From: Simon McClenahan (SMcClenahan@ATTBI.com)
Date: Fri Mar 15 2002 - 12:36:00 MST
----- Original Message -----
From: "Eugene Leitl" <Eugene.Leitl@lrz.uni-muenchen.de>
To: <extropians@extropy.org>
Sent: Friday, March 15, 2002 6:44 AM
Subject: Re: [META] email clients, YET AGAIN
> On Thu, 14 Mar 2002, Simon McClenahan wrote:
>
> > If you did this deliberately, then I would like to bitch back at you,
>
> Of course I did that deliberately. I wasn't even nice enough to put an end
> at the end of my message, thus screwing up the rest of the list digest
> subscribers who also happen to be Outlook users.
If you also did that knowing it would affect some digest readers, then you
are an asshole^H^H^H^H^H^H^H^H^H^Hunethical. Why discriminate against users?
Exploitations like that are not going to make them change anything.
>
>
----------------------------------------------------------------------------
----
Yeh, thanks so much for ending with an "end" that time, maybe you're not
such a bad guy after all. On behalf of all of us poor saps who currently use
Microsoft mail readers, we thank you.
> If people don't talk to me in plain text they obviously don't want to talk
> to me.
No, they do want to talk to you. Even if it's in broken English with poor
grammar and speelling misteaks they want to talk to you. It is you who
chooses to punish those who don't conform to your idea of how things should
be. We may disagree on few or many things, but the communication transport
is always open for us to express ourselves. You choose to sabotage that
transport by knowingly using those bugs.
> > If you did not do this deliberately, then to answer your question, most
> > mail readers (like Outlook Express) that send unformatted text are
> > configured to wrap lines when the message is sent. Some filtering may
>
> Um, so does my mailer. However, it
>
doesn'tbreakoverlongtextlinessuchasthose,whichisveryobviouslythecorrectproce
duretodoandnotwhatMSOutlookandafewotherbraindeadMUAsdo.
> An URI is just a special case for this.
>
> Meaning, if your MUA does this, it's broken. Simple enough.
An example of a URI with characters such as "?" and "-" would have
illustrated your point better, but nevertheless I agree with you. When
Outlook Express sends a message in plain text, it can be configured to wrap.
While I'm composing this message, everything looks fine and is wrapped
within the compose window with a horizontal scrollbar for that really long
word you created above. Wrapping is used to accomodate clients that don't do
their own wrapping when rendering the message. Maybe I should turn off word
wrapping from now on and see if anyone complains.
> The question of wrapped URIs is undecidedable. It requires a human mind to
> fix a preventable problem. I don't want my MUA come with a built-in mind.
> I want something which renders plain text messages. Which would run on a
> 32 MByte wearable, and render directly to an array of luminous pixels in
> my hud.
But we see several ways to render plain text - wrapped and unwrapped, URI
detection, email address detection, general hyperlink detection, etc. The
URI detection for example is a feature of the application that allows you
(hyper-)link to other text. If you don't want that feature in your HUD,
don't install or enable that component. You'll still have your split URI
problem, and you don't need a mind to join the lines together, you just need
an algorithm.
> Wrong attitude. I have a computer, I don't have to cut&paste, and to edit
> the URIs. I've shown above why the brokenness is at the end of earl
> wrappers, not at my end. I don't have any possible use for wrapped earls.
The problem is not with the wrapping, it's the detection of the full URI.
Your MUA's URI pattern matcher algorithm doesn't account for line breaks in
the raw text, simple as that. If you use a fixed-width character-cell
renderer and it is displaying a URL that's 200 bytes long, it doesn't matter
whether the renderer wraps it or not on the screen, so long as you get the
utility from it when you use the hyperlinking feature in your MUA. If
"clever" MUA's sent you a pre-formatted URI with line breaks encoded in it,
then approximately the same amount of cleverness on the renderer's side
probably could decode the true utility of it.
> > Looking at the mail headers of Brian's message, there is no X-Mailer:
> > header which is a de facto standard for specifying the MUA used to
>
> Where is the RFC requiring that to be inserted? Why do I need to specify
> my MUA, so that you can select the proper exploit from the vulnerabilities
> library? I think not.
How very unFriendly of you :-) I hope the first superhuman AI is not based
from any of your uploads ;-P. Why would I or any other Extropians want to
target you? Why do you want to target people with broken MUA's?
> Yes. I'm upgrading to mutt (currently pine user), as soon as I can get the
> DSL home box finder to work. The next evolution of this is a web front end
> to a SQL database containing a few millions of individual mails. It's just
> that this would take more time to set up than I have. In case anyone here
> has created something like this for their private use or know of any such
> solutions, please drop me a pointer.
More or less, "me too."
> > HTH.:-P
>
> EHFD, FOAD. HAND ;P
WTF?:-/
cheers,
Simon
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 09:12:59 MST