Re: I'm giving up computers if this is the future.
From: Mike Bartman (omni_at_foolie.omniphile.com)
Date: 09/12/04
- Next message: Larry Kilgallen: "Re: Impersonate"
- Previous message: Antonio Carlini: "Re: TSM hates my DECnet version (Was: TSM won't accept my circuit name)"
- In reply to: Undisclosed: "Re: I'm giving up computers if this is the future."
- Next in thread: Z: "Re: I'm giving up computers if this is the future."
- Reply: Z: "Re: I'm giving up computers if this is the future."
- Reply: Undisclosed: "Re: I'm giving up computers if this is the future."
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Sat, 11 Sep 2004 19:36:36 -0400
On Sat, 11 Sep 2004 03:28:39 -0400, Undisclosed
<nomail@dontbeaweaselspammer.com> wrote:
>Mike Bartman wrote:
>> There's a lot wrong with the PC (assuming you mean the IBM clones, not
>> the generic term).
>>
>> How about that whole IRQ thing? That bus that Apple once used
>
>I haven't had a problem with "plug and pray" since I gave up ISA and EISA.
I have. PCI is better, with the newer MBs, but it isn't perfect.
The first things I listed were mostly problems with the IBM PC. Some
of them have been improved considerably in the last 20 years, some
haven't.
>everyone uses PCI and the derivatives now. Even Apple(!) and Sun.
That's not because it's better. It's just more available.
>> How about the whole "can't address a drive over xxx megabytes" thing?
>> (that we went through more than once as drives *continued* to get
>> bigger...like nobody thought that would happen after the first two
>> times and came up with a scheme that wasn't closed-ended???) And all
>> the various ways different companies came up with to get around it?
>
>is there a generic solution to this?
Yeah, if you design things properly. The culprit is fixed length
fields in data structures. That may have been necessary 30 years ago
due to limited memory and CPU power, but there's not much excuse for
it today. If you need still more speed than the CPU can manage (find
this hard to believe...), use variable length fields to describe the
drive, and self-modifying code when you mount the drives...write the
fixed length version as you need it; once you know what sizes you are
dealing with on the particular drive in question. Don't hard code
them into the EPROMS and standards, and then run into a wall when
drives exceed the limits set by this, and then play all sorts of games
trying to get around them...just fix the problem up front!
>> How about that crappy Intel 80x86 architecture, that has no general
>> purpose registers at all (there are instructions that can *only* be
>> used with one particular register, and which register that is varies
>> with the instruction in question...every register gets used this way
>> at one time or another), resulting in *lots* of instructions that do
>> nothing but shuffle data from the register it's in to the one it needs
>> to be in for the next instruction to work.
>
>AMD-64 adds a comfortable number of registers. It's not the SPARC with
>32 GP registers, but it's not bad.
While it may be compatible with the Intel 80x86 family, it isn't the
Intel 80x86 family, so it doesn't change my complaints with the Intel
80x86 family.
>At least the non-real-mode
>> addressing has gotten rid of that rediculous segmented memory model
>> and the pain and suffering it exacted...you couldn't even compare two
>> pointers without "normalizing" them first (stuffing as much of the
>> address into the segment register as possible).
>
>dude, segments have been dead for everything but control by the OS
>kernel itself since the 386.
No, not really. The 80486 and MS-DOS (and early Windows versions)
used real mode too, and had the segment problems, and I suspect that
the current Pentium 4 still has them in there too...though I haven't
tried to program an Intel chip at that level since the Pentium
days...I let the compilers worry about it these days and don't write
any boot code. As I said, going virtual and using the "flat model"
eliminated them, but they are still there in the CPUs if you aren't
using that mode of operation.
>> And the flash upgrades of motherboards that, if anything goes wrong,
>> leaves you with a dead bit of chip-jewelry? How about a ROM with the
>> flash program in it and a reset so you can try again to load a new
>> BIOS??? That one's not IBM's fault...they didn't have a way to
>> upgrade the BIOS without plugging in a new chip...
>
>if you are buying a server or high-end PC, you can get one with two
>flash memories full of BIOS. One dies - you switch to the other.
That there are options on some systems, if you spend enough on them,
doesn't change the fact that PCs have problems...
>> For an overgrown cash register (which is what it was...), the PC
>> wasn't bad. As a good computer design it lacks big time. There were
>> better machines out at the time...they just didn't have "IBM"
>> stencilled on them. The Victor 9000 for instance. 800x600 monochrome
>> non-glare screen with dot-addressable graphics, VT-52 escape codes in
>> the BIOS for text output with special effects, 1.2 meg variable speed
>> double sided quad-density 5.25" floppies (the IBM machine had 160K
>> single sided, single density drives), and the screen brightness and
>> contrast were controllable with software or from the keyboard.
>> Standard memory was 256K (IBM was 64K). About the only place the IBM
>> beat it was in the expansion bus...the Victor had a connector on the
>> back that brought all the CPU lines and some MB lines out for an
>> expansion box, but no place for individual cards. The Sage was a nice
>> machine too...ran Unix on an M68000 and came stock with an MS-DOS
>> emulator that ran as fast as an original PC ran it. it was also
>> smaller than a PC (bit under half the size).
>
>ahh, a hardware geek. ;)
No, I'm a programmer...but software has to run on hardware to get any
real speed out of it...and as a computer owner and user, I do tend to
pay attention to hardware and its capabilities.
>never heard of those machines, actually.
You're under 30, aren't you? You're one of those people who came
along after the Wintel boxes took over the world, so you've never seen
the other ways it could have been if Joe Sixpack and the BOTASMs
hadn't perverted the computer revolution. You don't know any
better...
[the above is mostly a mini rant...may or may not really apply to you,
but it's true enough for too many...]
>as for the expansion bus... ISA is one of the major reasons the PC
>florished.
That it had an expansion bus, yes, I agree. That it had such a sucky
one is another issue. IBM tried to fix it with the Microchannel bus
in the PS-2, but the world was already too far down the rackety ISA
track to switch by that point. Too bad they didn't think things
through better before they foisted the PC on the world in a mad rush
of indifference to the whole concept of personal computers...
>Sage sounds quite a bit like the original Sun machines.
Somewhat. Was out about the same time.
>I agree with your general points that the PC, when it started out, was a
>wretched architecture.
>
>but it's had 20 years to rub away the sharp edges.
Some. It has improved, but some of the rough edges are still there,
they are just covered up a bit here and there. They still show
through though. Things like where you can install NT on a disk for it
to boot properly (first 2 gigs only I believe...due to a limitation in
the BIOS design I hear), the whole drive letter scheme that limits
total drive count (you can get around it sort of with RAID), etc..
-- Mike B.
-- Mike B.
'04 FLSTCI
----------------------------------------------------------------
To reply via e-mail, remove the 'foolie.' from the address.
I'm getting sick of all the SPAM...
----------------------------------------------------------------
- Next message: Larry Kilgallen: "Re: Impersonate"
- Previous message: Antonio Carlini: "Re: TSM hates my DECnet version (Was: TSM won't accept my circuit name)"
- In reply to: Undisclosed: "Re: I'm giving up computers if this is the future."
- Next in thread: Z: "Re: I'm giving up computers if this is the future."
- Reply: Z: "Re: I'm giving up computers if this is the future."
- Reply: Undisclosed: "Re: I'm giving up computers if this is the future."
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
