Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!purdue!decwrl!shelby!polya!shap From: shap@polya.Stanford.EDU (Jonathan S. Shapiro) Newsgroups: gnu.config Subject: What about the Adobe Lawsuit? Summary: FSF should think about a position Message-ID: <8480@polya.Stanford.EDU> Date: 16 Apr 89 00:30:54 GMT Sender: Jonathan S. Shapiro Reply-To: shap@polya.Stanford.EDU (Jonathan S. Shapiro) Distribution: eunet,world Organization: Stanford University Lines: 52 Note: I have sent this to gnu.config, because gnu.gcc clearly isn't appropriate. I vote for creating gnu.politics. As many of you will already know, fonts are not copyrightable under the current copyright legislation. Unfortunately, while the intent was to deny protection for font representations, the wording specifies that it is the analog representation of fonts that is not copyrightable. The *description* of a font *is* copyrightable. The copyright office has issued its opinion that the denial of copyright should extend to _digital_ representations of fonts too (the representations printed by laser printers and the like is digital). Both Adobe and BitStream have filed suit to block this. The copyright office has no legal standing to judge law, but the precedent is that the courts pay strong heed to their input on the grounds that they are experts in the field of copyright. The issue is as follows: The Adobe font descriptions are protected. The question is, should it be legal for me to make a printout of the fonts in some very large size and use a scanner of some form to reverse-engineer my own font description? Alternatively, should it be legal for me to use the rasterization hook in the LaserWriter to get the printer to send me back data about which dots are drawn in that size [just saves scanning and consequent inaccuracies]? The issue has been raised by a project at University of California at Berkeley, which has done the latter, and is reselling their version of the font descriptions for $10,000 complete to commercial institutions. Their descriptions are in postscript, but are not the same as Adobe's and therefore not a copyright infringement. [Even if they were the same, a good case could be made under the "limited number of representations" clause of the copyright legislation, that says that representations of a concept that has a limited number of reasonable representations cannot be copyrighted. Fonts are not copyrightable under this logic.] The issue is important to FSF because if fonts are not reverse- engineerable, the benefit of GNU-ghostscript is greatly diminished. As further data for the fire, the hard part about reverse engineering fonts is the "hints" - clues that tell the printer things like "all the legs on an 'm' should be the same width. This hints are terribly important in smaller sizes, where the line width begins to approximate the dot width. Hints are not easily obtained from the printed form, though automated systems with limited success exist. So should FSF take a position on this case, and if so, is there a concensus on what it should be?