On Mon, 10 Sep 2001, Rick Moen wrote:
> > Only those who have cheap junk modems, or who don't know how to set them
> > up properly.
>
> Well, look, folks, I hate to have to spell it out for you guys, but...
Please, spell. I'm anxious to see how good your english is.
> Have you ever looked, raw, at a binary file, e.g. cat'ed it to screen by
> accident? That sort of stuff is, of course, what comes whizzing past
> your modem every time you, for example, transfer a binary file over a
> telephone line. You will note that you can find just about any pattern
> you please of characters in there, if you scroll far enough -- rather
> like looking for patterns in clouds, except with less healthy exposure
> to the outdoors.
There's a couple of things wrong with this assumption.
1) The chance of a combination of binary code coming out in the exact
format of "+++ath0" is literally staggering. Winning lotto, by comparison,
would be an every day event.
2) IP transfers are, by design, inherently unencrypted. Which means that
if the text "+++ath0" is sent, that's exactly what you see. Exceptions to
this are, of course, things like ssh, IPSec etc, which perform some 3des
encryption on the enclosed text/data.
> So -- and I figure _some_ of you must surely see where I'm going with
> this -- a surprisingly large portion of the time, you can find odd
> little strings like "+++ath". Now, ask yourself, if modems were to hang
> up the line every time they happened to run across such a string by
> chance, would even the dimmest customers keep them for more than about a
> week before throwing them in the rubbish? I think not. Even junky,
> cheap winmodems aren't _that_ pathetic.
You are kidding yourself. Tell you what.
I'll run my sniffer on all my modem traffic for as long as it has buffer
space for. If you can find one instance of "+++ath0" in the capture
_except where it's specifically intended to be there_, like in this
message, I'll give you my modem.
> So, if you think modems all over the Internet are getting slammed off
> the line every one of the quite frequent times they encounter "+++" and
> "ath" sequences, you really need to think again.
These days, no. Most modern modem manufacturers actually pay Hayes an
appropriate licenseing fee, and avoid the bug. However, there are still
modem manufacturers who don't - and for whom this bug is still very real.
> But certainly you shouldn't take my word for it. By all means, have a
> blast trying to find and test modems that _are_ that pathetic.
It's not as hard as you think. In fact, I undertook just this exercise not
so long ago as a result of having exactly your argument with someone
elsewhere - and found myself red faced and with my foot in my mouth at the
percentage of modems which actually _do_ respond.
DaZZa
--
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug
