firstly maybe make the GUI and the Kernel act independant so when you want to move the app window the window doesn't have drawing problembs when entering a heavy msg loop etc etc.


Ever heard of asynchronous processing?
It is up to the programmer, not the OS. I wouldn't want my OS to decide for me when it will draw and when it will handle messages. The OS has no way of knowing anyway.

I would like to allow there to be a build in coding platform with a JIT debugger plus have all the libaries etc have comments linked to them to act as documentation ie compiled dll's etc have special resources geared at programmers.


You can download and install debug symbols for Windows.

Installing software will be compiling the software, so it is 100% configured for your system... so you would obviously only need 1 copy of the package to run it on diffrent arcutectures and runtime dependantcies - but hopefully keeping it as easy as MSI to install.


.NET can do this.

Namespace structure somethin akin to linux, yet Desktop still top of the Namespace ie no drive letters


While Windows still has drive letters, and probably will have them forever because of legacy support, it also allows you to mount partitions as folders on other drives, so you could create a single filesystem which allows you to access all drives.

mabe have the gui act in gl mode... give some real special effects and make it more interactive with the user.


GL? As in OpenGL?
No need for that, Windows Longhorn will have a GUI based on Direct3D.
Posted on 2004-06-16 04:54:44 by Scali

The OS has no way of knowing anyway.

Wanna bet! Although i don't use OpenBox, Kde3 is able to move the window and you can terminate it by clicking on the X without any problems.... explain that.


.NET can do this

Sofar as i know it is a platform for language portability (as in C#=VC++ and Visual Basic Forms) and not installation.
The other thing is that by th looks of it is that this it a *new* concept... HA!


No need for that, Windows Longhorn will have a GUI based on Direct3D

Btw that started in Win98! The nonclient area had a graded titlebar that is a Direct3D call DXImageTransform.Microsoft.Gradient sounds like something.

About Longhorn... seems kindof like OpenBox3 or KDE3.0.

Another thing, think about this i bet there are more enthusiastic programmers working for the opensource drive then for microsoft - hey even some of them do work for microsoft but what strikes me is the last screenshot of longhorn look like an KDE or apple (BTW a UNIX variant). with a touch of microsoft.

You being a enthusiastic programmer.... weigh it out. And also about the GUI and Kernel... the GUI is an extension of the Kernel and thus should it be treated.... independent (basic systems). Asyncronized programming... well yes but look that throws the problems for programmers to deal with a NODOUBT buggy platform. Shure i don't contest that Linux doesn't have it's downfall but look at the greater aspect of it... it's drive it for the better of the COMMUNITY and not for the pocket. Sadly the lack of it was simplicity and a late GUI and by then then Windows Drive was by far greater.

Admitting that by the looks of the last screen shot of codename Longhorn, they were in a dos box doing IPTABLES ROUTING something i have only been told in how-to in linux. So it is understandable that i feel that Windows is catching up with linux
and finnaly trying to contest with SERVER SIDE software because as of yet they are really lacking and so Longhorn is all in a hype but even now i trust that what is possible in longhorn is already done in linux/unix!


Posted on 2004-06-16 05:17:28 by Black iCE
Wanna bet! Although i don't use OpenBox, Kde3 is able to move the window and you can terminate it by clicking on the X without any problems.... explain that.


Let's not take KDE as an example of how a GUI should work, please.

Sofar as i know it is a platform for language portability (as in C#=VC++ and Visual Basic Forms) and not installation.


It can either compile at runtime or at install-time, so it can do what you want, as I already said. It doesn't hurt to look into technologies that you are not familiar with, you know.

apple (BTW a UNIX variant).


Apple uses its own GUI subsystem, which has about as much to do with X and *nix as Windows.

So it is understandable that i feel that Windows is catching up with linux


Really, Windows is not the one that has catching up to do.
You do not want to discuss *nix vs Windows with me.
Posted on 2004-06-16 06:08:41 by Scali
You can't change a blind zealot... you can only put them out of their misery :tongue:
Posted on 2004-06-16 06:22:40 by SpooK

In 1982 Apple paid UniSoft Systems (later UniSoft Corporation) to do a port of Version 7 AT&T Unix to the Lisa. At that time the Lisa had not yet been released. The port was done mostly by Scott Bryan of UniSoft and was finally released in 1983. I know this because I worked at UniSoft from 1982 through 1984. I remember the port being shown at the Winter 1983 Usenix conference in San Diego, CA. Apple never sold the software directly, but it was marketed by a company called UniPress (if I remember correctly).

In 1985 Apple had a prototype computer known as YACC (Yet Another Color Computer). The machine was never made commerically but the hardware side of YACC had some influence on the Macintosh II. UniSoft again did the port of Unix (AT&T System V). I wasn't at Apple or UniSoft for this one, but I heard about it when I went to Apple in 1986.

Many people have heard of A/UX. It actually started in late 1985. The initial idea was to port Unix to the forthcoming Macintosh II so that it could be used as a file server in a closet. I joined the group in early 1986. (I was the sixth person in the group.) The initial porting work was done, yet again, by UniSoft (AT&T System V release 2 with some Berkeley add-ons). The later evolution (after 1987) did not involve UniSoft.

http://www.cfcl.com/~eryk/weblog/archives/000183.html

Also i would like to say that while Windows is an operating system, and that KDE is a Window System and thus the Kernel and the GUI are kept seperate.

Multiplay Desktops support in Longhorn.... look around find some other similarities.mmm. How many people will @ corporate level put a Windows Server in for remote access or a Linux flavour? Ask yourself? Windows acts well as a domain controller and client pc but Server Side... well as i said there are better.

Spook to whom were you refering to... is it not my right if i would like Windows to be more like Linux. I did admit that Windows is a good GUI but there are problems. What was the topic of this form.... "if I had written Windows" so please don't be insulted by my views, i am allowed to have them aren't I?
Posted on 2004-06-16 06:26:13 by Black iCE
Nice Apple history, but I thought you were talking about Macs today, not 20 years ago. And they have their own GUI system as I said.

Also i would like to say that while Windows is an operating system, and that KDE is a Window System and thus the Kernel and the GUI are kept seperate.


I think everyone knows how KDE works. I just said that we shouldn't take this as a good example of how a GUI should be working. The reasons for this should be obvious.

How many people will @ corporate level put a Windows Server in for remote access or a Linux flavour? Ask yourself? Windows acts well as a domain controller and client pc but Server Side... well as i said there are better.


You'd be surprised how many people use Windows servers. Intel and MS run their websites on Windows, for example.
And yes, you said there are better. You didn't say why though, so it doesn't convince me at all. I can say Windows servers are better, and it would have just as much value as your words, without any additional backup being given, namely none.
Posted on 2004-06-16 06:53:22 by Scali
OK about mac's then the windows xp theme look is a bit similar to Mac Os 10.

May 27, 2004
Linux Servers Up, Unix Down: Survey
By Michael Singer

Linux servers are on the rise again thanks in part to low-end (under ,000) x86-based servers, according to a new survey by market research firm Gartner.

The report said server unit growth for the March quarter was up 27 percent (1.6 million units) over last year while worldwide revenues grew 9.3 percent to .8 billion. The survey confirms the now well-established hum of strong growth in the server sector.

"On a regional basis, the United States continued to be the largest server market, as it accounted for 37.8 percent of global server revenue," Michael McLaughlin, principal analyst for Gartner, said in a statement.

Despite legal threats from SCO Group and competition from Microsoft, Gartner's report said Linux continued to be the growth powerhouse in the operating systems server market, with a revenue increase of 57.3 percent in the first quarter of 2004.

Windows continued to be the operating system of choice for servers with revenue hitting 35.1 percent of total market share. On a shipment basis, Windows dominated all others with 69.4 percent of the OS server market.

As for Unix, the study said the OS suffered a decline with first quarter revenue down 2.3 percent from the first quarter of last year.

Gartner still found a strong need for mainframes. Revenue for the monster boxes totaled

May 27, 2004
Linux Servers Up, Unix Down: Survey
By Michael Singer

Linux servers are on the rise again thanks in part to low-end (under $5,000) x86-based servers, according to a new survey by market research firm Gartner.

The report said server unit growth for the March quarter was up 27 percent (1.6 million units) over last year while worldwide revenues grew 9.3 percent to $11.8 billion. The survey confirms the now well-established hum of strong growth in the server sector.

"On a regional basis, the United States continued to be the largest server market, as it accounted for 37.8 percent of global server revenue," Michael McLaughlin, principal analyst for Gartner, said in a statement.

Despite legal threats from SCO Group and competition from Microsoft, Gartner's report said Linux continued to be the growth powerhouse in the operating systems server market, with a revenue increase of 57.3 percent in the first quarter of 2004.

Windows continued to be the operating system of choice for servers with revenue hitting 35.1 percent of total market share. On a shipment basis, Windows dominated all others with 69.4 percent of the OS server market.

As for Unix, the study said the OS suffered a decline with first quarter revenue down 2.3 percent from the first quarter of last year.

Gartner still found a strong need for mainframes. Revenue for the monster boxes totaled $1.7 billion in the first quarter of 2004, a 12 percent increase from the same period last year.

Each of the major vendors of the low-end servers, IBM, (Quote, Chart) HP, (Quote, Chart) Sun, (Quote, Chart) and Dell (Quote, Chart), have something to brag about -- depending on how they read Gartner's tea leaves with respect to product cycles, industry segment emphasis and geographic demand patterns.

Although IBM made the most money from servers in the last three months, HP sold the most units, and Dell gained the most ground, said Mark Stahlman, technology analyst with Caris & Company. He told internetnews.com the only vendor who is positioned for significant upside in this market is Sun.

Why? For the simple reason that they are the most significant new entrant in this high growth market.

"Having mistakenly avoided this market over the past years, Sun is now aggressively selling into it -- particularly with its V20z 2-way Opteron box," Stahlman said. "As Sun ramps volume on this product and adds 4-way systems (expected to be announced at the upcoming Shanghai event) and expands its product-line with new designs from Andy Bechtolsheim's group, it is expected that Sun will gain substantial market share."

Sun has been working very hard to make up for lost time, bolstering its volume server lineup with support for x86 products. The company recently said its Hardware Compatibility List (HCL) for the Solaris OS on x86, doubled in size in the last six months. Key ISVs like BEA, Sybase, Oracle, and VERITAS have already signed on to support their solutions on Solaris x86.
.7 billion in the first quarter of 2004, a 12 percent increase from the same period last year.

Each of the major vendors of the low-end servers, IBM, (Quote, Chart) HP, (Quote, Chart) Sun, (Quote, Chart) and Dell (Quote, Chart), have something to brag about -- depending on how they read Gartner's tea leaves with respect to product cycles, industry segment emphasis and geographic demand patterns.

Although IBM made the most money from servers in the last three months, HP sold the most units, and Dell gained the most ground, said Mark Stahlman, technology analyst with Caris & Company. He told internetnews.com the only vendor who is positioned for significant upside in this market is Sun.

Why? For the simple reason that they are the most significant new entrant in this high growth market.

"Having mistakenly avoided this market over the past years, Sun is now aggressively selling into it -- particularly with its V20z 2-way Opteron box," Stahlman said. "As Sun ramps volume on this product and adds 4-way systems (expected to be announced at the upcoming Shanghai event) and expands its product-line with new designs from Andy Bechtolsheim's group, it is expected that Sun will gain substantial market share."

Sun has been working very hard to make up for lost time, bolstering its volume server lineup with support for x86 products. The company recently said its Hardware Compatibility List (HCL) for the Solaris OS on x86, doubled in size in the last six months. Key ISVs like BEA, Sybase, Oracle, and VERITAS have already signed on to support their solutions on Solaris x86.


yes i know that windows servers are in the demand. But as the report said, Linux is the growth powerhouse in the market.

http://www.internetnews.com/stats/article.php/3359981
Posted on 2004-06-16 07:01:42 by Black iCE
Damn... but I AM writting my own "Windows" / GUI

Is called SOLAR OS ;)

This way I can implement my own ideas while I am still using Windows and KDE as they are :D
Posted on 2004-06-16 07:31:04 by BogdanOntanu
Firstly i feel that i can't even start to fathom what to do. Starting with boot code -> os loader -> GUI and then memory management. etc etc.

I have looked around and i am to lazy to start to boot i have to first finish a windows app before i'll try doing a console os on my own.
Posted on 2004-06-16 07:41:28 by Black iCE
OK about mac's then the windows xp theme look is a bit similar to Mac Os 10.


Yes they both use windows, scrollbars, menus, mouse pointers etc. As do all other GUIs, what's your point?

But as the report said, Linux is the growth powerhouse in the market.


Yes I've been hearing about linux growth and improvements since the day the first version was released, and how it would replace Windows soon. I've never seen much of it though.

Why do you come here if you like linux so much? This is still a Windows forum, you know. I doubt that anyone is interested in unfounded linux zealotry. In fact, it could be considered trolling.
Posted on 2004-06-16 07:51:13 by Scali
To answer your post please reread all my above posts cause you first started with me! I was conveying my point and my wishes with the topic consirned. If you would like to do a search on my previous posts then you will find that i do support windows. It is my own wishes and you wanted me to give you facts about what i was saying so i did. what else to you want - i am here to try and give help and ask others some questions and if i wish to take part in the heaps "I wish list" them i may.

This is not to offend you it is just that right now i do have good things to say about os's and bad aswell. if i post about it yes you may argue then i, if i whish, go forth and prove my point.

Unfortunatly you viewed this as a *nix vs windows arguement based upon my desires for the topic at hand.

Kindly,
Black iCE
Posted on 2004-06-16 08:00:29 by Black iCE
Scali: Of course, if the only effort you make to reduce your program size is using the loop instruction, you won't have much success. There are often many places where you can use different instructions and save a few bytes. Added together, all of this can cut a program's size by a significant fraction.

Also, you seem like a naive programmer who has only ever seen x86 up close. x86 has flags that may or may not exist on other CPUs, and vice versa. It would come down to having to create a pseudo-flags register in memory somewhere, and having to emulate certain flags on certain CPUs.

It wouldn't take much intelligence to figure out that functions defined with a BOOL type could return their result in a flag on systems that support it (and where it's practical) and as a number in a register on other systems.
The usage of extensions such as special calling conventions could very well be hidden away in conditionally compiled blocks in the header files pertaining to the API.

Whatever, bx had a special function in 16 bit. And that is why it is still saved today, regardless of whether the PASCAL calling convention was the most popular or not.

If that's the case, then why did they choose to use BX as scratch precisely for that reason? And what does it have to do with anything anyway? In PASCAL, EBX is preserved in 32-bit code. I've never encountered any calling convention used in 16-bit compilers where BX is preserved.
Posted on 2004-06-16 08:25:45 by Sephiroth3
Added together, all of this can cut a program's size by a significant fraction.


Nonsense. It may save a few kb as I said above (hardly significant), but the size is mainly dictated by the data used (GUI resources, strings, bitmaps, tables etc), in many cases. That is where the real gain is, and that is a fact.

The usage of extensions such as special calling conventions could very well be hidden away in conditionally compiled blocks in the header files pertaining to the API.


What's the use? Only compilers on platforms that support it would use it, and the gain would be 0.
Let me put it this way: if it really was a good idea, do you really think that nobody ever thought of it before, in all those years of OS and compiler design by people much smarter and more experienced by you? No, it is too obvious an idea. We live in an age where all the obvious ideas are known already. To think otherwise is both naive and arrogant. It is also typical for people who have not coded in other languages than x86 asm much.

I've never encountered any calling convention used in 16-bit compilers where BX is preserved.


I have.
Posted on 2004-06-16 08:36:05 by Scali
My last 2c ...

if it really was a good idea, do you really think that nobody ever thought of it before, in all those years of OS and compiler design by people much smarter and more experienced by you? No, it is too obvious an idea. We live in an age where all the obvious ideas are known already


Mabe so, but never underestimate the fact that there are ppl out there who would like to take a diffrent approach and how is an idea concieved? Mabe others have thought of it but the credit goes to those that have made it a reality.

Also Even the smater and experianced can faulter... we are all human.
Posted on 2004-06-16 08:44:24 by Black iCE
What's the use? Only compilers on platforms that support it would use it, and the gain would be 0.

I'd say the PC with it's x86 is a very important platform. Remember that Windows was developed originally for the PC, and almost every household has one. There are many other processors that allow you to set the result in one instruction and branch with another, such as the ones listed in this table:


Processor Set cond. Sz Clear cond. Sz Use Reg. Sz Branch Sz
--------- --------------- --------------- ----------- -----------------
680x0 moveq #1,d0 :2 moveq #0,d0 :2 tst.l d0 :2 beq false :2
65xx sec :1 clc :1 cmp #1 :2 bcc false :2
R series addiu r0,zr,1:4 addu r0,zr,zr:4 - :0 beq r0,zr,false:4


Let me put it this way: if it really was a good idea, do you really think that nobody ever thought of it before, in all those years of OS and compiler design by people much smarter and more experienced by you?

Judging by the code their compilers generate, they can't be that smart. When I look at the code in various components of Windows I see that the code is many more times larger than what a human programmer would write.

There would be no reason for Microsoft to base their decision about their calling convention on what was most practical in 16-bit code or what was used by some little-known compilers. There's no way someone would write a Win16 program in a strange language that didn't support the PASCAL convention, much less call a Win32 API from it.
Posted on 2004-06-16 09:34:14 by Sephiroth3

Judging by the code their compilers generate, they can't be that smart. When I look at the code in various components of Windows I see that the code is many more times larger than what a human programmer would write.

Even WinXP is compiled with one of their older compilers, afaik - I guess Microsoft wanted to be sure there weren't any unknown bugs in their newer compilers, before they build the entire OS with it... XP SP2 is supposed to be compiled with a recent compiler, probably vc2003.
Posted on 2004-06-16 09:37:57 by f0dder
There are many other processors that allow you to set the result in one instruction and branch with another, such as the ones listed in this table


Yes, a table that includes 68k and 65xx, two CPUs that are a few decades old, and haven't been used in desktop PCs for ages. What useful info!
Get a clue, please.

Judging by the code their compilers generate, they can't be that smart.


If it were as easy to make compilers generate optimized code as it is to write it by hand, we wouldn't be having this discussion.
You obviously have never written a compiler before. I suggest you don't try to judge people that do.

I see that the code is many more times larger than what a human programmer would write.


I'm a human, and I don't care one bit about code size. I only ever use asm if I need fast code. Small code is of no use for regular PCs. And small code is significantly slower.

There would be no reason for Microsoft to base their decision about their calling convention on what was most practical in 16-bit code or what was used by some little-known compilers.


They made the decision to have 3 volatile registers, that means one of the 'x' registers had to be made non-volatile. The most obvious choice would be ebx. And it is not Microsoft that made this decision afaik. You also find it with many non-Microsoft products for x86.
Posted on 2004-06-16 12:31:29 by Scali
I dont know what I would change in Windows! It has everything needed!
Posted on 2004-06-16 15:00:31 by x86asm
I'm not sure if these are implemented in Windows or not, but I would reprogram, the Interrupt Controller to map the IRQs above the 20h interrupts. Intel reserves the first 32 interrupts and there were always problems from the superposition of the IRQs with the Intel interrupts/exceptions. You had to test if an interrupt is an exception or an IRQ.
And I would reprogram also the PIT's frequency to 10 or 1 or power of 2 millisecond, to eliminate that 18.2 millisecond shit.
Posted on 2004-06-16 15:47:06 by bszente
a teacher of mine said there were programs with hundreds of megabytes of code.

i found it hard to believe and asked if all was code.

he said yes.

but he also said "you know, give amounts of resources to researchers and they ll always find a way to use it".
thats true for data and processing power, but...
i wonder if the things he was referring to really contained that much of machine instructions.

i mean, how much space is taken by binaries in, say, windows, linux, or quake, or autocad, which are pretty big systems? i m talking of machine instructions, not filesize. even with compiled code i find it hard to figure. and there are dlls...
anyone knows?

maybe generated code, unrolled loops? thats kind of dull, and i dont consider this being really what i meant.

do you know about such software or operating systems?
as scali said, almost everything is taken out by data files. (and resources in exes are data files)

what is the common space ratio from src text files to binaries, for:
-asm
-hlls

thanx
Posted on 2004-06-16 15:54:40 by HeLLoWorld