Amiga Multitask

[gertsy]

The first Amiga (to be known as the Amiga 1000) was special. Revolutionary.

Memory was the ceiling for multitasking for all PCs until relatively recently (Last 12 years) Because it was so expensive and usually programs were written to use all that a standard configuration supplied.
Windows 95 had 32 Bit preemptive multitasking with 32 Bit application protection. But you needed a lot of memory to run multiple programs. Luckily you couldn't install pizz poor marketing bullcrap onboard most MBs.

[psxphill]

Alternatively, you could just have efficient software that maximizes free memory for your actual work purposes and enough RAM to fit the task at hand, rather than churning data to and fro over a disk interface many orders of magnitude slower than the RAM which is itself likely not actually fast enough to keep up with the demands of the processor.
 
But, you know, that'd just be crazy.

Some software needs alot of ram because of the functionality it offers. The only compromise you can make here is to remove functionality, but then there would be no innovation.
 
Some uses more than it should because writing perfectly efficient software is alot more expensive & ram is cheap. The only compromise you can make here is delay the software and charge more for it, the odds are the developers would run out of money.
 
It's not limited to Windows, the Amiga & Mac had the same issues. They both had to increase ram during development, because their inefficient code was too bloated. Using high level languages to ease software development was the cause.

[KimmoK]

Some software needs alot of ram because ....

And that again brought something to my mind.
After we got enough RAM (16MB on classic, 512MB on bigger boxes) and especially nowdays I think SW code should never be swapped to disk. Multitasking is more smooth when only heavy data is swapped (the way how one could do it on 68k with ImageFX etc).

It's a shame that for many apps in the mainstream developers spend absolutely no time in optimizing and the end user pays the bill.
(a few days ago I spent two hours waiting for NI DAQ SW to install on a high spec workstation and that kind of driver monsters make any computer to crawl (when silly OS is used). I imagine it would take a few days to re-install everything, so we keep HDD images in safe place in case of HDD hazards.)

[desiv]

.... and especially nowdays I think SW code should never be swapped to disk.

Oh, don't go there..
The "swap or not swap" arguments in the Linux kernel threads are epic...
;-)

desiv

[commodorejohn]

Some software needs alot of ram because of the functionality it offers. The only compromise you can make here is to remove functionality, but then there would be no innovation.

Sometimes this is true, yes - but very rarely compared to the number of applications that just plain take way more than they need.
 

Some uses more than it should because writing perfectly efficient software is alot more expensive & ram is cheap. The only compromise you can make here is delay the software and charge more for it, the odds are the developers would run out of money.

Writing perfectly efficient software is one thing, but these days most developers don't even try. There are text editors now that take up 10-20MB just sitting idle with nothing open. That's inexcusable. It's not even that nobody hand-optimizes software anymore, hardly anybody even designs for efficiency on any level these days.

Also RAM is only cheap by comparison to how it used to be. Any money you have to spend upgrading a computer that would otherwise suit your needs perfectly well because bloaty software is thrashing the disk is not "cheap" by any measure.
 

Oh, don't go there..
The "swap or not swap" arguments in the Linux kernel threads are epic...

This absolutely baffles me...sometimes you do have no choice but to rely on disk swapping, but it is so monumentally inefficient that I can't even begin to fathom why you would not avoid it whenever possible...

[lsmart]

It would preemptively multitask DOS applications.

No, it wasn`t. There was some primitive form of virtual process thing, but multitasking is a different kind of beast.

Anyway, there is nothing really magical about preemptive multitasking [..] if you had a A1000 with just 256K of ram, you probably were not doing much multitasking anyway...

Open Notepad, open Amiga Basic, run the Bach-Tune-Demo and write a note while it is still playing. Try this under DOS.

[itix]

There are text editors now that take up 10-20MB just sitting idle with nothing open.

Only 20 MB? Must be efficient program. I just launched Word and it is taking over 40 MB right after launch. I am not sure if there is anything what good old Kindwords couldnt do? :-)

[commodorejohn]

Only 20 MB? Must be efficient program. I just launched Word and it is taking over 40 MB right after launch. I am not sure if there is anything what good old Kindwords couldnt do? :-)

Well, "text editor" is not the same thing as a full-fledged formatted word processor, but I won't argue that the point is just as applicable to your examples

[psxphill]

There are text editors now that take up 10-20MB just sitting idle with nothing open. That's inexcusable. It's not even that nobody hand-optimizes software anymore, hardly anybody even designs for efficiency on any level these days.

You think it's inexcusable, but you have a rather extreme point of view.
 
If you work in software development and aren't lucky enough to work for a billionaire that is as obsessed with efficiency as you are then you'll end up having to make tough choices.
 
After you follow these choices here:
 
[URL]http://en.wikipedia.org/wiki/Project_management_triangle[/COLOR][/URL]
 
 

"You are given the options of Fast, Good and Cheap, and told to pick any two. Here Fast refers to the time required to deliver the product, Good is the quality of the final product, and Cheap refers to the total cost of designing and building the product. This triangle reflects the fact that the three properties of a project are interrelated, and it is not possible to optimize all three – one will always suffer. In other words you have three options:

• Design something quickly and to a high standard, but then it will not be cheap.
• Design something quickly and cheaply, but it will not be of high quality.
• Design something with high quality and cheaply, but it will take a long time."

Which compromise would you make?
 
If you pick throwing money at it, then there is no guarantee that your product will ever recoup the money. Plus you run into Brook's law (http://en.wikipedia.org/wiki/Brooks's_law).
 
So you could decide to take longer but then as your project rolls on not making money, ram is actually getting cheaper & cheaper all the time. You're going to have to charge more and users will ask you to justify it. The software runs in 100mb you'll say, they will answer: but my phone has 100gb of ram in it & I've been running your competitors cheaper software that uses 1gb of ram for the last year.
 
Good is something you'll have to compromise on. Unfortuntaly design decisions made at the start of the project are often unfixable by the end. This is usually worse on software you've tried to optimise, object oriented code is often easier to change but it comes with a higher performance penalty to start with. I've worked on a project that moved from C to C++ and while some things got slower there was also alot of things that became faster.
 
Some people go the open source route as without a boss you can spend as much time as you like writing the software. But this is only a short term fix. It will only work while there are people that grew up with being obsessed about efficiency spending their free time to write the software. They need a job to pay the bills & open source software doesn't generate money for software developers. The only money is in web sites and scripting, which isn't the type of job that an optimisation junky will go for.
 
 
Either use what makes you happy or become happy about what you use, or you'll end up making yourself ill.

[commodorejohn]

You think it's inexcusable, but you have a rather extreme point of view.
 
If you work in software development and aren't lucky enough to work for a billionaire that is as obsessed with efficiency as you are then you'll end up having to make tough choices.

I do work in software development, and I do it backend, on processes only the IT staff will ever use once per day, for a company that pays me the same in any case and only cares whether things are up and running on-time. Even so, I at least put some kind of thought into making things reasonably efficient and not a huge waste of memory and CPU time. Someone designing software that many people will use multiple hours a day should be putting even more thought into designing software that's light and responsive and not bloaty crap, not less (and that goes for the companies behind them, too.)

[psxphill]

I do work in software development, and I do it backend, on processes only the IT staff will ever use once per day, for a company that pays me the same in any case and only cares whether things are up and running on-time. Even so, I at least put some kind of thought into making things reasonably efficient and not a huge waste of memory and CPU time. Someone designing software that many people will use multiple hours a day should be putting even more thought into designing software that's light and responsive and not bloaty crap, not less (and that goes for the companies behind them, too.)

Unfortunately the companies won't agree with you, the same as they don't agree with the people who work there that want more time to write better software.
 
There are programmers that are incapable of writing good software, but it's quite subjective so you can't sack someone because of it. Most managers in software development are clueless, they just see people standing at a production line churning out product and paying into the managers pension.
 
Basically decent software is doomed to fail all the time any humans are involved in some form in it's creation.

[commodorejohn]

Well, that's basically the problem, isn't it? Software companies (or, actually, pretty much all companies, these days) have zero interest in actually providing quality anymore except insofar as it can be used to justify a higher price tag, because the whole pirahna-pool atmosphere of the modern business world sneers at the idea of pride in a job well done, let alone any other reason for doing anything than making the most money with the least expenditure of effort possible. It's a disease...

[Thorham]

Well, that's basically the problem, isn't it? Software companies (or, actually, pretty much all companies, these days) have zero interest in actually providing quality anymore except insofar as it can be used to justify a higher price tag, because the whole pirahna-pool atmosphere of the modern business world sneers at the idea of pride in a job well done, let alone any other reason for doing anything than making the most money with the least expenditure of effort possible. It's a disease...

Very well said. Greed is what drives these 'people'. Especially the greedy trash at the top. These 'people' need to be kicked down into the gutter where they belong (time for another French revolution).

[psxphill]

Well, that's basically the problem, isn't it? Software companies (or, actually, pretty much all companies, these days) have zero interest in actually providing quality anymore except insofar as it can be used to justify a higher price tag, because the whole pirahna-pool atmosphere of the modern business world sneers at the idea of pride in a job well done, let alone any other reason for doing anything than making the most money with the least expenditure of effort possible. It's a disease...

It's always been like that, we were just delusional sheep in the past.
You describe Commodore extremely well.

[runequester]

The result that will make the person responsible for green-lighting it look as good as possible, and permit blame to be shifted to someone else, if it tanks, is the result that will be pursued.

If it actually works as intended, that's a convenient side-effect.

[Iggy]

All I'm saying is that preemptive multitasking isn't impressive from the computer hardware's point of view, because just about anything can do it properly

That last word invalidates your statement.
Properly would mean efficiently and with a measure of utility and many early microprocessor can not provide that.

The 6809, yes. The 68K, yes. Early Intel processors? Not well at all.
Minix was probably the only example of this that worked reasonably well (until the '386).