NetNews Usenet Archive 1992 #20/NN_1992_20.iso/spool/comp/os/linux/10764

Internet Message Format | 1992-09-15 | 3.2 KB

Path: sparky!uunet!mcsun!news.funet.fi!hydra!klaava!wirzeniu
From: [e-mail] (Lars Wirzenius)
Newsgroups: comp.os.linux
Subject: Re: Can you access a virtual console directly?
Message-ID: <[e-mail]>
Date: 15 Sep 92 21:26:42 GMT
References: <[e-mail]> <[e-mail]> <[e-mail]>
Organization: University of Helsinki
Lines: 58

[e-mail] (Kenneth Falck) writes:
>You know, that might be one of the reasons MSDOS has made it so big;
>your applications can have very powerful output when you don't have
>to hassle with some control codes that really belong to the externally
>connected terminals, not to the console screen.

Perhaps they don't have to hassle with terminal control codes, but
instead they have to hassle with different device drivers for each and
every application.  Does it _really_ make any sense to have a few dozen
drivers for different video hardware for each major application.

And just because the console happens to have a memory mapped screen, it
doesn't make any sense to use a different protocol to talking to it and
to other output devices.  If you talk to the console differently than to
terminals, you need to write two versions of the program, which is
definitely Not Good.  Just forget it, it isn't worth it.  You don't buy
anything important by having direct access to video hardware, it only
makes it more difficult to have several programs that work nicely.

>How about if the kernel had some special services that would make it
>easy and portable (within Linuces) to use the virtual consoles without
>termcap? I mean something like this:

You solution has serious problems.  Among other things, it assumes that
the video memory is laid out in a special way.  Not everybody uses the
same kind of hardware you do.  If you use termcap, you just don't have
to worry about it.

>Does this seem like a stupid idea

Yes.  (Or did you guess that already? :)

>or would it bring some more power into the Linux-applications?

No.  I can't think of anything that you can do with poking around in
video memory directly that you can't do with a well designed set of
control codes and a working console terminal driver.  In fact, I assume
that the sets of possible operations can be shown to equivalent.

> I think it's a fact that 386 is a personal computer, and many people
>would like to use Linux as a multitasking OS only, and have no need to
>run programs remotely.

The users don't care how the programs have been written.  They only want
something that works.  If you write your programs so that they go
mucking with the intimate parts of the hardware, they are not going to
work as well with all hardware and software configurations.  It is the
job of the operating system (and other system software) to muck with the
intimate parts of the hardware, and to provide a convenient interface so
that the applications can use the capabilities of the hardware in a way
that is not directly dependent on the details of the hardware.

Also, the issue is not only that people are going to have problems with
external terminals, they are also going to have problems with running
the program in X xterm windows.

--
[e-mail]