When software vendors become liable for data loss, and the associated costs, then they have a very strong financial incentive to fix bugs.
In the current model, even with full disclosure, the most they risk is sales loss due to bad PR, and to modernize the old saw, "nobody ever got fired for buying Microsoft".
Having had the pleasure of programming the Z8K, it was not a 16 bit version of the Z80.
The Z8000 was available in four flavors:
Z8001 Segmented (8MB address space)
Z8002 Non-Segmented (64KB address space)
Z8003 Segmented (8MB address space), Virtual Memory Support
Z8004 Non-Segmented (8MB address space) Virtual Memory Support
The Segmented CPUs had a flag bit that allowed them to run in non-segmented mode.
The Z8000 was much closer architecturally to the 68K family than the Z80/x86 family. It had 16 orthagonal, 16-bit registers (R0-R15), which could be paired up as 8 32-bit registers (RR0-RR14). R15 (non-segmented mode) or RR14 (segmented mode) was the stack pointer.
The opcode names were similar to the Z80, but the architecture was vastly different. The Z8000 series was popular in embedded and military applications. Unfortunately, I don't believe Zilog ever built the Z8070 FPU for the processor, which also hindered it's acceptance as a mainstream CPU.
Anyone out there remember the Zilog ZEUS System 8000? It was a Unix System III variant.
Yeah... and when they make the closing credits unreadable so the talking heads can say, "We'll scare the sh*t out of you - film at 11!". What if I *WANTED* to see the closing credits... for example, I thought the actor who played Fred the Butler looked really familiar, but I couldn't place him...
If you ever watched the late lamented UPN series "Nowhere Man", you know that the logos are there for mind control! They're to turn us into good little citizens^H^H^H^H^H^H^H^Hconsumers.
Now if you'll excuse me, I need to go put on my tinfoil hat:-)
BZZZT! And thank you for playing! Here's your lovely parting gift.
MS did not get in trouble for being a monopoly -- that's perfectly legal. They got in trouble for abuse of said monopoly... using their monopoly in the OS space to obtain a second monopoly.
including proof of concept coding (which we got management approval to throw away).
This is critical. When you are doing something completely new, and you need PoC, then the odds are your first cut (even with semi-formal or formal design methods) will be bad/buggy/suboptimal. Being able to trash it, without fear of repercussion is very important.
I've seen many cases where we were told "this is demo, don't worry", and then the demo became the shipping product.
I've had the luxury of discarding built code exactly once. This actually cleaned up every single bug because I was able to work from a (re)design that came from what I learned in PoC phase.
Mod this guy up. In my 18 years of experience, I've *NEVER* seen enough time allocated to test, and this is in DoD-land, where they actually realize the need for test.
If test isn't allocated at least 30% of the man-hours (50% is better), then you're going to have delivery problems.
Whenever I had to make an estimate, I'd figure out the best value (we had COCOMO estimation software), and then I'd pad it...
Not because I thought the estimate was off, but because Marketing would *ALWAYS* shave the estimates, and that's what would be in the contract. Well, that plus Hofstaedter's Law.
In addition to man-hours, I'd often try to tack on N calendar-months for learning curve if we were doing something new (new platform, problem domain, etc...).
Wasn't always successful, but it worked pretty well... usually we worked out to the original man-hour estimate plus the learning-curve time. Of course, Marketing always low-balled it, and the learning-curve time always went away.
Slashdot Mirror
ZEUS (Zilog Extended Unix System) came out before System III was officially released. Pretty sure it was based on V7, not System III.
I'm pretty sure that uname indicated SysIII. Of course, that was ZEUS 3.1 and 3.2. YMMV for earlier versions.
Was there really any difference between V7 and SysIII?
When software vendors become liable for data loss, and the associated costs, then they have a very strong financial incentive to fix bugs.
In the current model, even with full disclosure, the most they risk is sales loss due to bad PR, and to modernize the old saw, "nobody ever got fired for buying Microsoft".
I *LIKED* the Z80's alternate register set. It made interrupt handling much easier!
Nope, the 8008 was. The 8080 was the child of the 8008.
The Z8000 was available in four flavors:
Z8001 Segmented (8MB address space)
Z8002 Non-Segmented (64KB address space)
Z8003 Segmented (8MB address space), Virtual Memory Support
Z8004 Non-Segmented (8MB address space) Virtual Memory Support
The Segmented CPUs had a flag bit that allowed them to run in non-segmented mode.
The Z8000 was much closer architecturally to the 68K family than the Z80/x86 family. It had 16 orthagonal, 16-bit registers (R0-R15), which could be paired up as 8 32-bit registers (RR0-RR14). R15 (non-segmented mode) or RR14 (segmented mode) was the stack pointer.
The opcode names were similar to the Z80, but the architecture was vastly different. The Z8000 series was popular in embedded and military applications. Unfortunately, I don't believe Zilog ever built the Z8070 FPU for the processor, which also hindered it's acceptance as a mainstream CPU.
Anyone out there remember the Zilog ZEUS System 8000? It was a Unix System III variant.
Here's a good "plain english" explanation of what happened.
Karma whore disclaimer: This link was stolen from the other article about this experiment.
the vast majority of people prefer the "sexy" new name.
Remember when the Apple marketroids tried to get everyone to pronounce SCSI as "sexy" instead of "scuzzy"?
As a proud Californian, I'm willing to admit that the MA Attorney General (who's name escapes me) deserves a heck of a lot of credit too.
How about killing that 20 second timer? That's pretty lame...
And under certain browsers (*cringe*IE5.5*cringe), when you hit the back button, you've lost what you typed if you blow the 20 second filter.
I also think that adding a new story type "Poll" along the lines of Ask Slashdot would be nice (for user suggested polls).
Nah. It should be the Trojans Navy...
"Protecting you from enemy seamen!"
According to CBS News, the "Big Three" are CA, MA, and MN.
GRC(R) Windows
Would that be the version that makes strange claims in really big oddly colored text?
What I like about that letter is dragging Jim Allchin's "Open Source Is UnAmerican" FUD back to bite MS in the a**.
Bill Lockyer: www.ag.ca.gov
He has a page for submissions on the MS case, but the deadline was noon yesterday (I just barely go t mine in on time!)
California, I'm proud to say, is reported (by my local news -- KNX 1070 -- as being just as strongly against the settlement as is MA.
Good for you Bill Lockyer!
A Toshiba Libretto also would work nicely.
Let's type to kill the 20 second timer now that this is the third time I've tried to post this.
Yeah... and when they make the closing credits unreadable so the talking heads can say, "We'll scare the sh*t out of you - film at 11!". What if I *WANTED* to see the closing credits... for example, I thought the actor who played Fred the Butler looked really familiar, but I couldn't place him...
If you ever watched the late lamented UPN series "Nowhere Man", you know that the logos are there for mind control! They're to turn us into good little citizens^H^H^H^H^H^H^H^Hconsumers.
:-)
Now if you'll excuse me, I need to go put on my tinfoil hat
They got in trouble for being a monopoly.
BZZZT! And thank you for playing! Here's your lovely parting gift.
MS did not get in trouble for being a monopoly -- that's perfectly legal. They got in trouble for abuse of said monopoly... using their monopoly in the OS space to obtain a second monopoly.
Where did you get the $4000? That's *SINGLE-SEAT*.
A full-bore installation of Rose and associated toolsets for Ada can easily run to $125,000 + 10%annually for maintenance!
including proof of concept coding (which we got management approval to throw away).
This is critical. When you are doing something completely new, and you need PoC, then the odds are your first cut (even with semi-formal or formal design methods) will be bad/buggy/suboptimal. Being able to trash it, without fear of repercussion is very important.
I've seen many cases where we were told "this is demo, don't worry", and then the demo became the shipping product.
I've had the luxury of discarding built code exactly once. This actually cleaned up every single bug because I was able to work from a (re)design that came from what I learned in PoC phase.
Mod this guy up. In my 18 years of experience, I've *NEVER* seen enough time allocated to test, and this is in DoD-land, where they actually realize the need for test.
If test isn't allocated at least 30% of the man-hours (50% is better), then you're going to have delivery problems.
Whenever I had to make an estimate, I'd figure out the best value (we had COCOMO estimation software), and then I'd pad it...
Not because I thought the estimate was off, but because Marketing would *ALWAYS* shave the estimates, and that's what would be in the contract. Well, that plus Hofstaedter's Law.
In addition to man-hours, I'd often try to tack on N calendar-months for learning curve if we were doing something new (new platform, problem domain, etc...).
Wasn't always successful, but it worked pretty well... usually we worked out to the original man-hour estimate plus the learning-curve time. Of course, Marketing always low-balled it, and the learning-curve time always went away.
I never (knowingly) allow any site to keep my CCnumber and why I always use a "temporary" CC number (for example Amex Private Payments).
who can't even tell you why qsort beats a bubble sort?
Only in the average case. For certain pathological cases, qsort is O(N^2).
Heapsort has a larger coefficient, but is O(N log N) for all cases.