Beginner Question

[Gerard]

> On Wednesday November 21, 2007 at 11:52:58 (AM) Scott Silva wrote:

> on 11/21/2007 4:33 AM Gerard spake the following:
> >> On Wednesday November 21, 2007 at 04:16:04 (AM) Andreas Kasenides wrote:
> > 
> > [snip]
> > 
> >> I was very suddened by the discussion in the Postfix lists. In my view   this
> >> sort of thing should happen in the commercial software world, not on open
> >> source projects. I would appreciate an answer (even if it is obvious) from
> >> somebody that nows enough of the internals of MS.
> > 
> > I honestly do not agree with that assessment. If these were commercial products
> > with millions of dollars, or whatever currency you are dealing with involved,
> > I can guarantee that Wietse and Julian would have worked out there differences
> > long ago. Microsoft could easily write code that would prevent any but its own
> > applications from working correctly; however, that would be suicide.
> 
> Microsoft has done this in the past. They made windows 3.1 crash on DRDOS is 
> the first one I remember. They caused problems in a few versions of 
> Wordperfect in the Windows95 days, although Wordperfect might have shot 
> themselves in the foot on those.

1) Win31 did not crash with DRDOS; it issued a warning message that an action
could not be completed in certain circumstances. The error message was a false
flag since the operation was completed. I agree that Microsoft probably did
plant this in their source code; however, they were far from the only software
distributor pulling stunts like this in the early days of home PC use. In any
case, this is nearly a 20 year old event.

2) Word Perfect was sold to Novel who had no idea how to market the product.
While they floundered, Microsoft spend an estimated $10 million dollars to
study Word Perfect and improve on it. In the end, they produced the finest
word processor on the market. Word Perfect simply died, not because of any
overt act by Microsoft, but rather through neglect by its owners. Even though
Word Perfect has since changed owners again, it has never regained its
dominance in the market place. Word Perfect v. MS Office is a perfect example
of how the market place, and not some scummy government bureaucrat, should
drive software development. Unfortunately, it is all too often the other way
around.

>   Symantec
> > could develop its AV program to crash any MUA it did not approve of.
> Have you tried a recent version of Symantec 360?

1) Not personally. I use "ZoneAlarm Suite" on my Windows machines. I just did
a quick Googling for "Symantec+360" but did not find anything relevant.

>   That
> > would cost it money, something commercial enterprises are not to keen to do.
> > 
> > Yes, incompatibilities do arise from time to time; however, they are usually
> > attended to quickly. The old adage, "Money Talks" is relevant here. With no
> > monetary gain involved, or board of directors to contend with, the motivation
> > to remedy a situation declines proportionately.
> > 
> > Who is to blame for this diabolical situation is debatable. Postfix is the
> > program that Mailscanner is accessing, so therefore there could be made a case
> > that Postfix is the one to make the rules. Wietse has not made any overt
> > changes to his product that would prevent Mailscanner from working with it;
> He does change it regularly, although it is his right to do so, and it breaks 
> MailScanner often.

1) It is called 'updating' or 'product maintenance', if you prefer. There is a
world of difference between updating/maintaining a program and placing a bomb
in one. It is very magnanimous of you to give him your approval to code his
program as he see fit. You assertion that he is breaking Mailscanner is absurd.
If there was an API in place, then I would agree with you (perhaps); however,
sans such a path, Mailscanner is simply using a back door method to access the
output of Postfix. Wietse has stated that he wants Mailscanner to access
Postfix using SMTP as is commonly done by other applications that work with
Postfix. Julian has chosen not to go that route. If the two of them cannot
come to a mutually agreeable solution, then Wietse is free to do as he wants
without regards for Julian and Mailscanner.

> > however, there is nothing that would prevent him from doing so. It would seem,
> > at least from my view point, that an acceptable API should be developed,
> > similar to what Dovecot and Postfix achieved, to alleviate this problem.
> > Unfortunately, that would require a dialog between the parties involved. From
> > what I have deduced, that does not presently exist.
> If you look at some of the old dialogs, there was only one approved way to do 
> things, and that was Wietse's way. That is more of a diatribe than a dialog.

1) Again, it is his program. Has has the right to make decisions on how it
should interact or be accessed by other application. You either play by his
rules, or attempt to circumvent them. It must be remembered, that if you
devise you own unapproved methods then you are accepting full responsibility
for any and all actions that evolve. I am not saying that anything disastrous
will ensue; however, if it does, you have no recourse and no one to blame but
yourself.

> > 
> > Just my unsolicited 2¢.
> I see your 2¢ and raise a nickel.

Wow, a whole 7¢.

-- 
Gerard