svn commit: r39233 - head/en_US.ISO8859-1/books/handbook/ports

[Benedict Reuschling]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am 19.07.12 12:22, schrieb Doug Barton:
> On 07/19/2012 03:05, Benedict Reuschling wrote:
>> Am 19.07.12 11:17, schrieb Doug Barton:
>>> On 07/19/2012 02:12, Niclas Zeising wrote:
>>>>   Suggested by:	Roger Magana <ram0042 [AT] hotmail [DOT] com>
>>
>>> Please don't obscure e-mail addresses in commit messages. It does
>>> nothing to confuse the spammers, but it does make it more difficult to
>>> see who has contributed what over the years; not to mention copying and
>>> pasting addresses into e-mails for responding to those contributors.
>>
>> I'm a bit surprised by this.
> 
> I'm not sure why, since you've been around long enough to have seen me
> ask people not to do it half a dozen times at least. :) I don't do it
> every time obviously, just periodically.
But isn't that highly inefficient in trying to achieve the goal? Instead
of contacting individual committers and asking to change their
behaviour, why not start a thread on developers about it or ask core
about defining a global policy that committers must adhere to?

> 
>> I've been doing this email-obscuring thing
>> as long as I'm with the project. My mentor did it also, and as far as I
>> can see it is done virtually by everyone,
> 
> That's not even close to true.

Everyone is a strong word and hard to verify, how about "most committers"?

> 
>> We probably should discuss this on a project level if there are reasons
>> not to do it (like you said, that spammers might not be driven away by
>> it)
> 
> It isn't "might." It's been true since day 1 that there are no patterns
> which are recognizable by humans that the spammers cannot trivially
> harvest. All obscuring games do is make people feel better, and in our
> case reduce the amount of useful information.

True, humans can do the reverse-lookup (if you would describe it that
way) fairly easily. But spammers don't surf the web, they use
bots/computer programs because they can harvest more pages more quickly,
but are having a hard time (unless they are told what to look for) with
obscure emails. Of course, you can teach a program to recognize these as
well, but that is not the point.

> 
>> and find other (read: better) ways to balance the attribution of
>> contributors in commits and not making them spam-targets at the same time.
> 
> There isn't one. If you publish an e-mail address on the Internet, it
> will get harvested.
Yes, that it true and a fact we have to live with at the moment. But
aren't you also trying to harvest the email address, but for a different
(non-spam) purpose? We cannot assume everyone has good intentions when
we publish information for everyone to see.

Let's say we omit the email address completely from the commit mails and
just attribute the submitter by name. Now, you wrote you need to do copy
and paste if you want to contact that person. How do you do it in that
case, go into the actual PR and grab the mail address from there (which
is not obscured in GNATS btw and is a much better hunting ground for the
aforementioned spam harvesters)? If that is so, then propose such a
solution so that all committers will comply with it when there is a
common consent.

Maybe it also gets better with the new bug tracking system, but I
haven't had the time yet to look into it further.

Regards

Benedict
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.8 (Darwin)
Comment: GPGTools - http://gpgtools.org
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAlAIJsAACgkQTSZQLkqBk0hwLwCfWBKJLKGxrfvUuuOsx3uh5FQ0
tqgAn0SGH0NGnBfHp8sN0HVPcHwbyw3Z
=AkiP
-----END PGP SIGNATURE-----