[rfc-i] draft-rfc-image-files-03

Iljitsch van Beijnum iljitsch at muada.com
Sun Apr 8 14:49:18 PDT 2012


On 8 Apr 2012, at 23:30 , Larry Masinter wrote:

> for document reviewers and users (who want support on a wide variety of devices with a range of screen sizes) vector art (as supported by pdf and svg) is more legible than pixel art (as png, jpeg, etc)

What do you base that on?

To my knowledge, there are no display devices in common use that natively display vector images. So vector images need to be rasterized into bitmaps at some point anyway, with a somewhat unpredictable loss in quality as a result, unless the resolution of the target bitmap is very high.

With a relatively low resolution bitmap the author needs to make tradeoffs explicitly rather than depend on high resolutions to sweep any problems under the rug.

[bitmap]
> images can be rendered from vector art but not vice versa reliably.

Which is a good reason to avoid vector formats in the first place, making this particular issue moot.

> device viewing capabilities are changing rapidly... do not design the archive format to optimize for a small subset of current capabilities.

Remember that we're not defining a general purpose file format that needs to support all the capabilities of current and future devices. What we need is a file format that allows us to express internet related standards in a clear and future-proof (and to some degree past-proof) way. If that means that we're stuck with images that are of a lower resolution than current and certainly future screens, I don't see any issue with that if the tradeoff is that we can unambiguously determine if images are useful (3 pt text in a 72 DPI image is unreadable, 3 pt text on an iPad 3 is also unreadable but it's displayable so some people may choose to use it) and edit them with extremely simple (to the point of being homegrown) tools.

But as I've said several times now, I don't see why the authoritative RFCs need images in the first place. Yes, diagrams can be helpful, but if the text and the diagrams disagree the text needs to take precedence and I certainly can't imagine RFCs with images that stand alone with no text describing the same thing.


More information about the rfc-interest mailing list