> There was the `DVI standards' committee. It refused to address the > real issues including \special{...} at a point where it might actually have made > some difference. Hence lost a great chance to prevent a huge mess. > (They did however discuss how many angels can fit into one scaled point). i think they were overawed with the effort required to write a standard. i told 'em so when they started: i also told 'em about the angels/pin head danger... the standard they produced was intended to be extended into areas where it might actually have been useful (in berthold's sense) but i guess they were `whacked' after what they did do. i'm not surprised: writing standards (with any degree of rigour) requires stamina -- it's not something i would undertake as a voluntary, spare-time, project. incidentally, i would contest william hammond's curious assertion that dvi is in some sense a `higher level' format than pdf or ps. in any document modelling i've ever read (e.g., oda, for which i gave a reference a while back), all three fit at the bottom of the tree, being non-revisable[*] forms. that no package exists to create dvi from pdf or ps merely represents the futility of making such a transformation -- it doesn't suggest that the transformation is impossible. i agree with berthold: let's stay at the top level -- the revisable (la)tex input file. arguing about output formats, whether they be aimed at the dot-matrix printer or reasonably generic gets us absolutely nowhere imo. robin [*] which is not to say they _couldn't_ be edited, merely that no-one in their right mind _would_ edit them in the ordinary course of events.