jzavgren@bbn.com (John Zavgren) (10/21/88)
I am looking for programs for plotting data that produce high-quality output. I want the capabilities of: labeling axes, inserting legends, plotting characters, including titles and subtitles, splines, etc. I have tried the PiCTeX package. The resulting LaTeX programs are extremely slow to process; the plotting character is a period. Perhaps thousands of periods must be plotted to create even a simple curve. It has every capability I need. But, it is not practical. I mention it only as an indication of the capabilities I am looking for. I beleive the way to go is PostScript. After all, all our printing is done on PostScript printers. So, we might as well eliminate as many intermediate steps as possible. Are there any good UNIX plotting packages that produce PostScript? RSVP John Zavgren
majka@grads.cs.ubc.ca (Marc Majka) (10/22/88)
In article <31190@bbn.COM> jzavgren@bbn.com (John Zavgren) writes: >I beleive [sic] the way to go is PostScript. After all, all our printing is >done on PostScript printers. So, we might as well eliminate as many >intermediate steps as possible. [...] PostScript is an adequate page description language. It is a poor graphical description language. Sure, all your printers understand PostScript, but printers are hardly the only kind of graphic output devices. Producing good graphics in PostScript is, frankly, a pain in the butt. As the fledgling "industry" of desktop publishing grows up, it will begin to discover that there are more exciting graphical concepts than those offered by PostScript. Expect either a new page description language to replace PostScript, or big changes in the graphics it offers. Writing PostScript-based graphics utilities will get you two things: soon-outdated software, and lousy graphics. --- Marc Majka
richard@gryphon.CTS.COM (Richard Sexton) (10/22/88)
In article <4357@ubc-cs.UUCP> majka@grads.cs.ubc.ca (Marc Majka) writes: >In article <31190@bbn.COM> jzavgren@bbn.com (John Zavgren) writes: >>I beleive [sic] the way to go is PostScript. After all, all our printing is >>done on PostScript printers. So, we might as well eliminate as many >>intermediate steps as possible. [...] > >PostScript is an adequate page description language. iNo, HP's PCL is adequate. PS is excellent. >It is a poor graphical description language. Oh ? You have a graphic that cannot be described by PS ? >Sure, all your printers understand PostScript, but printers are > hardly the only kind of graphic output devices. Ok, what else do we have: plotters, CRT's, film recorders, phototypesetters. Plotters - good for 8 colour var charts - low quality. Crt's - usefull for preview and little else. You can videotape form an NTSC signal, but thats not of great quality. However, high quality, professional output is obtained almost exclusivly from: Phototypesetters - Now were talking. If you are going to produce something of high print quality, this is how you are going to do it, either monochrome or colour through colout seps. No, you can't take a TIFF or GIF or GKS or PHIGS file to a printer. Film-recorders - You wanna make a movie ? This will be your output device. Input ? Really dumb. Usually a bitmap, or an analog voltage into the device. Yucch. Having a PS compatible film recorder would be a boon. I seem to recall one being discusses somewhere, but can't rememebr when. >Producing good graphics in PostScript is, frankly, a pain in the butt. This is true. Sitting down with a PS printer hooked up to a terminal is a pain. Most of us use programs to generate the PS. > As the fledgling >"industry" of desktop publishing grows up, it will begin to discover that >there are more exciting graphical concepts than those offered by PostScript. Ok. Like what ? >Expect either a new page description language to replace PostScript, or big >changes in the graphics it offers. Writing PostScript-based graphics >utilities will get you two things: soon-outdated software, and lousy graphics. Sheer, utter, and unadulterated bullshit. The software will not be outdated. When you and I are dead, there will be machines capable of rendering PS files. And as for lousy graphics, PS will not prevent you from drawing a bad graphic. Only a good sense of design and a smattering of talent will. But DO tell us. What is the alternative ? Or are you just a frustrated LaserJet I owner ? -- Point of light #999 richard@gryphon.CTS.COM {backbone...err, well connected site}!gryphon!richard
wen-king@cit-vax.Caltech.Edu (King Su) (10/23/88)
In article <7999@gryphon.CTS.COM> richard@gryphon.CTS.COM (Richard Sexton) writes: >In article <4357@ubc-cs.UUCP> majka@grads.cs.ubc.ca (Marc Majka) writes: <>In article <31190@bbn.COM> jzavgren@bbn.com (John Zavgren) writes: >>>I beleive [sic] the way to go is PostScript. After all, all our printing is <>>done on PostScript printers. So, we might as well eliminate as many >>>intermediate steps as possible. [...] <> >>PostScript is an adequate page description language. < >iNo, HP's PCL is adequate. PS is excellent. < >>It is a poor graphical description language. < >Oh ? You have a graphic that cannot be described by PS ? Since you are very sure of your self, you won't mind sharing your solution for my problem, would you? I have no quarrel about the suitablility of PostScript as a printer language, except one thing. I am into chip plotting; I want to plot the various layers of an integrated circuit design using different stipple patterns so that when two or more layers overlap the union of the patterns produces a unique stipple pattern. Goemetries in each layer are ploted as polygons. Since PostScript does not allow transparent writing of polygons (OR-ing a pattern onto a page instead of painting an opague pattern over the page, a white dot does not erase a black dot), I could not accomplish what I want with PostScript short of rasterizing the whole thing and sending it to the printer as a bit array or compute intersections of the layers and decompose all the polygons, both of which are un-acceptibly slow. Even the most ardent supporters of PostScript here has to admit that this is a very serious short-coming of the language. Any suggestions? I will be very grateful if you can help me. -- /*------------------------------------------------------------------------*\ | Wen-King Su wen-king@vlsi.caltech.edu Caltech Corp of Cosmic Engineers | \*------------------------------------------------------------------------*/
tonyw@microsoft.UUCP (Tony Williams) (10/27/88)
In article <8392@cit-vax.Caltech.Edu> wen-king@cit-vax.UUCP (Wen-King Su) writes: | I am into chip plotting; I want to plot the various layers of an | integrated circuit design using different stipple patterns so that | when two or more layers overlap the union of the patterns produces | a unique stipple pattern. Goemetries in each layer are ploted as | polygons. Since PostScript does not allow transparent writing of | polygons (OR-ing a pattern onto a page instead of painting an opague | pattern over the page, a white dot does not erase a black dot), I | could not accomplish what I want with PostScript short of rasterizing | the whole thing and sending it to the printer as a bit array or | compute intersections of the layers and decompose all the polygons, | both of which are un-acceptibly slow. Even the most ardent supporters | of PostScript here has to admit that this is a very serious short-coming | of the language. Any suggestions? I will be very grateful if you | can help me. Set the polygon as a clip region. Draw the black portions of the stipple pattern either explicitly with lines/dots etc, or using the imagemask operator. I admit, for certain patterns this might be slow and/or require replicating the imagemask pattern many times. Tony