OK!

[DanReed]

Hey all, just visiting. Looking good Tony!

[Undertoad]

Dan! Dan'l! JUST visiting? I'm ashamed and offended! etc. etc.

[alphageek31337]

Hi everyone, my name is steve (insert warm, alcoholics anonymous greeting here), and I'm from pittsburgh. Heard about the cellar from a post on slashdot, and i honestly thought I was getting in at the ground floor of something cool, and, as it turns out, I'm falling into what seems to me to be one of the few real communities still available on the web. I have a quick question or three for mr. mbpark.

1. How much are you selling that amiga of yours for?

2. Do you have a buyer yet?

3. Would you be willing to ship it to Pittsburgh if we came to a mutually beneficial agreement?

I've never had the chance to screw around with an amiga, and I'd like to see what they can do...I figure that this is the perfect opportunity to do just that.

Later

Steve Gazzo

[wst3]

by all means buy that Amiga!!!!!!

The Amiga is an interesting system, the zorro bus looks an awful lot like what PCI turned out to be except that it properly handles interrupt sharing.

And the OS... imagine if you will an OS that provides pre-emptive task switching, and fits in less than 256K of memory. If I had to pick on the OS I would point out that they did not include any kind of memory management, but then they had a pretty good argument that they didn't need it<G>!

I managed to use an Amiga as a digital audio workstation quite effectively, and didn't need memory management until I tried to edit a 90 minute recording. At that point it was a pretty simple matter to add on a third party product.

Another real plus was their layered approach, and the fact that their shared library architecture really worked.

The interface was also very cool, providing a very nice shell environment and a GUI. The GUI was noun oriented, in other words when you clicked on a document the OS figured out what action to apply... may not sound particularly earthshaking right now, but it was back then.

The hardware/Operating system combo was pretty remarkable, although it is showing its age now. The fastest machine I had (still have) was a 68040 running at 30MHz (I'm not sure about the processor speed, but it was 10'x not 100's). The co-processor chip architecture allowed the CPU to do CPU things, while graphics, I/O, and audio chips did their things. I was able to record and playback 8 tracks of audio with 48 channels of MIDI without a glitch. If I tried to play back full motion video I would get ocassional stutters, but I could configure things so that the video stuttered, not the audio.

The development environment was very unix-like, and a great deal of the GNU collection was ported to the Amiga. Then there were folks who had just too much time on their hands. One guy, Dillon, wrote an entire, integrated C development environment, including compiler, and it was no slouch. He also wrote a sliding window network protocol to allow a server running on one amiga to run clients on other Amigas or Sun 3's. And it was usable even at 9600bps.

If you get an Amiga be sure to find a copy of Mind Walker. It was a game that came with the original Amiga 1000, and it is very very cool.

Bill

[alphageek31337]

I haven't been this excited about low-level technology since I ran tomsrtbt on a 286 w/ 4 megs of ram just to mess around with old Infocom games (god, Infocom was great).

So, the only remaining questions are what are the specs on this beast, how much will she set me back, and where can I buy fence for a dinosaur pen?

Looking forward to doing business (!)

Steve

[DanReed]

When?

Really?

[Undertoad]

Not a /. story (that would be outrageous, but it's not really warranted), but the link is in my sig, and I, uh, kinda post an awful lot to /. (It doesn't hurt that I have user #333 and some people find that to be impressive. I don't, but some people do.)

[MaggieL]

Originally posted by wst3
by all means buy that Amiga!!!!!!

The Amiga is an interesting system, the zorro bus looks an awful lot like what PCI turned out to be except that it properly handles interrupt sharing.

Bill

Close friend of mine was an LSI designer at Commodore Engineering. She went to Compaq just before that folded. Now she has her name on some of the key patents for hot-pluggable PCI.

[adamzion]

[B}
Close friend of mine was an LSI designer at Commodore Engineering. She went to Compaq just before that folded. Now she has her name on some of the key patents for hot-pluggable PCI. [/b]

<b>Hot-pluggable PCI?</b> Now, there's an interesting thought... are you supposed be able to open the PC's case and remove the cards while the system is up and running? That would be quite a trick in the first place, although I'm honestly sure why someone would want to try it.

Or does this not refer to the internal PCI slots but something else entirely?

Gee, let's switch the video card without shutting down,
Z

[MaggieL]

[quote]Originally posted by adamzion

[B}

<b>Hot-pluggable PCI?</b> Now, there's an interesting thought... are you supposed be able to open the PC's case and remove the cards while the system is up and running? That would be quite a trick in the first place, although I'm honestly sure why someone would want to try it.

I had that reaction when she first told me about it, too.
But think about servers where you could add and remove disk controllers and network cards in-flight. Especially in RAID arrays. Ever see a big Compaq server? The disk drives have handles on them, and slide into bays in rack-mount units. The keyboard is a rack-mounted drawer. The monitor is on a rack-mounted shelf.<p>
Ghoddess, what a geekgrrl I am...hardware gets me squishy.

[adamzion]

Originally posted by adamzion

<b>Hot-pluggable PCI?</b> Now, there's an interesting thought... are you supposed be able to open the PC's case and remove the cards while the system is up and running? That would be quite a trick in the first place, although I'm honestly sure why someone would want to try it.

Originally posted by MaggieL
I had that reaction when she first told me about it, too.
But think about servers where you could add and remove disk controllers and network cards in-flight. Especially in RAID arrays. Ever see a big Compaq server? The disk drives have handles on them, and slide into bays in rack-mount units. The keyboard is a rack-mounted drawer. The monitor is on a rack-mounted shelf.<p>
Ghoddess, what a geekgrrl I am...hardware gets me squishy.
[/b]

One problem: I already manage three servers with hot-swappable SCSI hard drives making up a RAID. The tech for that has been around for quite a while before now. Unless, that is, I misunderstand entirely what you mean in your quote.

Which is, indeed, possible,
Z

[MaggieL]

Originally posted by adamzion

One problem: I already manage three servers with hot-swappable SCSI hard drives making up a RAID. The tech for that has been around for quite a while before now. Unless, that is, I misunderstand entirely what you mean in your quote.
[/b]

Only to the extent that this applies at the PCI bus level. So instead of just drives, you can replace drive *controllers*. Or NIC cards. Telephony interfaces. Or whatever you have attached to the PCI bus that uses this technology.<p> It's not really *new* tech in terms of internet time either...this was a few years ago.

[tw]

Originally posted by MaggieL
... So instead of just drives, you can replace drive *controllers*. Or NIC cards. Telephony interfaces. Or whatever you have attached to the PCI bus that uses this technology.<p> It's not really *new* tech in terms of internet time either...this was a few years ago. [/b]

Hot swap busses are old technology even for the military. One that I reviewed was a communication system for a new line of nuclear submarines. Any computer could and had to be removed or replaced while the entire communication system remained operational. It used an old (1970s) bus structure called something like IEEE 1324 (number long since forgotten) which is also used in aviation - and before fly by wire was yet common in commercial airplanes.

The most important problem in hot swapping are voltage differences - especially if the board has multiple voltages. It is critical to establish ground before making any other power or bus connection. If a board is powered by +12V, +5V, and -12V; but has no ground, then PN semiconductor junctions, normally used as isolators, can become conductors. IOW current flow in ICs where the IC is not designed to handle current flow - damage.

HP computers (pre-PCs), although not hot swappable, assumed that humans would make mistakes. Therefore ground was located on both ends of a board and ground pins were longer - to guarantee that ground would always be 'first to make' and 'last to break'.

USB is also a hot swappable bus. Current to a newly installed board must be current limited so that the board does not overdraw from the bus. Then software must recognize the new hardware, initialize the hardware, and inform all other systems (usually just the OS) that a new device is in existance. Another whole and different scenario is required for hot-swap removal. It requires coordination between software and hardware - something not always easily accomplished by the software side of engineering.

As a result of hot swapping and other features, the code for USB is about as complicated as the entire MS-DOS package. Linux had trouble implementing complex USB. The complexities of a hot swappable USB is why Transmeta is so important to Linux's future.

Hot swapping is not easily implemented since humans can vary the hot swapping process so easily. Code and hardware must be designed to withstand every 'innovative' hot swapping human.

Feature articles on hot swapping busses were in electronics magazines some years ago as newer, hot swap interface chips were developed.

[mbpark]

eBay does wonders for your life when it sells in 40 minutes!

However, I've gotten a couple of TRS-80 machines (Model 100 and Wp-2) to play with for data entry devices. The WP-2 has the same keyboard as my 1200.

My interests lie more with portable computers and remote data entry these days. One of my co-workers wants to try and port .NET to a Tandy Model 100, and I think he could at least get a cross-compiler from the CLR going.

However, It's been a busy week this week, and internet connectivity sucks ass.

Mitch

[lumberjim]

Holy shit!

is this the very first thread???!!!

more importantly, do you old timers think I've defiled it by posting to it? don't answer that, dagney.