In message <48096A22-98BD-49F3-8D80-713070256165@frobbit.se>, at 21:26:22 on Sun, 30 Jun 2013, Patrik Fältström <paf@frobbit.se> writes
In the paper I wrote that was not accepted I explained exactly what Lars-Johan explain above, and concluded that as SMTP and IMAP connections are involved data about the messages will be retained.
I.e. webmail consists of one "web" transaction between client and web server, and one email transaction from the web server to the mail server. If now mail is to be retained the mail transaction has to be retained. As web transactions are not to be retained that leg should and will not be retained -- even if it is a web transaction that as a result will generate email.
Whereas I am suggesting that a suitable compromise would be to retain the details of transactions between client and web-servers-known-to-be-mail-systems. Like Hotmail, Gmail and so on.
That was not acceptable by the parties that to me obviously where not happy with the result during the discussions in the European Parliament where it to be at least was clear that web should not be retained.
And I get somewhat bad taste in my mount when I heard people after decision is taken in the European Parliament that web is not to be retained still try to retain "some" web transactions.
I don't think that an "Internet e-mail service" should be excluded simply because it takes place on port 25. They are only very tenuously "web" transactions[1] anyway, rather than http[s] transactions. And no, this isn't splitting hairs, it's trying to assert some clarity in the terminology. The lack of clarity isn't helped by organisations like the BBC constantly inviting viewers to "email us via our website". By which they mean fill in a web form, which might or might not then get successfully emailed to someone inside the BBC by the web server. NB. In this context, I'm not advocating logging and retaining which pages at http://news.bbc.co.uk people browse to, although I know some people who would, especially if it's http://www.howtomakeabomb.com ps The reason I'm especially interested in this is that the list of data types in Article 5 is partly based on some work I did in 2001 (and subsequent pre-Directive Data Retention laws in the UK). [1] I can't remember the last time I clicked on a hyperlink pointing to www.gmail.com, it's either an app on my Windows task bar, a different app on my Android phone, or a parameter entered into my POP3 client (yes, you can access gmail by POP3 too, as I'm sure you know; and why does the access protocol make any difference to whether the transaction should be logged/retained, either in terms of common sense, or what it says in the directive?) -- Roland Perry