C[omp]ute

From: andrew cooke <andrew@...>

Date: Mon, 23 Apr 2012 20:20:01 -0300

I was brainstorming some ideas for a "worthy" project; I finally decided it
wouldn't work, but thought I might as well write things down in case I have
a change of heart.


The idea that email needs "fixing" seems to be common at the moment (it was
included in a list of "problems" by Paul Graham).  Now, personally, I manage
my email locally, because I am unhappy with Google (or anyone else) having
access to so much information.  And it works quite well.  So the core of the
idea was that I could make that approach available to others, packaged in a
way that didn't require any expertise.

Email would be pulled from existing mail providers over IMAP and stored
locally.  There would be an embedded SQL database for search (like mairix).
The client would probably be in the web browser, running against a local
server.

Taking things further, you could extend the client to automate encrypted
email.  The idea I see working is based on ssh - you don't try to guarantee
that the initial key exchange is perfect, but you cache it and warn of
changes.  So every email would include a private key; these would be extracted
and cached by my software when it receives email from others; sending email to
people with a known key would automatically trigger encryption; a change in
keys would flag a warning.

There were some more ideas about UI, implementation, and searching /
cataloguing email, but that's the general idea.

But there are two problems.

First, sending email requires an SMTP gateway.  You can't just send email from
your own machine these days.  And while you can pull email from web providers
you cannot push it.  So there would need to be a central SMTP server.  That's
not so terrible - adding a central IMAP server for receiving email would help
avoid sharing data with the big players, and you could imagine people paying
for this service.

Second, and more seriously, I realised that I was stuck thinking of a PC-based
solution.  And really, these days, it needs to support mobile devices.  Which
don't have the resources to do this.  Email really does have to be in the
cloud, in a sense.

Coincidentally, the title "From Personal Computers to Personal Clouds" caught
my eye -
http://www.windley.com/archives/2012/04/from_personal_computers_to_personal_clouds.shtml
I haven't read the article, but given the above you can see an argument for
some kind of personal cloud platform...

Andrew