Actually, BLINK as implemented by Netscape is a wholly useless construct for
inclusion in HTML 3.0, a ** structural ** markup language. We need to provide
precise mechanisms to indicate the author's meaning in a variety of media:
GUIs, text terminals, audio over the phone, even Braille.
In its place, we offer the general purpose <FADE> tag:
<FADE FREQ=integer_hertz PHASE=integer_degrees WAVE=type>
Where acceptable WAVE types are SQUARE, SINE, and TRIANGLE
<FADE> may contain: any element in <BODY>, as well as <BODY>
Suggested GUI rendering:
------------------------
All FADEd elements shall sample from a single time source, and the FADE value
shall be considered an opacity ratio for compositing ONLY the included
comments against the browser background
Suggested Text rendering:
-------------------------
The same rules apply as above, though there are alternate suggested
mechanisms for FADEing. While 1.0 and 0.0 are conveniently represented on text
terminals, intermediate values of obscuration with background may be achieved
by subtracting 127*FADE from the character code of the displayed value. More
thorough implementations may wish to use MD5 [RFC1321] to obscure the data, by
using successive values of
MD5(rendered text | FADE, in IEEE 80 bit float packing)
Rest assured that MD5 can be used for any reasonable frequency on even the
most meager hardware [RFC1820, "Report on MD5 Performance", J. Touch, ISI]
Suggested Audio Rendering:
--------------------------
The obvious mapping seems to be FADE <--> volume, with linear scaling.
Unfortunately, the serial nature of this media prevents users from perceiving
any of the author's intended beat-frequency effects between FADEd sections.
Suggested Braille Rendering:
----------------------------
Unfortunately, the current generation of Braille rendering hardware does not
afford the programmer control of pin height, conventionally restricted to up
(1.0) and down (0.0). Until variable-pin-height devices become widely
available, it is acceptable to round FADE values, with a value of .50 being
rounded up and down in accordance with the output of some cryptographically
random source (preferably a coupled oscillator or, better yet, a cosmic ray
detector).
Additional GUI Considerations:
------------------------------
Client developers are strongly encouraged to access native timing and
even-export services, in order for author to interface reliably with 3D
Shutter goggles, in order to properly render 3D Magic Eye (TM) content.
Additional Parsing Considerations:
----------------------------------
While it is technically correct to include an entire HTML3 BODY element
within a FADE container, when doing so care must be paid to nested FADE
elements. Nested FADEs oscillate to the same clock, but composite against the
parent FADE element, so beat effects can indeed arise.
Content Provider Considerations:
--------------------------------
Rather than placing an entire BODY in a FADE, authors may with to generate a
wider range of page-wide effects with the DISSOLVE element (which includes
WIPE, RANDOM, and IRIS).
So, how about it, Omni gang? Care to blow the competition out of the water?
Rohit Khare
PS: :-) :-) :-)
Begin forwarded message:
Date: Sat, 10 Feb 96 20:01:54 -0800
Reply-To: omniweb-l@omnigroup.com
Sender: omniweb-l@omnigroup.com
From: Scott Anguish <sanguish@digifix.com>
To: Multiple recipients of list <omniweb-l@omnigroup.com>
Subject: Re: Actually, OW is missing a crucial Netscape feature...
X-Listprocessor-Version: 6.0 -- ListProcessor by Anastasios Kotsikonas
X-Comment: OmniWeb Discussion
>
> In its continuing quest to superset the functionality of
NS2.0b3Mac, OW2 has
> pulled off some amazing maneuvers, including the extra-special
PNG interlacing
> renderer I 'found' today.
>
> But it doesn't do <BLINK>. Such a shame, to come so close, and
not far enough.
>
> Sadly, we shall have to discontinue all use of OmniWeb at our
organzation...
>
> Sincerely,
> Rohit Khare
>
> (A :-) for the archives)
>
Well if you people down at w3 would get off your thumbs and add
<BLINK> to the HTML 3.0 spec...
:-)