mam@gvgspd.GVG.TEK.COM (Mark A. Matthews) (11/15/88)
We've got some PC users here who covet the output of our Localtalk-connected Laserwriter. We also have recently installed a Kinetics FastPath 4. The PC's each have 3-Com ethernet cards. Therefore, a path exists between the Laserwriter and the PC's. Is there Appletalk software available to take advantage of this sort of network? (i.e., the PC speaks Postscript to the LaserWriter via Appletalk over ethernet, through the FastPath, then the Localtalk.) -Mark (mam@gvgspd.GVG.TEK.COM - or - ...!tektronix!gvgpsa!gvgspd!mam)
morgan@JESSICA.STANFORD.EDU (11/17/88)
tektronix!gvgpsa!gvgspd!mam@ucbvax.berkeley.edu (Mark A. Matthews) writes: > We've got some PC users here who covet the output of our > Localtalk-connected Laserwriter. We also have recently installed a > Kinetics FastPath 4. The PC's each have 3-Com ethernet cards. > Therefore, a path exists between the Laserwriter and the PC's. Is > there Appletalk software available to take advantage of this sort of > network? (i.e., the PC speaks Postscript to the LaserWriter via > Appletalk over ethernet, through the FastPath, then the Localtalk.) The traditional answer to the "EtherTalk for PCs" question has been no, but perhaps there is hope on the horizon. According to yesterday's InfoWorld, The Sun Microsystems TOPS division (TOPS for short) has just announced that its new version of TOPS/DOS will work on Ethernet-attached PCs, as well as LocalTalk-attached. The initial release claims to support 3Com EtherLink II and EtherLink/MC (for MicroChannel machines) [what, *not* the original EtherLink? grumble, grumble]. Our TOPS salesperson was willing to guess that their NetPrint package would share in this support, which would meet this particular need. Calloo callay O frabjous day. We'll see if it works. It's still a wonder to me that no other company has leapt into this game. (I have no connection with TOPS, except that they brought the donuts to a meeting here once.) - RL "Bob" Morgan Networking Systems Stanford
samb@REYES.STANFORD.EDU (Sam Brain) (11/17/88)
We also have lotsa PCs with 3Com cards in them, connected to ethernet, and Macs and LWs on LocalTalk. Access to LWs is easy if you have a UNIX machine on the ethernet. We run PCNFS, Sun's PC version of NFS, on the PCs, and NFS on the UNIX machine, a VAX in our case. We also run the CAP code on the VAX and use papif as spooler. The PC users just issue a "NET USE lpt1: \\hostname\printername" command (usually in the autoexec.bat) where "printername" is the UNIX name of the LW printer (the name in /etc/printcap), and bingo! all the output to lpt1: ends up going to the LW. The PCNFS package comes with a copy of a daemon, pcnfsd, which handles the print requests on the UNIX box. It all works very well, here. The LWs are accessed by PC- VAX- and Mac-users alike. Sam Brain Department of Radiation Oncology Stanford Medical Center Stanford, CA 94305.
A.Eric@GSB-WHY.STANFORD.EDU (Eric M. Berg) (11/21/88)
>> We run PCNFS, Sun's PC version of NFS, on the PCs, and NFS on the UNIX >> machine, a VAX in our case. We also run the CAP code on the VAX and >> use papif as spooler. >> The PC users just issue a "NET USE lpt1: \\hostname\printername" command >> (usually in the autoexec.bat) where "printername" is the UNIX name of the >> LW printer (the name in /etc/printcap), and bingo! all the output to lpt1: >> ends up going to the LW. We have a similar setup (except that we're using TOPS rather than CAP on our Unix system); however, in our experience, it isn't quite as easy as this. A major problem is that many DOS applications don't generate fully conforming PostScript; in particular, the PostScript they generate doesn't begin with the two character "magic number" of "%!". (Microsoft applications, such as MS Word or "Write" under MS Windows, are examples of applications with this "feature".) Now, the Unix-based LaserWriter spooling software (i.e. Transcript from Adobe, or TOPSprint from TOPS) looks for that "magic number" in order to decide whether the print job is already PostScript (and should thus be sent to the printer as is), or is plain text (and should thus be prefixed with a header of PostScript code which essentially says "take the rest of this file after the header and list it on the printer"). As a result, the DOS user prints a MS Word document, the output goes via PS-NFS to the Unix server, the print spooler looks for and can't find the magic number, so it decides the file is really plain ASCII text, and you get pages and pages of PostScript code being listed on your printer, rather than the document you wanted printed. Another problem is that applications often don't generate real PostScript; instead, they generate something which contains application-specific commands, and then send the printer a set of dictionary routines which define these commands in terms of PostScript primitives. So you have to make sure that this initialization file is sent to the printer before your document is printed. Again, this isn't always as smooth as one would like; the DOS applications seem to be have been constructed on the model of a printer connected directly to a single PC, rather than on the basis of a network printing environment where the printer's state can vary because of jobs printed by other users. We've been able to get around the problems by (1) having our users print to a file instead of directly to the printer; and (2) then having them run a DOS command file which takes sends first the initialization file, then the document, to the printer, making sure that the first line of each is the characters "%!". Needless to say, this is kind of a nuisance. If anyone is aware of an easier way to make this all work, I'd be delighted to hear of it! I tried modifying Microsoft Word's printer driver for PostScript to put the "%!" at the beginning of each output file, but there wasn't enough flexibility in the driver to be able to do it. Eric Berg Price Waterhouse Technology Centre -------