Re: Last Call: 'Proposed Experiment: Normative Formatin AdditiontoASCII

On Fri, Jun 16, 2006 at 10:24:26PM +0200, Iljitsch van Beijnum wrote:
> On 16-jun-2006, at 19:25, Lyndon Nerenberg wrote:
>
> > > As far as I know, support for SVG or _any_ vector image format is  
> > > much, much less common than for bitmap formats such as PNG or GIF.
> > Yes, but SVG is catching up rapidly.  As a W3C standard, it *will*  
> > be widely implemented.
> We'll have to wait for that before we could use it. And I don't  
> believe for a second that it will come close to bitmap formats.
SVG is currently exportable into most major bitmap formats.  
http://librsvg.sourceforge.net/ the converter is rsvg.  Runs out of the
box on most free unices.

Of course, you can't go back from the bitmap to the vector, which is
exactly the point.  Once you've sampled the image to put it in a bitmap
you've lost information.  That's why we're advocating starting from a
maximum information remresentation.  (Yes it's only maxmimum information
for diagrams and the like, but that's what appear in RFCs.)

> > But we are describing an *archival* format.  It's not important  
> > that they be editable,
> Yes, it is.
It's useful, but the primary issue is to be able to view them.  That's
what archival means.

Even if you care about editing, it really depends on what kind of
editing you want to do.  Edge detection is not so easy in a vector
representation, but selecting a box or hunk of text is much easier.  I
think I'm much more likely to do the latter on an RFC document than the
former.


> > it's important that they be renderable on the widest range of  
> > output devices as is possible.
> And you think vector graphics fit that requirement better than bitmap  
> graphics???
Any time you scale a bitmap to a new resolution you lose information.
This is not true of a vector representation.

At the very worst, I can render an SVG (or a pic diagram - choose your
favorite vector representation) into a PNG and display it however you
would display the PNG you want to store.  But the SVG user can pick the
resolution of their PNG to match the output device and that PNG will be
undistorted beyond the natural sampling limits.  An SVG user always has
the highest possible resolution bitmap available as well as the
thumbnails free from scaling artifacts.

No matter what resolution or encoding future devices use, a vector
representation starts from a very high (if not maximal) amount of
information about the image.  That high content description can be
sampled into a raster at whatever resolution or color model is
approriate.  If you start from the sampled raster you will never be able
to generate a more detailed image free from artifacts.

Breathe deeply and think about that.  That's the point, not how many
browsers will render SVGs today, or whether the gimp or photoshop likes
them.

-- 
Ted Faber
http://www.isi.edu/~faber           PGP: http://www.isi.edu/~faber/pubkeys.asc
Unexpected attachment on this mail? See http://www.isi.edu/~faber/FAQ.html#SIG

pgpYSfdHOUIba.pgp
Description: PGP signature

_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf