Unstable releases!

[sunshine]

ok, correct me if I'm wrong but aren't odd numbered releases considered unstable releases? Not the major release numbered but the minor and alpha numbers. The odd numbers denote an unstable release or code change, or unconfirmed bugfix.

Isn't it true than that the very first version of windows released to the public was unstable?(v3.1) and doesn't that make most IE releases unstable as well (IE , 3.3, 5.5) But it looks like Microsoft has been hiding their alpha level patches as "Service Packs" minor or major Kernel Patches(eg win NT4 sp6 = win nt5 or win2k) not to mention the latest windows release Win 5.1 ...

damn those foo-mongers....  :eek:

[badkarma]

I don't know if you ever noticed this ... but *every* release of windows/IE is unstable  :rolleyes:

seriously ... version numbering is no indication on wether it's a stable release. Usually you can take the following rules as a guideline (but they are definatly not set in stone)

major version changes (i.e. 1.0 -> 2.0) indicate feature enhancements to a particular program

minor version numbers (i.e. 1.0 -> 1.1) indicate bug fixes

some companies use a third decimal to show minor bug fixes (i.e. 1.0->1.1 fixes a major showstopping bug, 1.0->1.0.1 fixes minor bugs)

but like I said ... this is not set in stone, and some companies assign version numbers at random (or so it seems)

[saddlemagic]

Odd numbered releases are unstable... That's a new one on me. Here's one for ya: "If you stand on your left foot when you sneeze, you will not catch a cold."

The release number has nothing to do with the stability of the software. It is just a number.

And, "Windows 1.0" (as far as I know) was the first public release of Windows. I'm not sure the whole history and changes but, it was followed by "Win286" for the new 80286 processor, and "Win386" for the 80386. Then "Win3.0" was released, hearlded as the best OS you will ever see, then "Win3.1" came out as a bug fix.

Then M$ promoted "Chicago" for the next couple years, after that they changed the name to Win93, the next year it changed to Win94, then the successor to win3.1 finally came out. After more than four years of development it should have been perfect. "Win95" the triumph of M$ over.. itself?

Oops! I'm talkin' too much drivel.

[badkarma]

hmmm ... after some digging I encountered the Free Software Project Management Howto which covers choosing a method of version numbering

 

quote:
Linux kernel version numbering:

The Linux kernel uses a versioning system where any odd minor version number refers to an development or testing release and any even minor version number refers to a stable version. Think about it for a second. Under this system, 2.1 and 2.3 kernels were and always will be development or testing kernels and 2.0, 2.2. and 2.4 kernels are all production code with a higher degree of stability and more testing.
Whether you plan on having a split development model (as described in Section 3.3) or only one version released at a time, my experience with several free software projects and with the Debian project has taught me that use of Linux's version numbering system is worth taking into consideration. In Debian, all minor versions are stable distributions (2.0, 2.1, etc). However, many people assume that 2.1 is an unstable or development version and continue to use an older version until they get so frustrated with the lack of development progress that they complain and figure the system out. If you never release an odd minor version but only release even ones, nobody is hurt, and less people are confused. It's an idea worth taking into consideration.

so you weren't entirely incorrect, but it's not a widespread standard....

edit : quoting is soooooo difficult ... *sigh*

[ July 24, 2002: Message edited by: BadKarma ]

[beltorak0]

That's a good recap of the history of it, saddle; where did you dig that up?
Do you know when / how Win286 became Win386?  Two M$ coders at an M$ party decided to try to place the win (?)kernel(?) in protected mode and "_"blow away the 1 MB barier"_"... so they used a mini-disassembler / debugger and stepped thru the code, one instruction at a time, hacking away at each General Protection Fault.... one, by, one.  And you wonder why it's so unstable?  It hasen't had a re-write since the begining (I'm guessing); just bits of code used over and over.... that's probably the real reason M$ won't release the code -- too much code that survived from mid 1980's....

-t.

[saddlemagic]

Windows 386 was the first Windows I ever bought. I tried running it on my 386DX20/MS-DOS 4.01 and it would just not work, couldn't even start it up.

Thank you, I didn't have to dig that up, I remembered it. I have this horrible problem, I remember everything. I mean "Everything!" It would be nice to forget a few things every now and then.

[Master of Reality]

I have never herad of win386. I have only heard of win 1.X/2.X/3.X/95/98/ME/NT/2000/XP

[saddlemagic]

Windows 386 was a modified version of 2.x that was optimized to take advantage of the 80386 processor.  That way it could crash in record time. I believe it was WinME that broke it's record.

Likewise, Windows 286 was optimized for the 80286. I'm not sure what it was, my best guess is just plain 1.x with a different name.

[Master of Reality]

i had win98 crash within 1 second of typing in 'setup' to install it. I am scared of what ME might do.

[lazygamer]

So does this all mean that Windows XP has tons of code(more like cold) in it that is over 13-15 years old?

[badkarma]

quote:Originally posted by lazygamer:
So does this all mean that Windows XP has tons of code(more like cold) in it that is over 13-15 years old?

probably even older seeing it's based on 2k which was based on NT which was based on some UNIX version which was probably based on something else ....

[lazygamer]

WOW that's amazing! But surely ancient code can't all be bad. Linux is based on Unix, so that means that Linux would have code in it that dates back to the 70's. But that code it has is probably code that you don't mind having next to your modern cutting edge code.

After all, stuff that's designed to run on REALLY low power CPU's and RAM must be pretty optimized.

[voidmain]

Well, not really. The "age" of code doesn't really mean anything. If you use a "printf()" function, that function as been around since the beginning of 'C' so does that mean your code is really old?  No.  And if a snippet of code is written properly in the first place, you would write it the same today as you did 10 years ago.

Now when Linux came about all of the source was written from scratch but it was written in such a way to function similar to and be compatible with UNIX. On the kernel side once you have a driver written for a 3Com 3C503 card, theoretically that driver should never have to be rewritten unless there was a problem. Now for every new piece of hardware that comes out a new driver has to be written. And some/all of the old drivers may have to be rewritten if there is a complete overhaul of the kernel (when the Linux kernel went modular for instance).

On the non-kernel side of Linux, if you wrote your own IMAP server today for instance, you would have to write it to function in a specific way according to the IMAP specifications. You could have two different sets of code that perform the same functions. The "ls" command in Linux functions very much the same as the "ls" command on old UNIX systems, plus some added functionality.  The code for that command was written from scratch but it performs similar functions to the original version and is compatible with it. But unless you need to add functionality there is little need for changing the code.

The gcc compiler (thing that turns that source code into machine code capable of running on whatever processor you are running) is continually evolving for many reasons. To compile the source code more optimally. To add new and changing processor types.  To fix bugs. etc.  So every time you compile source code with a new compiler, even if it is the same source code you basically have different object/machine code.  Would that be considered old code or new code?