OT: SCO Forum

[Fairlight]

On Fri, Jun 23, 2006 at 04:30:54PM -0700, Bill Campbell, the prominent pundit,
witicized:
> I suspect that the end of the support time is from the time of the next
> major SLES release, not the first release.

That's for Enterprise, though, yes?  Or do you mean all products from both
sides.  So when they release the next SLES, we'll see an EOL for 9.x pro?

> We're running apache-1.3.36, and haven't tried apache2 yet.  I generally
> avoid the Latest & Greatest(tm) until there's some compelling reason to
> switch (e.g. software that requires it).

There actually seem to be less modules available for 2.x than there are for
1.3 last I checked.

> >How do you judge whether a dist is ready or not when you're not running a
> >good percentage of their packages?  :)  Smiley included, and no offense,
> >but it's an earnest question in the end.
> 
> I don't are about their server packages that we don't use.  One of the main
> reasons I moved things to OpenPKG when we made the switch from Caldera
> Linux to SuSE was to minimize our dependence on the vendor's packages.

Okay, but then why would you prefer SuSE over any other vendor?  You've
cited good engineering in the past, but you're by and large not utilising
most of it.  The kernels can be far more authoritatively obtained
from kernel.org...  Why even use "a distribution"?  For that matter, why
doesn't OpenPKG have their own linux dist, in this event?  I can see the
portability advantages of the architecture, but it seems to partly devalue
one's evaluation of any given vendor's distributions when one isn't making
use of much past the basic core system (kernel, fileutils, binutils, etc.).
What's your advantage to using SuSE over any other dist, in this scenario?

> If SuSE Linux became unavailable tomorrow, it would take me a day or so to
> get everything running on a different distribution, FreeBSD, or Solaris.

I'd just find a new dist.  Given the unlikely nature of that event
transpiring anytime soon, I haven't lost sleep over it.

> The hardest part of a transition away from SuSE Linux would be redoing the
> automatic installations we're doing with autoyast now.

Haven't messed with autoyast.  I'm a big fan of having manual control
anyway, though.  I'm one of those people that always (with -one- exception
that I can think of, and that's my AV definition updates) chooses "custom"
for any installation, even for Windows Updates.  I don't like being out of
the loop.

> Again, that's why we're using the OpenPKG server components, and the
> Enterprise versions of SuSE.  Keeping up with a new distro every six months
> is too much of a PITA.

Been on 9.0 for several years.  It's that once every three or five years
you have to switch that gets to me.  It doesn't matter to me if they slap
the label "Enterprise" on it or not; if they have something out, by the
time it's mature enough to be comfy, it's already probably exceeded 1/6 to
1/3 of its lifespan.

> I know just enough php to be able to configure horde/imp/..., and really
> don't like the language.  They've taken the worst of perl's syntax, and
> obfuscated it by burying it amongst a bunch of HTML.

Ah, so it's NOT just me.  :)  I'll debug problems with other people's code.
Otherwise, it's ALP software--Avoid Like Plague.

> That type of API changing seems to me to be an indication of lack of
> understanding of basic design principles.  

Agreed.  Unfortunately, it seems like an increasing number of packages are
starting to become "infected" by this MS-type mindset.

> David Korn did a presentation at a Seattle Unix Group meeting several years
> ago, talking about the development of AT&T's uwin, a system to provide Unix
> capabilities in a MS-Windows environment.  Korn said that one of the major
> issues they had to deal with is that Microsoft Windows often has different
> APIs, not only between major Windows versions, but between patch levels
> within a major version.  The same system call might take different
> arguments and/or return different results.
> 
> When I took an advanced Samba class from John Terpstra, member of the core
> Samba development team and author of the Official Samba 3.0 HowTo and Samba
> 3.0 by Example, he said that the RPC protocols used in Microsoft's SMB/CIFS
> were as bad or worse than Korn's description.

That's truly frightening.

> I don't know the answer to that as I've never played with Office macros
> (the last time I did extensive macro programming was when I wrote Model II
> Scripsit macros to take VisiCalc output, and transform it to print on Radio
> Shack Daily Report forms :-).

Lotus 123, here.  I was proud to be Home of the 20-Screen Macro[tm]. :)
I programmed Lotus in 30min to take 15min to make revisions to an entire
project (46+ spreadsheets) what would have taken an operator most of
a day to do manually--and saved my boss significant embarrassment.  I
actually wrote a non-malicious worm, now that I think about it.  It was
a self-propogating macro that went through the entire set of sheets for
a project, embedded itself in the next sheet in line, ran, and moved on
to project completion.  Gee, I never really thought of it as a worm until
right now, but it pretty much was.  The application of such practises does
have its merits, I suppose.  It's the abuse of such features that's at
issue.

Had they just given me the correct matrix formulas in the -first- place...

"You know, math really -isn't- my strong point."

"Don't worry, I'll give you the formulas you need."

*time passes, presentation is completed over several days, Exec VP calls
FROM the trade show for which they were made*

"These spreadsheets are all WRONG!"

"I used the formulas you gave me..."

"I gave you incorrect formulas then.  Here are the correct formulas.  I'll
be by in an hour and a half.  Have the new booklets ready when I arrive."

<thinking>How can I do in 1.5hrs what took 3 days???</thinking>

*code, execute, grab pot of coffee and wish for a tranquilizer, print,
review, put feet up on desk and wait for Exec VP's arrival, deliver stack
of revised booklets, observe look of amazement and accept compliment*

That's pretty much how it went, too.  I used to really -love- 123 macros.
They were amazingly powerful.

> Since I run OpenOffice.org programs on OS X or Linux, I'm not terribly
> worried about macro attacks that infect Windows.

I'd be worried if they had an abstracted opendir/nextdir/unlink featureset.

mark->