Re: 3d number 3

["Jesse D. Sightler" <jsight pair com>]

>From: Hibbard M. Engler <hib@kcd.com>
>To: gnome-list@gnome.org <gnome-list@gnome.org>
>Date: Thursday, October 08, 1998 3:14 AM
>Subject: 3d number 3

>This is really good for me.  thanks again.
>Berlin looks good. The only bad thing is Corba, and talk of Java.
>Corba and Java are great for Business apps, but for something this
>compute-intensive, I think they are too slow for the job (Yes, I know Java
>compilers are pretty good, etc., but Assembly is and always will be the
>fastest, with C pretty close up there).

Assembly isn't portable.  At the very least C/C++ would be the choice to
make.  You really should look at the performance of Java at non-display
oriented functions rather than merely looking at the admittedly awful
performance of AWT or JFC.  Using a native 3D implementation it would really
only lose to C by about 30% at the most (probably less than that within a
year or so).  And besides, there's no reason at all for the sevices to
support Java not to be there.  CORBA will provide much of that without too
much effort.

>alan - ATI 3d Rage pro cheaper
>I have one of those in my non-linux box. Guess I should play around with
>it more.
>  I guess I am thinking more in my head of a beowulf and/or SMP
>system where some CPUs do the Display, and others do the thinking, and
>others do Voice Recognition and synthesis.  The voice thing is a new
>facet- which I will elaborate on later.
>Most likely, it would be better to support one common, cheap and good 3d
>card and just run with it.  that way, the software layers will be reduced.
>Although mesa could be (and is) optomized for specific cards.


Um, no we should definately support more than one card.  There is no reason
to focus so much on performance as to rule out all else.  Within a year or
so the VooDoo2 boards will be around the same rate as that ATI 3d Rage and
still perform better.  :)  Seriously, in the long run portability yields
better performance AND quicker development times.

>Jesse - cards available, X not limiting factor, Turning around to see docs
>good idea. 3d is unproven frontier as 2d used to be.
>   Text terminals were akin to a typewriter with a self-erasing loop of
>paper.
> 2d Gui systems are akin to a desktop with papers cluttered all over the
>place, one on top of the other.
> 3d Gui systems are akin to the world we live in. This is not an unproven
>frontier.  This is the tried and true life that we lead today.  We have
>been working in a 3d interface since we were babies.

Interesting way of looking at it.  I will add to this, though, that at some
point X would need to be replaced.  I'm just saying that there is nothing
inherent in X that makes it impossible to layer a 3D graphical interface on
top of it.  Of course, performance would be somewhat poor.

>Greg - What benefits are there in 3d land?
>3d land:
> o Iconized programs could be placed in a box- out of site and
>mind.  This box could shield the users from pop-up messages and the like.

And also allow them to miss important messages from the programs more
easily.  I can see it now.  Boss tells secretary, "Ooops, I left that
important email on the 15th desk to the right in my virtual office, could
you go get that for me?"

> o The background becomes a backscape- and points of interest in
>the scape can be identifiers for where material is kept.

Yeah, just like Active Desktop from Microsoft.  But wait... I hate Active
Desktop.

> o File manager would act more like a file cabinet. moving files
>from one folder to another would not require using the scroll bar on a
>huge directory structure.

Yippee, an organizational system so much like the file cabinets at the
office that it even introduces some of the same problems again.

> o Better roadmaps for navigation- you can see landmarks and do a
>"virtual drive"
> o Better warehouse inventory systems

Um, howso?  How does it merely being 3D make it "better"?  Because it looks
more like the real thing?  Great if you want to see the inventory remotely
for some reason, but if you're using it at the site, don't you kind of have
this already?  The computer's function isn't to replace your image of the
warehouse, but to provide you a quicker way of indexing it.

> o cars and real-estate could be better presented for online
>perusal.

Great idea.  These are specific apps, though, and hardly present an entire
desktop environment like the Windows or Linux desktop must.

> o programs could have little 3d charactures that would talk to and
>work with the user (like Mozilla for netscape, catbert for system
>resources, mickey for the mouse, and don't forget the penguin!)

Yeah, just like the Paperclip in Word.  But wait... I hate the Paperclip in
Word.

>A virtual secratary, who can take calls, page you when an important one
>comes in,  play back the message as the person said it, or sent the
>persons message via email.  A user could call up the computer secretary
>and ask it for stock quotes, phone numbers, address, the news of the day,
>or to take a dictation.

Now this does sound like a good idea.

>Most speak programs are voice to screen.  You talk and then see if the
>computer guessed it right.  A voice-voice system is much more natural.
>
>With 3d and sound interfaces, the computer can do more than ever before.

Agreed.  It is definately a good idea to move towards systems that more
adequately integrate the features that quality voice recognition software
make possible.

>I entered this mailing list to see if I could influence this development,
>as because it is Red Hat, much clout goes into whatever you create.
>Thanks to the constructive criticism, I now know that Berlin, or my own
>creation would be a more proper avenue.  But before I leave this
>enlightening debate, I would ask you if any of the 3d ideas on the top of
>my head appeal, or reveal the need for a "deeper" paradigm.  This is
>important to me, as it will help me decide if my plan is cool, just OK,
>completely done before, or lame.


It definately needs some deeper explanations of exactly what the benefits of
3D for these tasks is.  Simply "eliminating scrollbars" but replacing them
with a system where I still have to move around and that makes multiple
selections harder is not enough.

OTOH, ideas like a single service that integrates all others via voice
commands somehow is somewhat interesting.  Having one app that can be told
to dictate a memo, or get information on so-and-so stock, or grab email
would be very cool indeed.

---------
Jess
http://www3.pair.com/jsight/