[NCLUG] Cringely on Microsoft's plan for the Internet

[Sean Reifschneider]

On Thu, Aug 02, 2001 at 05:08:01PM -0600, Evelyn Mitchell wrote:
>The interesting part is the discussion of how MS could take over
>TCP/IP without anyone being able to stop them. - efm

Isn't most of this article talking about how bad the "new raw socket" stuff
in XP is?  You know, the thing that got so much press a month or so ago
until people pointed out that it's not "new" and that current Win 95/98/NT
boxes can easily do this too?

He says that MS can make their own protocol to run the net and say that
it's to fix the problems that they've been causing.  And get "100 million"
people to install it over the course of a year.

I find that pretty suprising considering that people seem to be getting
*LESS* likely to upgrade to new MS products -- how many people rushed
out and bought '98?  Remember businesses sitting back and saying they
weren't going to run out and upgrade to '95?  How many people do you
know running W2k?  Hell, the brand-new, high-end notebook I got a few
months ago didn't even come with w2k...

Now, he says that MS can close this new protocol to other vendors, making
it impossible for Sun to implement it.  *BUT*, he says that the
infrastructure changes aren't going to be that big -- just flash upgrade
some Ciscos.

HelLO!  Why would Cisco get it but not Sun?  Don't forget that a number of
those infrastructure devices are running Open Source, which he claims will
not be able to get access to the new protocols for a few years...  "Yeah,
you can upgrade your workstations to TCP/MS, but we'll have to buy a new
Cisco PIX firewall to replace it.  So, $3000 for Windows upgrades, $18,000
for the firewall upgrade.

Oh, and I think he should examine his idea about having the authenticated
connections a bit more.  So, he won't execute attachments people send him
unless the sender is authenticated -- giving an audit-trail.  It seems to
me that it would make it *EASIER* for these worms to spread.  Sure, he
won't execute an unknown attachment from somone he doesn't know, but what
if one of his friends does?  "Oh, I got this attachment from Bob, and it's
authenticated...  It must be ok..."

Love his idea about closing down things so tight.  "No program other than
my mail program should be able to make outbound SMTP or POP connections, or
talk to my address book."  Except that "automatic dialer" program, the
caller ID program, my PDA conduit program all no longer function...  He's
not talking about a few small tweeks to the current software, he's talking
about a complete redesign of quite a number of fundamental parts of the
OS...  Which most people will disable anyway because they just end up
clicking "ok" to all the dialog boxes that come up anyway...

The simple fact of life is that people *LIKE* the conveniences of having
things all open.  They *LIKE* that from their spreadsheet program they can
say "Hey, mail this to somone", and it looks like there's a mail program
built in.  They like all their software tightly-coupled.  Right, Emacs
users?

MS already provides ways of not running attachments, executing controls and
that sort of thing.  It's not helping.  There are also ways *RIGHT NOW* of
authenticating e-mail that you get and the like.  If they're such a silver
bullet, why aren't people using them?

The idea that MS may be promoting having a system that allows the easy and
fast replication of viruses, to allow them an opening for other things is
interesting.  What I find MORE interesting is that it also leaves an
opening for everyone else, and it's actively promoting people to look at
other alternatives like Linux, legislation, etc...

Mostly I think the article in question is a crock.  It's probably worth
examination to see if there is an opening that the Open Source community
can exploit before Microsoft does...

Sean
-- 
 I think you are blind to the fact that the hand you hold
 is the hand that holds you down.  -- Everclear
Sean Reifschneider, Inimitably Superfluous <jafo at tummy.com>
tummy.com - Linux Consulting since 1995. Qmail, KRUD, Firewalls, Python