OSDev.org

Schol-R-LEA, i like the fact that you putting a lot of efforts into your comments related to the Dawn operating system, i see from your attempts that you focusing serious attention into the conception to try creating a deeper understanding about the realibility of solutions like this.

but i guess its not a question of arguments, i guess i told everything from the system, critics has told everything, every standpoint is clearly can be seen and can be understanded. from now, we will see after hardwares becoming more available, how properly the OS will fullfill its purpose.



also, update:

2017, march 29. - Optimizations and bugfix
The error line handling in compiler is improved.
Imrpovements to fix mouse and network packet losses on low rendering speed
Operating system speed is improved by 10%

Korona: today the most popular risc technology is the arm, and probably the only notable and existing cisc architecture is the x86. by sales, ARM outperforms x86 by 95:5, also there is prety much more arm manufacturer and creator compared to x86, so currently we can say that risc/arm conception is winning over x86, and x86 is surviving only on a small segments of desktop computers for historical reasons.

HOWEVER

its not so simple, becouse the ARM architecture is not really much risc any more.

even at first arm cpus were just almost risc, they was somewhat a bit more complicated when we compare to other risc cpus. then they got new kinds of instruction extensions, they got two kinds of floating point unit, then they got SIMD units (arm neon). then they got (and removed) hardware JAVA instruction compatibility called JAZZLE, then they got variable instruction length (THUMB instruction set), and finally they obtained support for 64 bit binaries, while still keeped the hardware compatibility for 32 bit binaries. arm instruction set is still not as much complex as x86, but its closing up. we can prety much say its not RISC any more, which is funny, becouse even the company name is advanced RISC machine.

Geri wrote:mikegonta, how not creating a bootable x86 emulator by myself will rip my OS, when the os is not even x86-related?

The same way that releasing the source code will prevent it from being implemented on an x86 PC.
The interest in doing a simple bootable game loop to run your open source GUI is totally different from that of doing a closed source
obfuscated code emulator.

mikegonta wrote:

Geri wrote:obfuscated code emulator.

you have totally right in the point when you say that once a code is compiled into subleq, its very hard to morph it back to high-level language forms. x86 originally is also some type of hardware obfuscator, due to the variable instruction lenght and due to the fact that on x86 you must decode the instruction to be able to access the next instruction. this is one of the reasons why software corporations liked x86 at the early times.

subleq binary code shows somewhat similar behavior, the difference that the approach for cpu is the opposite in any ways.

Geri wrote:you have totally right in the point

That's not the point and you are totally missing the point.
Had you come here with a closed source x86 binary OS that requires the making of an OS to load and run your OS you would have gotten
the same response (and the same response on the hardware forums).
The discussion has gone way beyond the OS and that's why I'll let it Rest In Peace.

mikegonta wrote:

Geri wrote:you have totally right in the point

Had you come here with a closed source x86 binary OS

this is not an x86 OS, please read the description

@korona: I agree with your point entirely, I hadn't really intended to imply that RISC was better in any absolute sense, especially now - the argument RISC could potentially have superior performance was more persuasive in 2006, but even then, not by much. The RISC vs. CISC debate is sort of meaningless now.

However, I would say that if I had to choose an architecture for widespread adoption, x86 would be the last one I would want - it is simply a poor design, regardless of whether you favor CISC or RISC. Even Intel agree on this point. As I said, they never intended it as a long term design; it was meant to fill a gap in their release schedule until a product that never got finished was ready for sale in the then-booming workstation market. It was only meant for use in embedded systems, in any case, so they didn't care if it was an awful monster to program since the people who would be using it wouldn't care, either.

They were more dismayed than anyone when IBM chose it for the PC, if only because home computers didn't fit into their plans in the first place (they had even stopped selling 8080s to microcomputer manufacturers, which is why the Z80 overtook the 8080A in that area; they only let IBM browbeat them into selling 8088s for the PC because IBM was computing's 800-lb. gorilla at the time). Its success - and the effort that they were forced into applying towards keeping it going long after they meant to, for the sake of avoiding a massive switch-over and tons of consumer hostility - is purely a historical accident.

Anyway...

Beyond that, I think Mike is right. It would be best if one of the mods locked this thread, IMAO.

Geri wrote:

mikegonta wrote:

Geri wrote:you have totally right in the point

Had you come here with a closed source x86 binary OS

this is not an x86 OS, please read the description

I know you aren't a native English speaker, so perhaps you haven't run across the word 'analogy' yet. What Mike was saying is that someone who brought in a closed-source x86 OS that yadda yadda... would have gotten the same reception you have. His point was that if we were being hostile, it was not so much towards the idea of a SUBLEQ based system, or Dawn in particular, but the specific ways in which you were presenting and defending it.

2017, march 30. - Various bugfix
Toolset control in paint and text editor is improved.
Bugs fixed in pow, log, exp.
Bugs fixed in calculator

2017, march 31. - Disk handler bugfix
Disk read/write stall glitch fixed when using fread/fwrite .
A bug fixed in the task handler.

2017, april 1. - Sheduler bugfix
Minor optimizations in the sheduler.

2017, april 9. - Touchpad and low-spec optimisations
Dawn can handle the touschreen hardware as an absolute mouse better.
On low-speed hardware, Dawn will disable the wallpaper on default.
Bugs fixed in the desktop settings menu.

I think you should add anti aliasing to the text. It would greatly improve the interface.

obiwac: thankyou for your hint. aliasing is implemented, but disabled due to performance issues.
since dawn runs in an emulator at the moment, and subleq have no multiplier units to render the gui more quicker, antialiasing at the moment canot be enabled.

(for the note, when aliasing is enabled, it produces very similar smooth fonts quality like your OS does)

2017, april 10. - Low-spec optimisations
Added more optimisations on weak hardware.

Hi,

Geri wrote:2017, april 10. - Low-spec optimisations
Added more optimisations on weak hardware.

You need to add a "change log" to your website. This forum is for announcements, and it is not intended to be used for your personal "revision control system commit history".


Cheers,

Brendan