Re: tlug: Re: djb [was: ibm.net with LINUX (Red Hat)]

[Karl-Max Wagner <karlmax@example.com>]

> You do if you've got crackers or email pirates in your machine and
> that's all you've got to go on to choose the software to crush them
> with.  But that was exactly my point.  Karl-Max said "Dan's right, the
> old programs suck, the old protocols suck, and he built a new

And I'll keep doing so. It reminds me of my years of being a Packet Radio
Sysop. We used "The Box" by then, using mostly a W0RLI protocol,
which is similar to RFC821 / RFC822. We had no end of trouble.
Then we switched to F6FBB-Box, which is in many ways similar to
qmail. Almost all trouble just vanished. No more crashed
mailqueues etc. In other words: I've seen that all already. And
I know from my own experience that djb is right - every word of
it. Actually, I wonder whether djb knows the F6FBB stuff....

> 1.  arrogance is costly (if I had anything except advice to

Well, the guy who pays is just not djb, but rather the one that
refuses to listen and ponder. Ultimately his problem, not
djb's....

>     contribute, I'd be a lot less arrogant and a lot more persuasive); 
>     this is a minor theme, though 

I don't consider him that arrogant in fact....

> 2.  technical progress is not the be-all and end-all; there is such a
>     thing as good enough, and in the real world you continue using

What is "good enough". Good enough today is totally inadequate
in five years. 10 years ago FAX was "good enough", today email
is a requirement. And so forth.

>     systems that work---Karl mentioned "current requirements"; well, a 
>     hacked smail satisfied my then-current requirements, and if I had
>     screwed up the qmail config or the smail queue recovery, I would

It's actually difficult to screw the qmail configuration.

> and I still don't see how QMTP can be both more forgiving and
> contribute to greater reliability of the overall system (ie, the
> internet).  Allow people to be sloppy, to code by example, to borrow
> from reference implementations that may not be accurate reflections of 
> the standard or may fail to handle important special cases---OK, the
> QMTP link may stay up, but what if that stream simply gets cat'ed to
> the next (non-QMTP) program?  (Been there, done that ;-)

In that case you definitely know that something is wrong because
the protocol has been violated. Just drop the connection and
report an error.

> the time qmail was too risky _for me_ to use).  I simply quoted him to
> support the statement that legacy systems and legacy languages are
> critically important in the Real World.

They are only critically important to those that failed to take
the right decisions at the right time. Particularly in the case
of COBOL it is obvious that a lot of people failed to do so.
They deserve the mess.

                                Karl-Max Wagner
                                karlmax@example.com
--------------------------------------------------------------
Next Nomikai: 18 September, 19:30 Tengu TokyoEkiMae 03-3275-3691
Next Meeting: 10 October, Tokyo Station Yaesu central gate 12:30
--------------------------------------------------------------
Sponsor: PHT, makers of TurboLinux http://www.pht.co.jp