changes in technology - yes, with solar power and improving batteries maybe a centralized power grid isn't quite as essential as it used to be - but also design of programs and systems to make the necessary tasks easy.  The only serious effort I know of to do with was Apple's Server versions of its software - which they've pretty much abandoned.  Apparently there was insufficient interest; or maybe it just wasn't good enough.

                                                        -- Jerry