Summary: What's wrong with SCO (long)

[Sean Eric Fagan]

In article <1991Apr17.210627.4517 at beaver.cs.washington.edu> pauld at cs.washington.edu (Paul Barton-Davis) writes:
>I support the notion that for word-processing it might work, but for
>systems programming and serious Unix users, its a joke.

Uhm, I consider myself a serious systems programmer and unix user, and I
*like* sco's unix (3.2v2).  It's stable (I had kithrup up for three months,
before I decided to reconfigure the kernel).

I will admit, I picked up a lot of how to use it by osmosis, from over two
and a half years of working at sco, but I still find that it takes very
little to maintain kithrup.  (My latest endeavor, putting kithrup on a
network, has taken me about two days to get working properly.  All without
*any* documentation, only talking to some bsd-knowledgeable folks.)


>the libs keep saying
>"must be linked with the -lx option" (which links the xenix
>libraries).  

The libraries do that?  Wow.  Not on kithrup.  On kithrup, some of the
manual pages say that needs to be done with some things, but I largely
ignore them.  Note that this is done for old makefiles and the like:  xenix
(of which there are a *lot* of systems) had a seperate library (-lx:  the
xenix library) which had some extensions.  AT&T picked them up, finally, for
3.2, and there are good reasons for using libx instead of putting them in
libc.  (Personally, I'd like to see libc about 50% smaller, and would be
willing to live with a few dozen more libraries to get rid of all that
namespace pollution.)

>comp.unix.wizards has been talking a lot about bloated
>code: SCO has *two* shared memory systems, and two semaphore systems
>!!! 

Uhm, so?  If I remember correctly, internally, they're very much the same
code.  The interface is different, but they do the same thing.

Blame AT&T on that.  Xenix had shared memory and semaphores, and then AT&T
went with a different interface, and people like you complained about it, so
xenix went and had both.  Note that *all* versions of 3.2 for the '386 have
both, so why are you blaming sco?

>The security stuff is a joke - who ever heard of an "expire"
>option for users that can't be undone ? 

It can.  Try looking at SLS's; the information was posted here.

But, as always, I keep forgetting:  it's easier to run off at the mouth like
a damned fool than to actually see if your vendor is listening to your
complaints.

>The boot sequence sucks, although
>at least they let you choose single user as it comes up, which ISC
>does not. I have to sit in front of the system as it boots, which
>might be normal for many systems, but if you're writing a driver and
>reboot 30 times a day, its awful ... 

I find nothing wrong with the boot sequence.  Since I've done quite a bit of
kernel development, I've had to reboot often.  And it's never bothered me.
(Or don't you realize you can have a serial console, or can reboot from any
terminal?  No, of course not, that would require actually seeing if such
things work, and it's much easier to be an idiot.)

>The csh is broken, 

The csh is *old*.  And, yeah, it's broken.  I've ported tcsh, as have
others.  No real problem.  Tell people at SCO you want it, and, if they get
enough requests, they will make it available.  But, again, that's probably
more work than flaming.

>the default
>keyboard mapping is brain-damaged (by default they eliminate Alt-<key>
>combinations, and Ctrl-<arrow key> is not the same as <arrow-key> -
>try using emacs with that :-() and only root can reset the key mapping
>for a multiscreen (surely the default file should apply to ALL
>multiscreens, not just the console). 

1.  I agree about the alt, but that's because I use emacs.  I run 'mapkey
/usr/lib/keyboard/cv' at init stage 2.
2.  I like the fact that Ctrl-<arrow> is different.
3.  I know why only root can remap keys; there is a system-wide table (I
believe it works for sunriver stations and the like as well), and you
wouldn't want joe blow to change someone else's keyboard, would you?  Oh, I
forget, you would.
4.  I don't know what you mean by "ALL multiscreens, not just the console."
On kithrup, and almost every system I've seen, the only multiscreesn(tm)
were on the console.  (The exceptions had sunrivers and the ilk.)

>The startup file problem ("xxx:
>cannot execute") is ridiculous, 

Never seen that.

>Looking through the "value-added
>extensions", almost all fall into the "code bloat" category. 

So I guess you don't like job control, bug fixes?  Most of the extensions
(with the exception of the c2 stuff, which I dislike rather much, although I
don't notice it at *all* since I applied the c2 SLS) are for POSIX or some
other standard.  (I've grown to *hate* standards, because I've had to work
on implementing them.)

>Someone will decide to
>change something more fundamental, and suddenly, there will be even
>more significant differences than there are now.

Too late.  The POSIX stuff SCO added is incompatible with the stuff ISC
added.  Point in SCO's favor:  AT&T and Intel have adopted sco's
implementation for the iBCS thingy; as a result, ISC is not compatable with
the standard, and any isc-written COFF program will not work on future
releases.  Oh, well.  (*I* kept urging people to get in touch with ISC and
get some coherency.  Mostly I mumbled to myself, I will admit.)

There is a very violent message below, so you may not want to read it.

As I am no longer an SCO employee, I can finally freely say this:  you are
an asshole.  I (and my coworkers) worked *hard* on the product, and we
didn't always get to implement the decisions we wanted.  But we tried, and
we listened (even if we didn't always reply), and 3.2v2 is a *very* nice
(stable and quick) OS, and 3.2v3 will be even better (although there are
some things I don't like about it).  Most of your complaints are addressed
in SLS's; a polite note to someone (such as myself) who posts regularly is
usually answered politely, and I tried to push any major bug reports
through.  (I circulated Chip's complaints about the C2 stuff around quite a
bit, and I believe it actually influenced things.)  *You* on the other hand,
just inspire contempt, and drive people (at sco) from responding or posting.

How would *you* like to be told that something you've devoted the better
part of two years is a piece of shit?  Over and over and over?  And even
after working on it and making it better, all anyone ever does is complain.
And most of them are *idiots*:  they don't read the manuals, they don't try
thinking, they just *bitch*.

Most of the vocal complainers in this group make me mentally sick.  May you
get the OS you deserve.

-- 
Sean Eric Fagan  | "I made the universe, but please don't blame me for it;
sef at kithrup.COM  |  I had a bellyache at the time."
-----------------+           -- The Turtle (Stephen King, _It_)
Any opinions expressed are my own, and generally unpopular with others.