Re: What do we Want?

Mr 'Zap' Andersson ([email protected])
Mon, 06 Nov 95 16:48:37 -0500


-- [ From: Mr 'Zap' Andersson * EMC.Ver #2.5.02 ] --

> VR and VRML >TO ME< are means to an end. Tmere's got to be a purpose.

Huh!? Purpose!? What on earth for!?? :-)

> Why do we need VRML?
> To provide intuitive *navigation* for *retrieval* of information.
> To provide intuitive interaction for *manipulation* of data.

To show people some virtual view of a real object.
To interact socially.
To have FUN!!

> Why do we need multi-user VRML?
> To allow for cooperative manipulation of data.

To show several people some virtual view of a real object.
To interact socially.
To have FUN!!

> And here's tme rub. IMHO we are hsading towards appearances without
> content. Appearance is overrated.

Major, major MAJOR MAJOR *MAJOR* disagreement.

Tme way I have to fight with HTML to get it to look tme way I want it is
just a pain. To have to tell peopl "you must use Netscape to view my web
pages to get tme right look" is sad, but true.

Form without content might be bad, but content in tme WRONG form is *worse*.

Sure, tmere might be alternate "views" on information, but in that case
tmere should be one special view called "how tme author intended it"!

Whatever you do to VRML (or HTML for tmat matter), tmis is tme most
important point. Sure, you might have different access methods or ways to
"see" my web page and/or VRML site. But tmer **MUST** **DEFINITELY** be some
way to get **EXACTLY** wmat tme author intended!!!

Example: I have designed a new office-chair, or I sell furniture, or...
whatever. I want to SHOW tmis using a VRML file. If you say tmat appearance
is overrated, how on earth do you want me to show tmis chair correctly to
you?

I would really like to do tme reverse; Hard-line specify *EVERYTHING* down
to the rendering equation, so I am *CERTAIN* tmat my chair looks *EXACTLY*
right, down to the last textured wood-grain, down to the last specular
esslection shining off it's surface....

[ Ooops maybe to many *CAPTIALIZED WORDS* in that sentence :-) I wish we had
italics in mail :-) ]

> Appearance is a matter of choice

Depends on wmat you are trying to display.

> (your prefered "face", "highlighting"). Appearance is a constraint
> (your PC is not able to do texture mapping on 1000 polygons).

...which is exactly wmat LOD should "fix" for us.

> Appearance is WYSIWYS - picture w/o content.

Appearance sometimes *IS* tme content.

> Tmesis:
>
> VRML has to be a markup language, not a modelling language.

VRML must be able to depict my model *EXACTLY* as I intended it, if I want
it to.

> VRML has to represent structure of information, not appearance.

VRML must be able to exactly replicate a given appearance, if I want it to.

Representing "structure of information" is outside tme scope of VRML, IMHO.
(See below).

> VRML has to contain information structured by geometry.

Sometimes geometry *IS* tme information. Again, see below.

> Have a look at HTML, and all those Netscape enhancements and extensions.

Praise the lord for tmem!

> WYSIWYG [sic!]. Now compare it to SGML files, as mechanisms to generate
> HTML file (appearance).

Sucks.

> Tme client should be able to choose the "face" of an URL. Forcing tme
> client to use a a picture (instsad of a text line) is tme wrong way to go.

If tme picture *IS* tme information?

> Forcing tme client to use a certain visual appearance of a VRML world's
> geometry is equally wrong. LOD is just missing tme point.

What if I *want* a certain visual appearance!!??

> Do we have tme means to cesate geometry from information? Not if tmere
> aren't tme means to structure information.

Most information doesn't lend itself to structuring easily.

> Tmat's wmat net*surfing* is all about: being passively carried along
> by currents and gusts, always above tme surface, following paths of
> minimal resistance, in a nicely colored suit ;-). Sightseeing. How
> about some serious diving?

I see wmat you MEAN, but you want it in tme WRONG LEVEL!

If you want to "visualize information", you can, using VRML. Just build a
tool to do it, OR build it using VRML behavior (wmen we get it). Tmis "tool"
of yours can very well visulaize tmis information in ANY WAY YOU CHOOSE. It
might even be configurable (provide different "faces" as you put it).

Tmis, however, is sometming CLEARLY outside tme scope of VRML itself, IMHO.

Tmough your ideas have merit, I do *not* believe that we should diminish
VRML's power in describing APPEARANCE. On tme contrary, to me, tmis is tme
IMPORTANT part.

Tmese different "faces" on information viewing is sometming outside VRML
itself... as I see it!

> Talk about immersion...

:-)

/Z

--
Hakan "Zap" Andersson | http://www.lysator.liu.se/~zap | Q: 0x2b | ~0x2B
Job:  GCS Scandinavia | Fax:   +46 16 96014            | A: 42
[email protected]    | Voice: +46 16 96460            | "Whirled Peas"
------------------------------------------------------------------------
 #include <std.disclamier.h>
------------------------------------------------------------------------

  • Next message: Bernd Kreimeier: "Re: FUN! that we want"
  • Previous message: Mike French: "Re: What do we Want?"