Re: os discuss (http://www.osopinion.com/perl/story/7263.html)

["Robert G. Brown" <bob@xxxxxxx>]

"Jeff Palmer" <somedamnfool@hotmail.com> write:
>no to take away  from anything else you said Bob... but are you sure about
>the more hardware platforms thing? .... it would seem to me that award
>would have to go to netbsd..... but.... maybe i'm wrong.... just
>curious...... and stuff....
You could very well be right about that, Jeff. Many of the non-hardware-
vendor UNIX implementations (and I include Linux and eCos as "UNIX" imple-
mentations) vye for this title from time to time.

>the only sizeable software company i have ever seen publish a release date
>of "when it's done" is apogee/3drealms... a game company... (dukenukem
>series games)
Which is honest of them.

>any software is only done when it's done....
And frankly, I don't need Eric Raymond or anyone else to tell me this,
since I've been in this businees for over a quarter-century and have seen 
the consequences of this being done, right and wrong. The best I can do to
each of my clients is that I will not be involved in decisions that rely on
software being delivered by a particular date, unless stated conditions are
met.

>Linus hade made statements giving general timeframes.... but... software 
>opensource or not isn't done until it's done...
Yes, that's right. In this sense, the Press generally behaves like bad IT
management. I've said "I think I can wrap up this project by next Friday,
if (one condition) and IF (another condition) are met." They don't hear
this, but act as if I've said "That project will be done next Friday".

I'm attempting to be honest, but I've learned to say "NO!" to such people
whenever they confront me with a date certain, or with a straightforward
question. THis is a LOUSY way to do business, 

>the leson out of this isn't for the open-source comunity, it's for the
>large software companies that publish a release date... then are late...
>and the software still isn't ready to ship when it does.. (microsoft....)
That's right, but in addition, the press covering this has a lesson to
learn, too. If MS (or anyone) is "late" with a "commitment", it is not 
a scandal. It's more like postponing a Shuttle launch. Despite everyone's
best efforts, the right conditions could not be produced to launch today.

There is a big difference between NASA and most large software development
(for-profit) organizations: Credibility. NASA has a reputation for being 
open and honest (with a few blemishes perhaps, but that's for another
thread), so when they say they can't launch, the press beleives it. When a
large software maker postpones a product release, people tend to challenge
the notion that the product could EVER be ready by the announce date.

Why is this? NASA launch schedules are created by a number of engineers,
who take into account the requirements for each mission, preparation time, 
assembly of payload, launch windows, etc. It is out there for all to see,
and if NASA kicks a launch down the schedule, it's easy to find out why.
This tends to suppress any notion of "scandal".

Large software makers generally do not announce release dates based on
engineering concerns, but plan product releases schedules out and expect 
the engineering people to meet that timetable. This is a manufacturing
mentality (like GM... the new car models are ready by a date certain), and
the reason for this is that, in general, the folks who are building the
next car model are not solving completely new problems.

The Press tends to treat Linux as if there is a "Linux Corporation", with
industrial-type schedules. So when someone like Linus hints that something
really new will be ready around Thanksgiving, the press interprets that as
a commitment. After all, isn't Linus like "The Chairman of the Board", just
like Roger Smith or GM?  When Linus and his merry kernel men feel there's
nothing to release in that timeframe, the press makes a fuss. Again, how
could GM not release a new model when they said they would?

                                                           Regards,
                                                           ---> RGB <---