Hacker News new | ask | show | jobs
by chaganated 1969 days ago
Starting off, SVG is an XML encoding, which is a rather hairy thing to deal with. Just to get a feel for LOCs:

https://github.com/GNOME/libxml2

There's also the web's (horrible) tendency to make a diarrhea of special-purpose tags and attributes--many of which could be subsumed by basic turing-completeness. Compare this to postscript, which was a relatively simple stack machine to begin with.

And of course, if we start special-purposing things, we're no longer compliant with the standard, and it would be chaos to call it SVG at that point.
1 comments
InnerGargoyle 1968 days ago
yes i think if we have to write very customised version of svg rendering then it would be like writing another graphics system, why bother with svg then?

on second thaught if xml is used for encoding and declaration of a perticular drawing, then what is used to render actual svg? is this again going on OS's graphical system?

link
chaganated 1968 days ago
There are many SVG rasterizers out there. This one is okay:

https://en.wikipedia.org/wiki/Librsvg

Could just be that I'm a cross-platform guy, but it seems like most programs bypass the OS rasterization stack (rendering fonts/curves/polygons/etc. with open source libraries rather than native OS facilities, and sending the OS a finished bitmap). This choice probably addressed a mix of cross-platform consistency issues, as well as the lowest-common-denominator problem.

link