operating system in thinbasic architecture ?

[ioptic]

I am quite impressed that thinbasic works so well and am wondering if the architecture can be used to code a replacement operating system ?

It appears that it just loades the modules it needs to complete a specific task (recursive/repeatadly, if needed and apparently in an isolated spawned memory location- ie. no chance of mix ups etc...).
When comparing this with the Windows arcitecture which loades everything and is bloated, prone to dll and version mix ups, I was wondering if an
operating system could be written based on the thinbasic architecture ?
Theoretically it would be - not just thin- but skeletal :) the exact opposite of windows and probably more robust and a lot faster... leaving powewr to burn to do the fun stuff :)

so to load a word processor:-
load program, printer driver, connect to a shared memory clip board location..., perhaps a file/fat reader with network capabillities, a graphic interpeter, a tcpip internet connection and Bingo...

What does everyone else think ? Is windows realy such s difficult program to replace with something faster ?

rob
-o-o-

[ErosOlmi]

Better to not exaggerate.
thinBasic is just an interpreted programming language that needs a Windows operating system in order to work.
:D

[ioptic]

I know that:) but...

I was refereing to the architecture / method of using system resources.
When you look at the overheads of an OP and then look at what is actually used, most of the resources are not used- but slow the computer down, cause crashes etc... Then compare that with loading only what is needed to perform the task, and it's tiny by comparison... which beggs the question why OP's have to load everything and why they don't just load what is needed to perform a specific task?
The issue becomes even more interesting when we see the limits of CPU powers and their overheating issues/ costs etc... so it would make good sens to think about a lighter OP :) a bit like making smaller people to fit more into a house that can't be made any bigger. :) Everyone is talking about faster CPU's etc... so why not use logic and talk about making smaller OPS's ? :shock: ... should have the same effect ? :P

rob
-o-o-

[ErosOlmi]

I think we cannot reduce it to a simple considerations. Microsoft comes from many years of Operating Systems and in most of the cases all applications running many years ago under Win95 are still working under Windows Vista. Not many other OSs can share the same binary compatibility for so long time.

Of course this means bloating because you cannot put away all what you want but you have to keep it for compatibility purposes. I'm pretty sure MS would have rewritten many of the core engine libraries of its OS but they didn0t and we can still program in the same way but at the same time go into other directions, .Net for example.

This is also possible thanks to the increase of power recent CPUs have given to the market. CPUs are so fast now that even an interpreted language like thinBasic can be considered valid for quite big and complex applications. Few years ago it would be impossible.

So, my idea is to use what we have the best way we can following MS documentation.
Documentation is a must. I spent thousands of hours in studying MS docs and even now there is no day in which I'm not able to learn something new or correct something I erroneously implemented. And this is positive from my point of view.

Ciao
Eros

[ioptic]

Legacy programs are surly a big issue. But being a curious creature, I can't help wondering how big an operating system realy has to be? The other issue which facinates me is the emergence of open source code. Linux now has a flavour for everything, from a wrist watch chip to a super computer :) This appears to be down to evolution and being able to change and swap components and priorities. Windows is Ok in itself, but having closed itsef off from evolution, it has become an bloated unstable dinasour. Dare I mention Vista and its incompatibillity with many of Ms's own-old products? :)
I am sure that these hick-ups are not always deliberate marketing ploys to make us buy the latest stuff. But if we grant them that these issues are not deliberate, it also means that they have lost control over the complexity of their code and these glitches are unintented bugs. Which does not refelect well on them at all!

For me, it's always Ocams rasor "Keep it simple!" and an OP that only loads what is needed to run a program, reduces complexity and unintentional errors. Besides an OP which is modular can have every module improved individually- which opens up the dead-end of evolution- to competition and ultimatly better ideas.
It could also load older hock-ins as needed, giving total backward copatibillity. A good example of how well modular, interpreted stuff works, is the internet. Built as an Atomic war communication tool, the core has hardly changed in 50 years :) but, boy, does it have some hot add ons and improvements hanging of its small, old and uncomplicated thin frame :)

So perhaps- lets not be too hard on microsoft, let's just hope they untangle, modularise, simplyfy and downsize their mess before someone else kicks them of their pedastool :)


Best regards
ROb
-o-o-

[Michael Clease]

I heard somewhere that microsoft are looking at creating a base OS which then runs virtual machines.

[Charles Pegge]

When you consider what could be acheived with 16k bytes of software on the pre-PC home computers, it is evident that Operating systems have become massively inflated. XP/Vista is not 500,000 times more useful than a Commodore PET or a BBC micro system, as the multi Gigabyte expansion OS size would suggest.

But I suspect that much of this space is taken up with bulky multimedia files as well as duplicated library functions. Size and speed are very cheap and tight coding is considered expensive. Besides, the PC industry needs to sell lots of new hardware to support itself :)

It would be very nice to have a core operating system that booted up in a fraction of a second - as the original home PCs once did.

[ioptic]

16K, In think my wristwatch has more :) Just imagine how many resources would be available if the OP could be loaded on demand ! :)

I have an old charting program. Novachart vintage 1994. 192k, Zero dll's and still working in a flawless graphic mode on a 22" screen!
So I remain unconvinced that it can't be done. Also the Thinbasic team needs a mention here. They are 1st class pros and deserve our gratitude!


Just imagine what whopper sized, super fast games could be run on a thin OP?...

meanwhile back at bloat OP central ...:) we just have to keep the reinstallation Cd handy- when Win tripps up. I spend at least 2-4 weeks every year fixing Win- so waiting for the re-instalaltion to complete, I've had plenty of downtime to ponder OP improvenents :)

best regards
Rob
-o-o-

[Charles Pegge]

The main problem with developing a new operating system is drivers! Because hardware designers provide OS specific software to run their devices, you would have to offer fairly comprehensive binary compatibillity with MS-Windows or Linux. We are well and truly locked into the system. You can see how independent OS developers, like Haiku have fared over the years:

http://en.wikipedia.org/wiki/List_of_operating_systems

And here is a sobering thought:



This paper refines Ingo Molnar's estimate of the development effort it would take to redevelop Linux kernel version 2.6. Molnar's rough estimate found it would cost $176M (US) to redevelop the Linux kernel using traditional proprietary approaches. By using a more detailed cost model and much more information about the Linux kernel, I found that the effort would be closer to $612M (US) to redevelop the Linux kernel. In either case, the Linux kernel is clearly worth far more than the $50,000 proposed by Jeff Merkey.


http://www.dwheeler.com/essays/linux-kernel-cost.html

So our only practical option would be to provide a simplified virtual OS that sits on top of these planet-sized beasts.