The War for General Purpose Computing

originally posted by: dug88
a reply to: Gothmog

How about some references for any of your claims. I posted plenty for mine. You have zero.

You are in minus territory .
You take a bunch of facts and imagination and come to a conclusion .

a reply to: MykeNukem

The more I'm hearing the "experts" in this thread talk, the more I'm realizing they're not experts.

That is totally your responsibility .
My responsibility , on ATS , is to deny ignorance.

originally posted by: MykeNukem

originally posted by: Havamal
Sorry. My expertise is with larger and more complex computer systems. I always considered windows as a "toy" system for home users. Still is.

windows, at is base, is very primitive. Compare to IBM systems in th 1960s and they are just getting to a virtual machine. iBM 360 1964.

Ahem. But let the kids have fun with their toys.

You realize anyone can just run Hercules and emulate an IBM 360 OS which is Public Domain now at thosands of times the speed , right?

Can the IBM 360 emulate Windows? No, of course not.

The more I'm hearing the "experts" in this thread talk, the more I'm realizing they're not experts.

I don't suppose you have heard of z/OS, or the current level of prossers that run over 50,000 MIPs per box? Do you even know what what an MSU is?

It is 2021. You are boasting that toy computers can now run a 1960's ios?

Did you know that the z/series now runs about 90 percent of wall street, banking, large government (SSA, IRS, DOD) and represents most transactions in the business community world wide?

Toy computers. Last I checked it would take 60,000 commodity servers running window to meet the demand of one z/os mainframe .

See the problem? One box the size of a refrigerator vs. A building housing a rabbit farm, where stuff is always failing.

Mean time to failure on a z/os 15 - 32 years.

Big boys get big toys. Little boys who went to the local one year tech school only know about toys - Microsoft and maybe Linus. Good toys for small systems.

Want to play heavy metal?

IBM.

originally posted by: MykeNukem

originally posted by: Havamal
Sorry. My expertise is with larger and more complex computer systems. I always considered windows as a "toy" system for home users. Still is.

windows, at is base, is very primitive. Compare to IBM systems in th 1960s and they are just getting to a virtual machine. iBM 360 1964.

Ahem. But let the kids have fun with their toys.

You realize anyone can just run Hercules and emulate an IBM 360 OS which is Public Domain now at thosands of times the speed , right?

Can the IBM 360 emulate Windows? No, of course not.

The more I'm hearing the "experts" in this thread talk, the more I'm realizing they're not experts.

As a side note - why would anyone want to emulate windows?

When was the last time a mainframe crashed? 1959?

Get serious. Mainframe don't crash. They can have parts fail while running, they tell you, and you hot swap parts and never stop processing.

The last time I saw this happen was 2004, and it was a channel card. The engineer moved the plug to a spar slit. We list about 10 seconds of processing time. About 10 billion service unit cycles, but the machine made it back up in about 12 second. Last outage I recall.

Toy computers.

Sniff..

originally posted by: MykeNukem
a reply to: Gothmog

Here's a situation off the top of my head:

Could a packet sniffer be setup to retrieve TPM stack API packets and therefore the "key", which would enable you to spoof a piece of hardware and deliver a payload, taking over the complete machine?

Just a thought...

OP: thoughts on this?

Both Spectre and Meltdown may have been able to .
Yet , they are both protected from by hardware and software mow . (and were acknowledged in the 80s by the way)

a reply to: Havamal

Just because they haven't updated yet doesn't mean mainframe is a thing.

Almost all companies use systems until they are hopelessly outdated. Remember Y2K?, nothing happened but we still updated everything.

Not sure who would want to emulate Windows, the point is, your systems can't even do that.

When the reverse isn't true.

BTW- For the record, I don't use Windows.

Are you seriously trying to say that mainframe computing is superior to lets say Distributed Processing? Come on now.

There's a reason we moved on, multitasking being one of them. RAID arrays being another.

originally posted by: Gothmog

originally posted by: MykeNukem
a reply to: Gothmog

Here's a situation off the top of my head:

Could a packet sniffer be setup to retrieve TPM stack API packets and therefore the "key", which would enable you to spoof a piece of hardware and deliver a payload, taking over the complete machine?

Just a thought...

OP: thoughts on this?

Both Spectre and Meltdown may have been able to .
Yet , they are both protected from by hardware and software mow . (and were acknowledged in the 80s by the way)

They were acknowledged in the 80s pertaining to what?

TPM?

ok then.

Like I said, you guys liike to flex, but when the rubber meets the road it's all ad homs.

lmao.

originally posted by: Gothmog

originally posted by: dug88
a reply to: Gothmog

How about some references for any of your claims. I posted plenty for mine. You have zero.

You are in minus territory .
You take a bunch of facts and imagination and come to a conclusion .

You've already proven yourself wrong to the OP after stating you didn't have to read anything because you know everything there is to know.

Then, you were proven wrong.

Enough said.

First you said Network Boot wasn't enabled by default - Then, you admitted it is always enabled by default.

Then, you said TPM isn't cryptographically encrypted - Then, you admitted you hadn't read up on that. (after berating the OP for suggesting such a thing)

Tried giving you the benefit of the doubt, but I was wrong too.

One thing any good tech knows, is that he doesn't know everything.

Deny Ignorance.

originally posted by: MykeNukem

originally posted by: Gothmog

originally posted by: MykeNukem
a reply to: Gothmog

Here's a situation off the top of my head:

Could a packet sniffer be setup to retrieve TPM stack API packets and therefore the "key", which would enable you to spoof a piece of hardware and deliver a payload, taking over the complete machine?

Just a thought...

OP: thoughts on this?

Both Spectre and Meltdown may have been able to .
Yet , they are both protected from by hardware and software mow . (and were acknowledged in the 80s by the way)

They were acknowledged in the 80s pertaining to what?

TPM?

ok then.

Like I said, you guys liike to flex, but when the rubber meets the road it's all ad homs.

lmao.

As being exploitable (memory side addressing)
TPM does not have anything to do with it .
Nor does it have anything to do with drivers.

I don't know about flexing .
I am trying to educate , to Deny Ignorance .

originally posted by: Gothmog

originally posted by: MykeNukem

originally posted by: Gothmog

originally posted by: MykeNukem
a reply to: Gothmog

Here's a situation off the top of my head:

Could a packet sniffer be setup to retrieve TPM stack API packets and therefore the "key", which would enable you to spoof a piece of hardware and deliver a payload, taking over the complete machine?

Just a thought...

OP: thoughts on this?

Both Spectre and Meltdown may have been able to .
Yet , they are both protected from by hardware and software mow . (and were acknowledged in the 80s by the way)

They were acknowledged in the 80s pertaining to what?

TPM?

ok then.

Like I said, you guys liike to flex, but when the rubber meets the road it's all ad homs.

lmao.

As being exploitable (memory side addressing)
TPM does not have anything to do with it .
Nor does it have anything to do with drivers.

I don't know about flexing .
I am trying to educate , to Deny Ignorance .

I asked about TPM stack API.

That's a thing.

I can assure you, it has to do with TPM, as the name suggests.

Never mentioned anything at all to do with MSA or drivers in that post.

I thought you said Deny Ignorance?

a reply to: dug88

All those old programming languages, and many old programs still work. In fact, the modern frameworks and languages still execute the same old code underneath, for the most part. It's all still there, and all still available to those who want to write programs with it.

The thing is, way back then they also used far fewer resources and had simpler interfaces. But, at some stage 640 kilobytes of RAM, with a text only interface no longer cut it. So they have been incorporating new ideas and technologies as they do, and now we have graphical functions and operating systems that you can work in by manipulating little pictures on screen.

Most of the time, the stuff is monetized. They want to get paid. It's that simple.

But even the most complex of current computational tasks can still be described in terms of a Turing machine. And some would say that things such as graphical interfaces and modular code have vastly simplified complex tasks for the majority of users.

People are writing highly graphical games with no traditional computer programming background. So, I would argue that the 'universality' of computer platforms is converging, and well and truly alive.

originally posted by: chr0naut
a reply to: dug88

All those old programming languages, and many old programs still work. In fact, the modern frameworks and languages still execute the same old code underneath, for the most part. It's all still there, and all still available to those who want to write programs with it.

The thing is, way back then they also used far fewer resources and had simpler interfaces. But, at some stage 640 kilobytes of RAM, with a text only interface no longer cut it. So they have been incorporating new ideas and technologies as they do, and now we have graphical functions and operating systems that you can work in by manipulating little pictures on screen.

Most of the time, the stuff is monetized. They want to get paid. It's that simple.

But even the most complex of current computational tasks can still be described in terms of a Turing machine. And some would say that things such as graphical interfaces and modular code have vastly simplified complex tasks for the majority of users.

People are writing highly graphical games with no traditional computer programming background. So, I would argue that the 'universality' of computer platforms is converging, and well and truly alive.

Heh, we agree for once.

Turing and Boolean Logic rules the day.

BloatWare.

With handy API's for interconnectivity and Tx on the IOT.

It's a good thing and a bad thing. Dangit.

originally posted by: MykeNukem

originally posted by: chr0naut
a reply to: dug88

All those old programming languages, and many old programs still work. In fact, the modern frameworks and languages still execute the same old code underneath, for the most part. It's all still there, and all still available to those who want to write programs with it.

The thing is, way back then they also used far fewer resources and had simpler interfaces. But, at some stage 640 kilobytes of RAM, with a text only interface no longer cut it. So they have been incorporating new ideas and technologies as they do, and now we have graphical functions and operating systems that you can work in by manipulating little pictures on screen.

Most of the time, the stuff is monetized. They want to get paid. It's that simple.

But even the most complex of current computational tasks can still be described in terms of a Turing machine. And some would say that things such as graphical interfaces and modular code have vastly simplified complex tasks for the majority of users.

People are writing highly graphical games with no traditional computer programming background. So, I would argue that the 'universality' of computer platforms is converging, and well and truly alive.

Heh, we agree for once.

Turing and Boolean Logic rules the day.

BloatWare.

With handy API's for interconnectivity and Tx on the IOT.

It's a good thing and a bad thing. Dangit.

Bigiron is useful if you have teams of sysprogs and devops. But there is something to be said about those consumer 'toys'. Just look at all the newer tiny compute devices like tablets, phones and etc. Look at the richness of function in the dumbest of devices. There is a lot there.

Sometimes, it is the 'right tool for the job' and some disconnected and archaic mainframe (and the vast majority still are a bit archaic) is not the answer to getting a weather app on your smartphone.

With the birth of quantum computing, I imagine everyone will turn back to big-iron for a while, but hybrid and handheld are where the users will be, increasingly.

It is a commercial inevitability.

originally posted by: chr0naut

originally posted by: MykeNukem

originally posted by: chr0naut
a reply to: dug88

All those old programming languages, and many old programs still work. In fact, the modern frameworks and languages still execute the same old code underneath, for the most part. It's all still there, and all still available to those who want to write programs with it.

The thing is, way back then they also used far fewer resources and had simpler interfaces. But, at some stage 640 kilobytes of RAM, with a text only interface no longer cut it. So they have been incorporating new ideas and technologies as they do, and now we have graphical functions and operating systems that you can work in by manipulating little pictures on screen.

Most of the time, the stuff is monetized. They want to get paid. It's that simple.

But even the most complex of current computational tasks can still be described in terms of a Turing machine. And some would say that things such as graphical interfaces and modular code have vastly simplified complex tasks for the majority of users.

People are writing highly graphical games with no traditional computer programming background. So, I would argue that the 'universality' of computer platforms is converging, and well and truly alive.

Heh, we agree for once.

Turing and Boolean Logic rules the day.

BloatWare.

With handy API's for interconnectivity and Tx on the IOT.

It's a good thing and a bad thing. Dangit.

Bigiron is useful if you have teams of sysprogs and devops. But there is something to be said about those consumer 'toys'. Just look at all the newer tiny compute devices like tablets, phones and etc. Look at the richness of function in the dumbest of devices. There is a lot there.

Sometimes, it is the 'right tool for the job' and some disconnected and archaic mainframe (and the vast majority still are a bit archaic) is not the answer to getting a weather app on your smartphone.

With the birth of quantum computing, I imagine everyone will turn back to big-iron for a while, but hybrid and handheld are where the users will be, increasingly.

It is a commercial inevitability.

Agreed.

Millions or Billions of handhelds all interconnected to relay signals, even if they are "off network" is the future.

There will be no "off network" soon.

Also, yes, older systems, if they do their jobs, why update? Unless necessary or forced for compatibility reasons? It doesn't make good business sense to update for no reason. That's why we still have mainframes. For now.

a reply to: Gothmog

Maybe you should tell that to the people who actually wrote all those articles I referenced.

I've seen bits and bobs of this process over the years. The Apple II was great for open architecture; the Mac came along and closed it down.

Assemblers used to be available for 8086 (and descendants); I guess they are still out there, but the Windows security model has practically made them a thing of the past on that OS.

The locking down is more visible now and chafes in some cases.

Cheers

a reply to: Havamal

First, I am not very prolific at this topic BUT I am a person that asks a lot.

And I dare now to try to correct you on some things. My uncle had a mainframe and a normal stacked server tower with drawers in his barn and he explained me, after I pointed out the huge differences between the two, he explained to me:


The mainframe is built for redundancy and throughput. The server rack for many millions of different tasks. You can unplug single components from a mainframe and it will chooch on without loosing much, or no calculations while the server rack, the drawers can be pulled one by one too, by pulling it it's like you take a complete fraction of the whole system offline.

That's why your mainframe chooches for 30 years or whatever the timeframe is, because they are built sturdy and for redundancy. It's like comparing a truck to a fleet of race cars.

a reply to: chr0naut
Big Iron systems are the way to go for big business and governments.
The CPUs are RISC (or Cell) and there are DRAWERS full of em .
Most of the Top 500 are "Big Iron" .
There are a few Super Clusters on the list , but not many .

Big Iron systems won't play games though .
Wait , one could say they can play chess and Jeopardy rather well .

originally posted by: dug88
a reply to: Gothmog


Maybe you should tell that to the people who actually wrote all those articles I referenced.

The articles are (for the most part) not the problem .
It is when folks start extrapolating stuff on their own .

a reply to: Gothmog

I posted some direct quotes from Linus Torvalds supporting my claims, actual criticisms posted directly in Wikipedia, a long rant by Richard Stallman, several articles from Arstechnica with direct quotes, actual straight information about these technologies, where interpretation is not required, where anyone with common sense can see the problem.

Again, if you're going to keep #ting up this thread with baseless claims, please post references for all the claims you've made. You've posted things that are straight up not true, without any references, responded dismissively and just non-productively, please, post something reasonable worthy of discussion. Post something a grownup would post instead of unsubstantiated bull#.