OSDev.org

Dario wrote:

Chandra wrote:Why Linux worked? Simple because it was programmed to be _usable_ system by only _one_ programmer and after he made code available to others it was only the question on how to perfect it. And even then....I don't think that anyone expected to grow so big. Other then that...we are already too late...20 years to be precise. 80s and early 90s were arena of operating systems. Well, today thanks to the mobile devices they are somewhat back to mainstream, but still no newcomers.

I view Linux little of a disaster actually. Linux is almost 20 years old and it is starting to become usable as a modern operating system. Go back a few years and it didn't have a stable driver/kernel API. You had to compile the kernel in order to add new drivers to it. Even if I'm a software engineer, I would go ape **** if I had to do that in order to install a driver. Mounting was a mess and often had to be done manually. Early graphical UIs was ugly and almost unusable.

Linus Torvalds has written about 2% of Linux and the rest have been added by existing code or other developers. Linux is based on very traditional technology, basically a UNIX derivate. Many of the interfaces had already been designed as a result of supporting many of the POSIX interfaces, so Linux isn't really something new or groundbreaking. Sure there are interesting solutions inside the kernel, but no new paradigms. Linux developers tends to be more conservative when it comes to technology than your most conservative politicians.

Ubuntu finally made Linux somewhat usable I must admit, and those developers realized that you had to create a easy to use interface and also automate many tasks. Despite these improvements, I still prefer Windows as workstation OS. Windows have been nice for over a decade now and what Windows did easly and automatically ten years ago, required software engineer skills and in depth knowledge to do the same on Linux OS about that time.

Now, Linux have been very successful in embedded systems and one reason for that is that is that it is hidden. Either you don't see the system at all (a router for example) or you have a UI that completely hides all of the operating system features (Google Android for example). So the success of Linux is acutally when people don't have see it and deal with it.

All of this isn't the fault of the Linux kernel because after all it is just a kernel. Many of the bad things comes from added software around the kernel which often must be compatible with old legacy software originating from the 70s.

I see a bright future for new kernels, new CPUs and systems will require new ways of dealing with them efficiently. Embedded system world will not be as conservative as the PC market so there will be a nice variation of OSes.

OSwhatever wrote:Go back a few years and it didn't have a stable driver/kernel API.

Still doesn't. Driver binaries are still specific to a given kernel version.

You had to compile the kernel in order to add new drivers to it.

What you mean is the invention of kernel modules (instead of having to compile the drivers into the monolithic kernel proper). You still have to have modules (drivers) that match the kernel version...

All of this isn't the fault of the Linux kernel because after all it is just a kernel. Many of the bad things comes from added software around the kernel which often must be compatible with old legacy software originating from the 70s.

The whole concept of patching your system together from many different sources - kernel here, coreutils there, shell another place, and so on ad infinitum - places a severe burden on those who want to "make it happen", i.e. the distributors. That's fine for the developers though because they just want to play, unencumbered by concerns as to what their decisions mean for the "downstream" people. Add to that the amount of redundancy introduced by having literally dozens of distributions, and you see some severe manpower burned because the underlying concept sucks.

That's not talking about the amount of problems users experience because of "upstream issues". The past stories regarding WLAN drivers (madwifi, anyone?), device filesystems (devfs, UDEV), and video drivers (fglrx / ati-drivers) are bound to repeat themselves.

I know that there is no best OS. Every OS has it's own advantages and disadvantages.
But my question is: what is the most developed Operating Systems of OSdev members do you know for the current time?

Hmmm... BareMetal and Visopsys managed to go corporate to some extent, which is quite an achievement in itself. Then like (almost) everyone, I'm impressed by shiny graphics, so I have a soft spot for those OSs which exhibit pretty GUI in the OS screenshot section. Also, Valix and Metta (berkus' project) look pretty cool too, but from a theoretical standpoint this time (current implementation is still in too early stages to be exciting).

Let's say you want to make an OS that is better then Windows: You gotta start small

Windows had a big group of professional developers who lived probably close together and work months on the same project, here you may find (slightly) more people but we would be divided by great distances, therefore it'll be harder to communicate, you could use some sort of chat but when would you communicate if you live in a totally different time zone? And as you may know the key in teamwork lies in communication.

One needs to agree on the criteria first. In the end, "good", "better" and "best" mean how well an OS meets its design goals. There are no good or bad design goals, but there are good and bad designs. What are the design goals of Windows, Linux and your hobby OS?

I am surprised that no osdev members hobby OS, did not make it into the "10 best alternative operating systems of 2011" list .

http://www.techradar.com/news/software/ ... ems-934484

I'm sure I've seen Dex around here somewhere ...

Dex was here long ago. The banhammer has taken his place. And IMO only the art got that OS on the list. Technically its just inspired by DOS with all of its design qualities.

Combuster wrote:Dex was here long ago. The banhammer has taken his place. And IMO only the art got that OS on the list. Technically its just inspired by DOS with all of its design qualities.

What!, the only one on the list and you banned him/her

Or the ban is a clue as to how he managed to get there

Combuster wrote:Or the ban is a clue as to how he managed to get there

I think he may of done the same to this site
http://maketecheasier.com/dexos-the-tin ... 2011/03/24

and this..
http://royal.pingdom.com/2008/09/26/10- ... he-future/

Those pesky hackers.

Dario wrote: Why Linux worked? Simple because it was programmed to be _usable_ system by only _one_ programmer and after he made code available to others it was only the question on how to perfect it.

You are wrong.
a) Linux does not work (not the way it should). We have horrific issues day by day.
b) Linus never programmed a "usable system", all he did he wrote a kernel-skeleton, with mostly empty syscall handlers. See kernel.org. He never coded any userspace app, for example a shell or init that would make a system usable. He wrote kernel space only.
c) Linux kernel is not written by only one programmer. Never was (save the incomplete 0.0.1). See kernel.org.

Why it worked? I assume because it was bloated and huge enough to make big companies' managers happy (reading: http://news.cnet.com/8301-13505_3-10358024-16.html). In my several decades in IT I noticed that whenever a decidion had to made, always the worst solution (regarding to technical matters) was chosen and became mainstream (Amiga vs. PC, SCSI vs. IDE, OS2 vs. Win ...etc.). What's good and what's saleable is quite different.

Posted: 2011-01-05, 17:14

Let it rest.

Solar wrote:Posted: 2011-01-05, 17:14

Let it rest.

Oh, sorry. I apologize I missed the date.