Re: hypothetical question

Re: hypothetical question

[Ted Dennison]

In article <3b448a6d@news.iprimus.com.au>, Karl Fitz says...
>
>in my random thoughts as a 'newbie' programmer,
>I begin to wonder the possibillity of creating an Operating System entirely
>with ada. Or would it be better to use a variety of programming languages...
>many questions arise from this but it intruiges me,

Ada is a systems-programming language, and thus theoretcily well-suited for the
task of OS development. It has been done in the past. Outside of those two
statements, I don't have anything new to add that you couldn't find by going to
groups.google.com and looking through the many posts made on this same subject
in the last couple of weeks.

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
          home email - mailto:dennison@telepath.com

Re: hypothetical question

[Al Christians]

Larry Kilgallen wrote:
> 
> The cost of creating an operating system is extremely high - ask
> Microsoft or IBM.

But you should be able to do it for the odd $billion or less.  If 
you can add another $billion to develop a few desktop apps and come
up with a company that is in the top 10 worldwide in market value,
that's extremely cheap.  This is a very lucrative opportunity awaiting 
funding. Ada would be a good choice for a project like that.  

Al

Re: hypothetical question

[Larry Kilgallen]

In article <3b448a6d@news.iprimus.com.au>, "Karl Fitz" <vampyre_16@hotmail.com> writes:
> in my random thoughts as a 'newbie' programmer,
> I begin to wonder the possibillity of creating an Operating System entirely
> with ada. Or would it be better to use a variety of programming languages...
> many questions arise from this but it intruiges me,
> would it be more/less, stable (as a compasion to other Os) ?
> ''           ''              ''       , resource hungry.... ?
> Im sure that it is entirely possible but i guess the real question is ....
> would it be practical ?

It would certainly be practical, but the initial creation would not
be any cheaper than creating a new operating system in _any_ language.
The cost of creating an operating system is extremely high - ask
Microsoft or IBM.

Re: hypothetical question

[chris.danx]

> in my random thoughts as a 'newbie' programmer,
> I begin to wonder the possibillity of creating an Operating System entirely
> with ada. Or would it be better to use a variety of programming languages...
> many questions arise from this but it intruiges me,
> would it be more/less, stable (as a compasion to other Os) ?

Stability is dependant on design, the language used only affects predictability
and correctness which do effect stability.  Ada has strong typing, and a lot of
compile time checks, so this would improve predictability and correctness but
only to a point.  So Ada does improve stability but will only succeed with good
design.

Also you can't just do an OS in Ada, at some point there has to be some Asm
though it may just be one or two lines to invoke task switching etc.


> ''           ''              ''       , resource hungry.... ?

T o me this isn't really a language issue, it's an implementation one. This is
two fold, the implementation of the compiler and the implementation of the OS.
It does you no good having a compiler that has efficient resource management if
your OS design is resource hungry.  Careful design is necessary to avoid a
resource hungry OS, which isn't easy.  (As far as I know most Ada compilers are
quite good with resource management).


> Im sure that it is entirely possible but i guess the real question is ....
> would it be practical ?

Of course it's practical (if you allow for a little asm, if not then NO!).  It's
just difficult, and it will take a long time.


> I also ponder this... is there one 'better' programming language...
> and how can you compare them.....

It's a personal thing, isn't it?  To me Ada is really good for some things but
falls down when it comes to others.  It's the right tool for the job philosophy.
So no, there isn't "one 'better' programming language", there is only one for a
particular developer for a particular job.

example:
    Haskell/OCaml    -- Compilers (in FP compilers are simpler
                                   to write, or so i'm told)
    Ada              -- App programming/OS work



Chris

Re: hypothetical question

[Robert Dewar]

"chris.danx" <chris.danx@ntlworld.com> wrote in message news:<TW117.4820$Jk6.1116376@news2-win.server.ntlworld.com>...
> Also you can't just do an OS in Ada, at some point there has to be 
> some Asm though it may just be one or two lines to invoke task 
> switching etc.

nope, once you have your low level Ada compiler bootstrapped then
you can do the operating system task switching entirely in terms
of Ada task switching. Thinking you have to use ASM to do task 
switching is exactly like thinking you have to use ASM to do a
hardware addition. Yes, there is a bootstrap problem, but given
Ada compilers on other machines, that is easily solvable.

Re: hypothetical question

[B. Douglas Hilton]

Hey, don't overlook Hurd! Its a microkernel + multiserver OS like QNX
which currently uses the gnuMach microkernel. Functionally it is very
similar to older versions of Debian GNU/Linux.
    My goal is to replace gnuMach with a compatible microkernel written
in Ada, use it to boot the Hurd, and then replace the C-Hurd with Ada-Hurd translators
at my leisure. The end result would be a rock solid OS that is binary compatible
with all the Hurd packages written in whatever language they are written in.
It would be a bad idea to get rid of the gcc libc because so much great old
software is made for it and c hasn't died yet.
    There's no point in re-writing an OS because GNU has all the features you'd
ever want or need. All you need is an Ada microkernel to start the migration.
Use GNU as your OS, just make it work with an Ada kernel and we will blow
away the Windoze world in no time. Imagine how cool it would be if the Linux
kernel was written in Ada.... but the Hurd is even cooler still!
    What Linux did ( and we should be thankful ) is bring the GNU OS to the
people. There is no learning curve from a user's perspective from going from
GNU/Linux to GNU/Hurd because it is the same set of UNIX-like programs
that we use. It has BASH, ls, cp, mv, ln, startx, etc. Use Linux to learn GNU,
then help GNU grow away from just one kernel. In some ways, Linus sort of
inherited a full-blown UNIX OS when he wrote his kernel, but keep in mind
that it was thousands of university students that wrote the GNU operating
system, and those c-programs are not the same thing as the Linux kernel.

I don't mean to be ripping on Linux, because I'm using a state of the art Linux
system to post this message, but I personally can't understand the kernel because
it has grown so huge, and I always like to promote worthy new projects to keep
the competition fierce!

Cheers,
- Doug


Larry Kilgallen wrote:

> In article <3b448a6d@news.iprimus.com.au>, "Karl Fitz" <vampyre_16@hotmail.com> writes:
> > in my random thoughts as a 'newbie' programmer,
> > I begin to wonder the possibillity of creating an Operating System entirely
> > with ada. Or would it be better to use a variety of programming languages...
> > many questions arise from this but it intruiges me,
> > would it be more/less, stable (as a compasion to other Os) ?
> > ''           ''              ''       , resource hungry.... ?
> > Im sure that it is entirely possible but i guess the real question is ....
> > would it be practical ?
>
> It would certainly be practical, but the initial creation would not
> be any cheaper than creating a new operating system in _any_ language.
> The cost of creating an operating system is extremely high - ask
> Microsoft or IBM.

Re: hypothetical question

[chris.danx]

> > Also you can't just do an OS in Ada, at some point there has to be 
> > some Asm though it may just be one or two lines to invoke task 
> > switching etc.
> 
> nope, once you have your low level Ada compiler bootstrapped then
> you can do the operating system task switching entirely in terms
> of Ada task switching. Thinking you have to use ASM to do task 
> switching is exactly like thinking you have to use ASM to do a
> hardware addition. Yes, there is a bootstrap problem, but given
> Ada compilers on other machines, that is easily solvable.

I didn't know that, thanks.


Chris

Re: hypothetical question

[Ted Dennison]

In article <3B450EE0.8FC6D60D@engineer.com>, B. Douglas Hilton says...
>
>    There's no point in re-writing an OS because GNU has all the features you'd
>ever want or need. All you need is an Ada microkernel to start the migration.

Well, it doesn't support real-time.

There is supposedly a RT-MACH out there. It'd be interesting to see what could
be done with that.

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
          home email - mailto:dennison@telepath.com

Re: hypothetical question

[Bobby D. Bryant]

In article <cScfJwUl$Roy@eisner.encompasserve.org>, "Larry Kilgallen"
<Kilgallen@eisner.decus.org.nospam> wrote:


> It would certainly be practical, but the initial creation would not be
> any cheaper than creating a new operating system in _any_ language.

However, we'd all like to think that the even-more-expensive task of
*maintaining* an OS would be cheaper if the OS were developed in Ada.

Bobby Bryant
Austin, Texas

Re: hypothetical question

[Larry Kilgallen]

In article <9i6lkf$bqi$2@geraldo.cc.utexas.edu>, "Bobby D. Bryant" <bdbryant@mail.utexas.edu> writes:
> In article <cScfJwUl$Roy@eisner.encompasserve.org>, "Larry Kilgallen"
> <Kilgallen@eisner.decus.org.nospam> wrote:
> 
> 
>> It would certainly be practical, but the initial creation would not be
>> any cheaper than creating a new operating system in _any_ language.
> 
> However, we'd all like to think that the even-more-expensive task of
> *maintaining* an OS would be cheaper if the OS were developed in Ada.

But the tough part is getting to the start of the maintenance phase :-)