Re: AbiWord's printing architecture

From: Robert Roebling (Robert.Roebling@t-online.de)
Date: Sun Jan 26 2003 - 16:54:44 EST

  • Next message: Dom Lachowicz: "Re: AbiWord's printing architecture"

    Leonard Rosenthol wrote:

    > >Well, we do :-) I just tell the FreeType backend
    > >of Pango to produce outlines of the rendered
    > >text and write these to the PS file.
    >
    > 1) The "printer" is no longer able to using the
    > hinting data in the font to do better quality
    > output on raster devices and/or lower resolution
    > PS printers.

    How much of a reality is that? I use this for printing
    on Linux with ghostscript. I haven't even seen a printer
    which does less than 300dpi and most do far more. I'm
    also not sure what rendering engine GhostSript uses, but
    I'm not sure it is so much better than FreeType's and
    that is what Pango uses to create the outlines.

    > 2) If the user chooses to take the output PS file
    > and convert it to PDF

    I admit that this is a drawback. But I currently
    aim at printing, the rest is secondary.

    > As Dom noted, the correct thing to do is support
    > font embedding with subsetting.

    I didn't find any info on this yet, but doesn't
    that mean that the printing code has to scan the
    whole printout which glyphs have been used and
    it will then dump all those glyphs into the PS
    file? I am particularly concerned about the order
    of things as I suspect that the glyph data has
    to be at the top of the PS file and the actual
    text comes later. BTW, does AbiWord do this
    glyph subsetting?

    I am still undecided on what the best approach
    is, but it strikes me as stupid that so many
    Linux projects waste time on writing yet
    another variant combining Xft, FriBidi and
    PS with subsetting. I think Linux would look
    (well, print) better with a PangoPS library.

    Regards,

      Robert

    -- 
    Robert Roebling, MD <robert.roebling@medizin.uni-ulm.de>
    


    This archive was generated by hypermail 2.1.4 : Sun Jan 26 2003 - 16:57:29 EST