Patrick Elliott wrote:
> Yeah, so what went wrong? lol Lets see. Code bloat, 

Sure, bloat my OS code with printer drivers. I'll take that over having 
six disks of printer drivers distributed with each application any day.

 > people actually
> wanting to be able to transfer documents between unlike systems (with 
> filenames intact),

Never really had a problem with that...

> oh, and an endless list of cases where people 
> "tried" to make something run on alternatives, but MS changed theirs in 
> some way that either broke the alternate, or broke their code "on" the 
> alternate. 

Well, sure.  Even in the cases where that was deliberate, I'd still say 
"well, duh."

>>> Even with the need to make some adjustments, there is still a basic 
>>> standardization to internals, and commands.
>> As with various versions of Windows. Yes?
>>
> Umm. Not really.. Half the stuff that doesn't work between 3.1 and XP is 
> a result of hacks needed to make it work right at all on 3.1, but which 
> where bugs, hole or unintended interfaces.

Yes. If you take advantage of bugs or undocumented interfaces, or bypass 
the API to try to do the same thing yourself without using the API, then 
when the API gets upgraded, your program breaks.

It's not like nothing got rewritten since System 7 either. This is all 
normal computer stuff that happens on every operating system and 
application and framework everywhere.

> The other half are cases 
> where MS changed the underlying implementations, so you just *can't* do 
> it any more. 

Sure. See above. If you want a multi-user system, you can't take a 
single-user application and expect it to run without giving it undue 
privileges.  I mean, heck, even POV-Ray had to get beaten over the head 
before it stopped writing per-user configuration into per-computer 
directories. How long was Windows doing multi-user systems before Vista 
finally said "OK, enough, really, you have to deal with multi-user 
configuration"?

Things that are actually written to the documented APIs port 
surprisingly well. If you write your own INI file parser, don't expect 
the next upgrade of Windows to put your stuff in the right place in the 
registry for you.

> They are still doing that, releasing .NET, then basically 
> making it very very hard to code anything with MFC. 

In what way are they making it hard?

 > Sure, it will still
> *run*, usually, but its fairly clear that, if they could, they would rip 
> out all those old libraries and bury them, never mind what inconvenience 
> it might cause anyone.

And this differs from the desires of every single programmer inheriting 
a legacy system in what way?

>>> adaptions, most of its isn't going to flat out refuse to work right 
>>> because you plugged a Ford transmission into a Mitsubishi motor, in a 
>>> Chevy frame.
>> I see. That's why Apache and MySql and VI don't work at all under 
>> Windows, yes?  I was wondering why that was.
>>
> Sorry? Are you saying all of those just had some bits of code lopped 
> off, some new code tacked on, then recompiled, because... I get the 
> impression its a *tad* more complicated than that most of the time.

No. I'm saying none of those programs "flat out refused to work" on any 
OS, including Windows and variants of UNIX.

> Mind you, you get some of the same with Linux, depending on if its X, or 
> some other GUI you are running, but I get the sense that the gaps you 
> have to leap are "slightly" less cavernous.

Well, yes, so? And putting a Ford headlight in a Ford car is going to be 
much easier than putting a Ford headlight in a Toyota. (Note that there 
used to be two kinds of headlights: Round, and Square. Now there's a 
whole bunch.)

 > Besides, now you are talking
> about odometers, gas gages, or steering wheels, which is a bit higher 
> level than the "core" systems.

I haven't any clue where your analogy is supposed to be going. You seem 
to be arguing that since cars are built from physical parts, Windows sucks.

> Which version? lol Seriously though, forgive me if I would prefer to 
> avoid MS in my critical life saving devices. ;)

I certainly would too. But then, I'd avoid Linux in those same situations.

IME, hardware is much flakier than software. I have about 40 Linux 
servers I'm controlling right now. Between the colo "testing" their 
generator and frobbing the power on and off, the bakery dumping a 
50-pound bag of flour into the fans, the brand new expensive Dell 
servers just deciding the disk controller doesn't exist any more, I've 
had way more crashes than anything in software causes.

I've also worked at companies where the servers were running Windows, 
and they just never died unless the hardware did.

Indeed, the worst OS I've used is actually Solaris, where we had to 
rearchitect a number of services to account for the bugs in the OS that 
would (for example) not reap processes spawned by cron such that even 
kill -9 didn't stop them, or write file data over top of the inodes just 
because the disk got busy.

So I don't know what kind of professional experience you've had with 
long-running Windows machines, but "drunken sailor on marbles" is far 
from my experience.

> Ok, its not that bad, "anymore". It was, not that far back. 

I had no problems with keeping Win2000 servers running indefinitely. We 
had them locked down in a different state, with the only access being 
remote unless you wanted to hop on a plane for a couple hours.

That's 8 years back, now.

Sure, DOS could get locked up pretty easily. WFW worked surprisingly 
well for the time and the power of the machines it was on.

Try running Unix on a machine with no virtual memory, no memory 
protection, and the kernel in the same address space as the user 
programs, and see how well it runs.

> argue that we have, partly in XP, and hugely in Vista, traded stability 
> for the equivalent of some goon at the door saying, "Now, you know we 
> can't do nothing about the bad guys outside, so whys you want to leave? 
> Just stay here, nice and safe like, and let us decide if that packet 
> shood get sent or not." Security via not letting you do anything. 

Again, a bizarre analogy. How about explaining what you dislike about XP 
by means of a reference to computers rather than to the Godfather movie?

 > Or an
> admittance that they can't stop the stuff that "requires" that kind of 
> security. 

What kind of "stuff" are you talking about?

> Either way, I didn't like 95/98 because it robbed me of a lot 
> of control I *used to have* over my system. 

Like what?

> XP, is kind of getting on my 
> nerves, and what I have seen of Vista... Well, I am not the only one fed 
> up at this point.

Yeah, Vista is certainly a step back in many ways. Many of the problems 
it's trying to solve, tho, are people who are saying "I should be able 
to do anything I want, oh and when it breaks we'll blame Microsoft."

The thing I don't really understand is the number of people who use 
MacOSX, Linux, and Windows, and complain that Windows asks for 
confirmation for privileged operations, but equally croon about how 
great Linux is that you have to sudo a program so you know it's 
something dangerous.

-- 
   Darren New / San Diego, CA, USA (PST)
     "That's pretty. Where's that?"
          "It's the Age of Channelwood."
     "We should go there on vacation some time."

Post a reply to this message