[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Sun, 28 May 2000 11:04:19 +0100 (BST)
> into some kind of OCR program to recover text from these vectorized
> line-drawings *of* text, which in turn would have to be plugged
Like PDF, SVG uses scaleable fonts, so one would expect to find actual
characters. The problem is that, in many cases, people will place words
and characters individually, rather than have a sensible reading order.
The OCR-like problem isn't recognizing the characters, but re-constructing
the reading order, and possibly spacing.
I should also point out that vector graphics are legitimate when the
alternative would otherwise have been a GIF of a line drawing. Such
GIFs don't scale and often don't reproduce text clearly.
There is, in my view, a niche for a high level drawing package, like Unix
PIC, in which one states the relationships between elements but lets the
software do the detailed layout. However, I don't believe that would
cover all of even the legitimate uses of vector graphics on the web.
When I put this on the SVG list, the suggestion was to use powerful
style sheet languages (another problem for Lynx) to re-arrange a
user friendly description into SVG. I don't think this would work, because,
without a standard, you could only recover the logic by reading the page
Incidentally, I'd advise against direct campaigning on the mailing list;
I think that would just drive people into a corner - the developers are
graphics people and the consumers are flash web site producers. I think
the only approach is for anyone actually within W3C to bring pressure to
bear from behind the scenes.
; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden