Let's face it, the most important, truly alternative, hobby operating systems that are somewhat usable today are three: MenuetOS, SkyOS and AtheOS. All three are hobby, open source OSes, written from people who enjoy coding low level programming. Read more about the differences between these OSes and which one you might want to try out.
Post a Comment
because each of them are trying to do something different. Anyways, I'm just extremely happy I will be able to install atheos again (as syllable) and be able to expect some updates.
I feel kindve bad about the fork, as now the core parts of atheos wont be as similiar to each other as they would be developed fully by kurt (dev. styles, etc?) and how Kurt didnt want it forked, but perhaps it was destined to. It will at least be an interesting summer with the updates coming forth, hopefully quickly.
Cant wait to install syllable on my athlon 1900+, it was as visually more responsive than win2k on my celeron 800 with 320 megs of ram on a ppro 200 160 megs of ram. I just wish there was some way to compare it other then compile times, etc.
Installation: 6/10
Hardware Support: 5/10
Ease of use: 8/10
Features: 8/10
Speed: 8/10
Dev. Activity: 2/10
For Hardware Support, what have you compared the three OS's against? For AtheOS, have you included all of the available third-party drivers (Which Syllable includes in the base)?
Out of interest, how do the three different OS's compare for things like network support? (O.K, I know how AtheOS does, how about the other two?)
vs the other 2 OSes, Atheos is getting the short end of the stick. Compared to mainstream OSes Atheos is right on in that assessment. Id still say speed is a 9 on atheos. And installation is simpler then any linux distro Ive tried.
Well, interesting article.
Many alternative OSes have really low activity around develop.
We (4 world-wide guyz) started a project to build a "SourceForge-like" portal dedicated exclusively to OS development.
We belive that it could be a revolution around the alternative OS scene, 'cause it could create more co-operation between different developers.
We're looking for PHP-skilled developers, web designers, surfers, and whatever you can offer...
If you're interested, do not hesitate to contact me.
OS Factory Forum: http://more.at/osportal/
OS Factory Forum: http://more.at/osportal/
This website opens extra advertisement windows and tries to install some precision time component on my machine.
I couldn't find any messages to read about this new OS portal development.
The board info says there is 1 registered member.
Is the link correct?
#m
I'm sorry to nit pick on your review of Menuet. There's a statement about inconsistencies arising from procedural programming. I don't agree. I believe this is due to the architecture of the software. Of course I'm just speculating, I haven't seen the source. (It's probably very nice source too.)
Anything that can be done object-oriented style can be done procedural style.
I think OOP is very cool and the advantages are obvious: encourages consistency, higher level of abstraction, data encapsulation, simplification of code. Properly implemented OO code should have no or virtually no overhead compared to procedural code.
I still prefer procedural programming. Most mainstream processors by way of instruction sets and register sets are procedural-centric. If done incorrectly, procedural programming can seem tedious and error-prone. If done correctly, procedural programming is that exact same thing as OO programming except inverted.
The programmer has the impression the code actively operates on data. (This is my impression anyway.)
I feel procedural programming is easier to follow, debug, and is more logical - especially in lower level languages like Assembly. Object oriented assembly does exist; I heard lots (I think from here) about a form being implemented in GEOS back in the day.
Anyway this deficiency you speak of is more the fault of programmer than the language or the style of language. Sorry for nit picking this bad. I just thought I would come to the defense of procedural programming. It seems to really be under the gun lately.
Oh yeah, I don't think MMX, 3Dnow!, and friends would help an operating system much. These instructions are mostly used for recurring operations on larges amounts of data. This is something applications do more often than operating systems. Maybe in fancy graphics routines?
With almost all OOP projects, the biggest problem is tight coupling due to poorly designed hierarchies and inheritance itself. This is one of the major reasons that big Java and C++ projects have so many problems.
The larger OOP projects need very smart architects/coders to make sure the system doesn't end up tightly coupled. If it ends up that way, it is almost impossible to work with the code base. It becomes highly viscous.
#m
This is where you seperate the Men from the Boys. C programmers should stick to writing buggy crap like Windows and Linux.
I used to write tons of assembly way back when on my old 8086/286 for sheer speed. After all working with an 8Mhz clock demands maximum speed. But this was also in the days when one-man-bands could write entire games/apps on thier own in assembly. I still remember my first stab at 32bit protected mode coding, it was like heaven with such a massive address space.
But things change, there is no way I would go back to the hell that assembly becomes on largish projects. Its just not maintainable. I know its a great feeling knowing that registers and memory is being done as efficiently as possible but C/C++/et al do generate pretty good code now, the only advantage assembly really has is executable size.
Needs change, I used to assembly to speed up graphics operations (blitting/parallex scrolling and other tricks) but these days with the advent of GPU's more powerful than the main processor this is just not realistic.
Anyhow...how the hell do you code assembly for something like the Itanium??!!! 
> Anyhow...how the hell do you code assembly for something like the Itanium??!!!
Why should that be too hard? The only special thing that this processor does is executing
predefined groups of instructions at the same time. It is not hard for an assembly coder to identify
nearby instructions that can be done in parallel and make bundles of them.
I recall doing something similar even back with 8088/8087, where you could
write an instruction for the 8087 (the FP unit), then follow that with some integer instructions
that could be processed by the 8088 while the 8087 was working.
Why should that be too hard?
Because IA-64 is VLIW? Last I heard even the Intel compilers have trouble generating decent IA-64 code.
Exactly, anybody who remembers coding back on old processors and how coding was a matter of knowing the exact clock times for each instruction, i.e. knowing the tricks like using xor to clear values, knowing the hit for passing word boundaries. There was pages and pages of clock cycles per instruction. I bloody worn my book out, although I still have it!
Now, I wouldn't want to try the same thing with VLIW, it was hard enough choosing the right instructions never mind trying to put them in the right order to keep the pipelines full (both integer and fp).
Back in the days of the 8086/286, we didn't need to worry about the hit of branching (as much, no cache to flush IIRC), just how long an instruction took. Now things are a lot more complicated.
Once off the 6502, wrote a bit of 68000 asm for the Amiga.
After that, it went downhill. In Intel Inside world, wrote some 8086 asm, some 386 asm, and even some MMX asm. None of it was as simple and fun as that 6502.
Having gone through Intel training for Itanium... I cannot imagine writing asm code for vliw.
Nor can NEC ;-)
NEC's chief architect Len Tsai launched an astonishing attack on Intel's Itanium processor at Bert McComas' Platforms Conference in San Jose today.
Tsai stunned the room by describing Intel's approach as "brute force", and predicted that learning the new instruction set would take "half a generation". He said NEC - which shipped its Itanic2 servers last week, had not seen any demand for the ten-year old processor.
I better get started learning that stuff, eh? ;-)
#m
The article was strange. It completely missed the point of OO.
"not being object oriented each application has its own distinct look instead of inheriting widgets from a main object"
Windows is not strictly OO. Windows has a consistent GUI only because there is a defined API to create a Window, a Button, an EditBox etc and a set of guidelines as to appropriate GUI behaviour (that some developers still ignore.) The MFC is one OO wrapper to the WinAPI, but so then is OWL and VCL from Borland. This just goes to show that the comment itself is incorrect. Even an OS like BeOS has an underlying API within the Appserver which is wrapped by the BeAPI. Just because Be chose not to show this API (or it too is written in C++) doesn't make the OS have a consistent look and feel. A lot of criticism was placed on BeOS back in the DRx days *because* it was so inconsistent, especially where Browser did what Gnome now does with confusing secondary menu items in addition to a windows main menu bar.
"It is fully C++"
What on earth is that supposed to mean? Because it's written in C++ it's good? LOL! Look at Apple and MacOSX. Nice GUI with Objective-C API. (Objective-C is a lot like C++ done right in many respects.) At the end of the day, the OS can be written in COBOL so long as there is a clearly defined, well thought out API in a reasonable language (be it C, C++, ADA or Eugenia's nemesis, Pascal) that developers are willing to use/can understand.
I don't mean to be rude/judgemental, but this article is shoddy, inaccurate and rushed and gives no useful information what so ever.
"especially where Browser did what Gnome now does with confusing secondary menu items in addition to a windows main menu bar."
What do you mean? What is it confusing?
Having a menubar like the MacOS plus menubars on applications like in Windows. One of the Developer Release versions of BeOS (I forget which one(s)) had some menus on the local application, but say, other as part of Browser (browser being an older 'version' of Tracker.) This would be okay, I guess, if there had been consistency in the types of items that appeared, but there wasn't. Sometimes a certain menu was in one place, sometimes another. This is why they hired someone to write a GUI Guideline.
Gnome probably does this better, but I remember it being horrible last time I looked at it. Mac/Amiga style is fine on its own, Windows etc fine on its own - but the two are clunky when mixed.
The menus on Browser belonged to and changed with the current application, like Finder in MacOS...
Yeah right...
I get a large blank page, and a second full-screen page containing an ad.
I do PHP regularly and was kind of interested in what you might be trying to do, but it appears that you're trying to make money off a blank page via annoying (and full screen!) popup ads.
Not too cool for attracting potential developers...
Ok, I understand. I don't think that Gnome is mixing it up though, because there is only the global application menu and the global action menu at the desktop menu while every application specific (non global) menu is at the application's window. This makes the most sense to me.
> For AtheOS, have you included all of the available third-party drivers?
Yes.
> OO != OS GUI Design
Yes, we know that, thank you for your insightful correction (NOT). This is not what I mean. I mean that MenuetOS does not have "classes" or objects or whatever, that you can easily *inherit* a universal look for all the apps. And this is the reason why all MenuetOS apps look so different and isolated from each other.
> "It is fully C++" What on earth is that supposed to mean? Because it's written in C++ it's good?
No, it means that it is written fully in C++. Not more, no less. AND CUT THE FREAKING SARCASM.
> Eugenia's nemesis, Pascal
What are you talking about? Pascal was the first ever language I used at college and on my first PC.
If you freaking going to use my name on this board, use it correctly. Not by talking BULLSH*T.
You might think that I am the "editor" of this site that has no feelings and that I can be a target that can not be hurt. NEWS BREAK. I am also a reader of this site, and most of all, a human being.
So be careful how you talk about me or other editors around here. I had enough of all this mindless trolling the past 3 years on BeNews and here. If you talk, talk *nicely*. Or don't talk at all.
> but this article is shoddy, inaccurate and rushed and gives no useful information what so ever.
For you, maybe. But for users who have never used theses OSes, this a good oveall synopsis of what these OSes too. And I don't find it innacurate, not one bit. I am not taking back not even 1 word I wrote. I meant everything I wrote.
Before you post over here, make sure you have read here:
http://www.osnews.com/rules.php
Eugenia,
Don't let them get you down. The reason people keep coming back to this site is the quality of the content. I know plenty of poeple will back me up on that.
I don't understand why this guy is insulting you. Hint: If you don't like the article, don't read it. Its that simple. Eugenia doesn't get paid for this, she does it out of interest.
Personally, I found it an interesting comparison of the top "hobby" os's out there. Personally I am a big syllable/atheos fan and have decided to help out with the coding and improving of it.
A comparison between AtheOS, Syllable, OpenBEOS, and Cosmoe (hopefully all will reach the point of an simply-installable release soon). I have high hopes that one or all of these will become the next great enduser OS.
OpenBeOS is not done at all. It does not even have a GUI, it is not usable. Cosmoe barely works and it is simply a port of the AtheOS GUI over the well known Linux. There is nothing to compare there. Not yet.
I know that OpenBEOS and Cosmoe are not ready yet. But I keep hoping they will be someday soon. If I could program worth a crap, I would help.
*Tries to pick up the pieces of alternate OS dreams after they have been dashed upon the rocks of reality*
*Picks up copy of Linux Administration for Dummies*
Until then, I guess I have to keep trying to find some flavor of Linux that doesn't make me pull my hair out every time I try to update things like KDE. I have read about Gentoo and Debian, but I am only a Linux dabbler, and can't refer to myself as an experienced Linux user (which seems to be required for those two distros). Mandrake and Red Hat are the only distros I've tried. I've successfully managed to destroy the functionality of Mandrake by trying said KDE update. I installed so many KDE dependencies by RPM that the built-in admin tools and even urpmi stopped working.
I am waiting for somebody to make a software installer for Linux that I can use, and that won't take days for software to compile. I've been spoiled by Windows in that respect (even though it's installers are far from perfect).
By the way, does AtheOS/Syllable have as simple of a software installer as BeOS had? If so, I may have my wish granted after all.
For those actually able to write the code for these OSes, THANK YOU!!!
A.N.
What OSX done is the best compared to KDE desktop menu or the same in Gnome. However, for me all this is just wasting the screen space especialy for those cannot afford to buy a large screeen monitor.
Regarding the Re: OO != OS GUI Design I think Daryl Dudey is right Eugenia. Just close your eyes to anybody that attack other person's personal. Well at least he/she just a user at your 'forum'.
> What are you talking about? Pascal was the first ever language I used at college and on my first PC.
Why all people think that there is impossible to write good operating system in pascal? I am now writing NeoDOS (http://neodos.from.lv) - M$-DOS compatible operating system in pascal. Also there is Bens Pascal OS (http://www.thegaineys.fsnet.co.uk/). And Pascal P-system was alternative of M$-DOS on first PCs.