[unios] A monkey wrench....

Faustino E Osuna osuna@CTC.Net
Mon, 28 Dec 1998 16:18:58 -0500


From: Faustino E Osuna <osuna@CTC.Net>

This is a multi-part message in MIME format.

------=_NextPart_000_001C_01BE327D.C5FC4E20
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

I know this throws a monkey wrench into the brainstorming process, but =
has it occurred to anyone of creating multiple kernels? Maybe even =
dynamically loadable kernels (DLK)?  Where once the OS starts up, it =
fires up the UniKernel (default kernel) then according to the needs of =
the individual application, it would load the associated kernel?  Or the =
user (not the application) has the ability to unload and load a kernel =
on the fly.  Or give the user the option of starting up with something =
other than the "UniKernel", a kernel specifically designed for their =
overall wants and demands.

I would think the latter suggestion would be more feasible.  Having =
DLK's being loaded and unloaded every time the application runs would be =
somewhat redundant.  Instead, there should be multiple kernel =
architectures that allow the user to reap the most benefits from his/her =
system.  Of course, one major issue to this system would be the =
interface between the kernel and the application.  There must be a =
standardization (ugh I hate standards).

A good example of this would be, that a business might want to dedicate =
UniOS system entirely for networking purposes (lets say a print server). =
 They load the appropriate kernel that allows the application(s) and the =
network to benefit the most from the system (High-security and moderate =
speed).  But on the other hand, we a home-user (like the most of us) who =
wants UniOS' power as well but the security doesn't have to be as high =
but the speed is a must or flexibility.

In my novice opinion, this would make UniOS more diverse and flexible.  =
But it just might defeat the entire purpose.

Just an idea to chew on, I apologize for not elaborating too clearly.  =
Of course ideas, suggestions, gripes...etc are welcomed.

- Enrique

PS - I haven't formally introduced myself, I'm Enrique expert slacker.

------=_NextPart_000_001C_01BE327D.C5FC4E20
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD W3 HTML//EN">






I know this throws a monkey wrench = into the=20 brainstorming process, but has it occurred to anyone of creating = multiple=20 kernels? Maybe even dynamically loadable kernels (DLK)?  Where once = the OS=20 starts up, it fires up the UniKernel (default kernel) then according to = the=20 needs of the individual application, it would load the associated = kernel? =20 Or the user (not the application) has the ability to unload and load a = kernel on=20 the fly.  Or give the user the option of starting up with something = other=20 than the "UniKernel", a kernel specifically designed for their = overall=20 wants and demands.
 
I would think the latter suggestion = would be=20 more feasible.  Having DLK's being loaded and unloaded every time = the=20 application runs would be somewhat redundant.  Instead, there = should be=20 multiple kernel architectures that allow the user to reap the most = benefits from=20 his/her system.  Of course, one major issue to this system would be = the=20 interface between the kernel and the application.  There must be a=20 standardization (ugh I hate standards).
 
A good example of this would be, = that a business=20 might want to dedicate UniOS system entirely for networking purposes = (lets say a=20 print server).  They load the appropriate kernel that allows the=20 application(s) and the network to benefit the most from the system=20 (High-security and moderate speed).  But on the other hand, we a = home-user=20 (like the most of us) who wants UniOS' power as well but the security = doesn't=20 have to be as high but the speed is a must or flexibility.
 
In my novice opinion, this would = make UniOS more=20 diverse and flexible.  But it just might defeat the entire=20 purpose.
 
Just an idea to chew on, I apologize = for not=20 elaborating too clearly.  Of course ideas, suggestions, = gripes...etc are=20 welcomed.
 
- Enrique
 
PS - I haven't formally introduced myself, I'm = Enrique expert=20 slacker.
------=_NextPart_000_001C_01BE327D.C5FC4E20--