Article 73049 of comp.os.vms: Path: cs.utk.edu!emory!news-feed-1.peachnet.edu!news-feed-2.peachnet.edu!darwin.sura.net!europa.eng.gtefsd.com!howland.reston.ans.net!noc.near.net!news.Brown.EDU!stout!robinson From: robinson@stout.geo.brown.edu (Darrin Robinson) Newsgroups: comp.lang.postscript,comp.os.vms,comp.os.msdos.misc Subject: Re: Brain-Dead Auto-Sensing Postscript Printers Date: 22 Sep 1993 21:32:34 GMT Organization: Brown University, Providence RI 02912 Lines: 53 Message-ID: <27qg9i$9oa@cat.cis.Brown.EDU> References: <21SEP199323170949@ariel.lerc.nasa.gov> <27qdrv$cef@agate.berkeley.edu> NNTP-Posting-Host: stout.geo.brown.edu Xref: cs.utk.edu comp.lang.postscript:20020 comp.os.vms:73049 comp.os.msdos.misc:7943 In article <27qdrv$cef@agate.berkeley.edu> andywang@crown.berkeley.edu writes: >jgreely@dinosaur.cis.ohio-state.edu (J Greely) writes: > >>In article <21SEP199323170949@ariel.lerc.nasa.gov> uugblum@ariel.lerc.nasa.gov >> (Greg Blumers [3-6777]) writes: >>>I am a bit upset that the auto-emulation sensing mode doesn't work. > >>I found it amusing. After playing with the HP 4si/MX, I expected >>autosensing to "just work". Then I took a look at the QMS 1725, which >>failed dramatically; reading the manual, QMS says: >> "Using a form of artificial intelligence, ESP technology >> analyzes incoming file data from any of the printer's >> interfaces. ESP technology, which works with _most_popular_ >> _commercially-available_applications_ [emphasis mine], selects >> the appropriate printer language from those installed on the >> printer and processes the print job..." >> -- QMS Document Option Commands, page 1-2 > >>The real problem is that the "document options" commands look like >>PostScript DSC comments, so the printer attempts to parse any file >>that begins with "%!", stops when it finds something that doesn't look >>like one of their commands, and *then* runs the auto-sensing code. >>-- >>J Greely (jgreely@cis.ohio-state.edu; osu-cis!jgreely) > >I spent hours trying to get our QMS PS-410 to autosense using >its EGP (Emulation Guessing Processor). The printer always >ended up spewing out pages of Postscript code in ASCII, jamming, >or doing nothing. QMS even gave me a firmware update early this >year for free. So what, it didn't work. > >I finally replaced it with a HP LJ4m. Works great! > > >Andrew Wang While evaluating more than 8 median range laserprinters over the previous winter, we ran into that alot. I won't name names, but we purchased about 58 HP Laserjet 4si MXs with duplex option installed for many reasons. The 4si MX's JetDirect Card (Multi-IO) is very nice... and the PS autosensing works 99.999999% of the time. (In about 800,000 pages (or 150,000 jobs), it messed up once... it printed postscript code... go figure... but the other 149,999 jobs and 799,990 pages it did just fine! We only print Postscript here, using Transcript filters, and nearly all of the printers are in Personality=Auto mode) , Darrin E. Robinson (DER31) Hamnet N1LLV 146.700-, 146.880- MHz /| Systems Programmer Internet darrin@MIT.EDU \| Dist. Computing & Network Services robinson@Planetary.Brown.EDU |\ M.I.T. Information Systems ICBMnet 41 29 24 N 71 18 48 W (NPT) |/ 1 Amherst St. - Rm E40-338 SPANet PGGIPL::ROBINSON (7132) ' Cambridge, MA 02139, USA AT&Tnet (617) 253-0131 Article 73051 of comp.os.vms: Newsgroups: comp.lang.postscript,comp.os.vms,comp.os.msdos.misc Path: cs.utk.edu!ornl!news.ssc.gov!linac!uwm.edu!math.ohio-state.edu!cs.utexas.edu!uunet!imagen!weingart From: weingart@imagen.com (Phil Weingart) Subject: Re: Brain-Dead Auto-Sensing Postscript Printers Message-ID: <1993Sep22.215548.17416@imagen.com> Sender: usenet@imagen.com Nntp-Posting-Host: napa Organization: QMS Inc., Imagen Division References: <21SEP199323170949@ariel.lerc.nasa.gov> Date: Wed, 22 Sep 1993 21:55:48 GMT Lines: 50 Xref: cs.utk.edu comp.lang.postscript:20021 comp.os.vms:73051 comp.os.msdos.misc:7945 In article jgreely@dinosaur.cis.ohio-state.edu (J Greely) writes: >In article <21SEP199323170949@ariel.lerc.nasa.gov> uugblum@ariel.lerc.nasa.gov > (Greg Blumers [3-6777]) writes: >>I am a bit upset that the auto-emulation sensing mode doesn't work. > >The real problem is that the "document options" commands look like >PostScript DSC comments, so the printer attempts to parse any file >that begins with "%!", stops when it finds something that doesn't look >like one of their commands, and *then* runs the auto-sensing code. This is not the problem. The problem is twofold: 1) The QMS emulation sensing algorithm works on the first N bytes of the job. In order to do so, it has to know where the job begins. On bytestream interfaces (serial, parallel) this is not easy. The printer will end a job due to timeouts, but on a shared printer receiving jobs from a spooling server, it is likely that one job will immediately follow another much of the time, so timeouts don't work. The answer here is to insert the documented End-of-Document string at the end beginning of each job, which is fairly easy to do in Netware using PDFs which QMS will supply on request. On the current version of QMS Crown printers, use %%EndOfDocument. On older revisions, and onthe PS 410, 815, 825, etc., use 0 statusdict begin 0 200 setemulation end This will work on a serial interface. On parallel, it's "1 200 setmeulation" instead of "0 200 setemulation". 2) QMS' emulation sensor takes a statistical sample of character pairs from the first N bytes of each file and compares that profile against profiles compiled by examining files of known language. With this approach, we have a 97% + success rate on customer files. The most common failure is on text files with special formatting conventions (/etc/printcap is a good example) or special characters (C program listings, for instance); these files tend to look like postscript or HPGL to ESP. The problem described above was an issue two years ago, but on the product revisions going out the factory door today, I almost never see the printer miss on a postscript file, as long as the files are properly delimited. ------------------------------------------------------------------------------- Philip K. Weingart, Systems Engineer QMS Philadelphia philw@sales.imagen.com 735 Chesterbrook Blvd. Phone: 215-296-4900 Fax: 215-296-2191 Wayne, PA 19087 Article 73119 of comp.os.vms: Path: cs.utk.edu!emory!ogicse!news.tek.com!gazette!frip!andrew From: andrew@frip.wv.tek.com (Andrew Klossner) Newsgroups: comp.lang.postscript,comp.os.vms,comp.os.msdos.misc Subject: Re: Brain-Dead Auto-Sensing Postscript Printers Message-ID: <27sr4v$mdf@gazette.wv.tek.com> Date: 23 Sep 93 18:50:07 GMT Article-I.D.: gazette.27sr4v$mdf References: <21SEP199323170949@ariel.lerc.nasa.gov> <23SEP199310390402@almach.caltech.edu> Reply-To: andrew@frip.wv.tek.com Distribution: world Organization: Tektronix Color Printers, Wilsonville, Oregon Lines: 30 NNTP-Posting-Host: frip.wv.tek.com Xref: cs.utk.edu comp.lang.postscript:20039 comp.os.vms:73119 comp.os.msdos.misc:7980 Regarding the HP4m: "When connected via a serial port to a VMS print queue, I can fool it by stacking a postscript job followed by a non-postscript job in the queue." You're not "fooling" it. You're concatenating two jobs into a single one. The printer won't auto-sense a language change within a job. PostScript jobs sent over a serial port must be terminated by control-D, the end-of-file character. If the serial port is idle for a long enough period, a "timeout" error will happen and the job will terminate. Note that this is an error, not a normal termination, and any error handler will be invoked. "Eventually we figured out that we can telnet to port 9100 on the printer and send files to it this way, as long as you follow the text you want printed with 56000 non-printing characters to flush the buffer." The printer doesn't speak the telnet protocol at port 9100. It just wants a raw stream of TCP/IP bytes, followed by a TCP FIN. 56k is roughly the size of PostScript's input buffer, but PostScript would much rather see a clean EOF, which it gets from the network stack when it sees FIN. In a network program that uses sockets, call shutdown() after sending the last byte so that FIN will be transmitted. -=- Andrew Klossner (andrew@frip.wv.tek.com) Article 73123 of comp.os.vms: Path: cs.utk.edu!avdms8.msfc.nasa.gov!europa.eng.gtefsd.com!howland.reston.ans.net!math.ohio-state.edu!cis.ohio-state.edu!dinosaur.cis.ohio-state.edu!jgreely From: jgreely@dinosaur.cis.ohio-state.edu (J Greely) Newsgroups: comp.lang.postscript,comp.os.vms,comp.os.msdos.misc Subject: Re: Brain-Dead Auto-Sensing Postscript Printers Date: 23 Sep 1993 15:31:55 -0400 Organization: Ohio State University Computer and Information Science Lines: 28 Sender: jgreely@cis.ohio-state.edu Distribution: world Message-ID: References: <21SEP199323170949@ariel.lerc.nasa.gov> <23SEP199310390402@almach.caltech.edu> NNTP-Posting-Host: dinosaur-o.cis.ohio-state.edu In-reply-to: shoppa@almach.caltech.edu's message of 23 Sep 1993 10:39 PST Xref: cs.utk.edu comp.lang.postscript:20040 comp.os.vms:73123 comp.os.msdos.misc:7981 In article <23SEP199310390402@almach.caltech.edu> shoppa@almach.caltech.edu (Timothy D. Shoppa x4256) writes: >We just got a HP LaserJet 4m, and its format-guessing is not quite perfect. >When connected via a serial port to a VMS print queue, I can fool it by >stacking a postscript job followed by a non-postscript job in the queue. Sounds like your spooler is broken. It should be appending a Control-D to each job, so that the printer knows where each job ends. Otherwise, it has to rely on the timeouts. >We are less than happy with the 4m's "network interface", however. We bought >it thinking that we just hook it up (via thinwire or twisted pair) and >it would look just like a lpd ("networked printer") device (this is what >the HP salesperson had suggested, without guaranteeing!) Personally, I wouldn't *want* a printer that spoke lpd; I rather like the HP4's approach of listening on TCP port 9100, and letting you speak to it interactively with telnet (hint, hint). We didn't buy their software; I just wrote a small program that opens the port and spits the job across. >as long as you follow the text you want printed >with 56000 non-printing characters to flush the buffer. You're kidding, right? Why not just close the port and reopen it? Works for me... -- J Greely (jgreely@cis.ohio-state.edu; osu-cis!jgreely) Article 73147 of comp.os.vms: Newsgroups: comp.lang.postscript,comp.os.vms,comp.os.msdos.misc Path: cs.utk.edu!gatech!howland.reston.ans.net!europa.eng.gtefsd.com!uunet!imagen!weingart From: weingart@imagen.com (Phil Weingart) Subject: Re: Brain-Dead Auto-Sensing Postscript Printers Message-ID: <1993Sep24.040105.18180@imagen.com> Sender: usenet@imagen.com Nntp-Posting-Host: napa Organization: QMS Inc., Imagen Division References: <23SEP199310390402@almach.caltech.edu> Date: Fri, 24 Sep 1993 04:01:05 GMT Lines: 20 Xref: cs.utk.edu comp.lang.postscript:20049 comp.os.vms:73147 comp.os.msdos.misc:7991 In article jgreely@dinosaur.cis.ohio-state.edu (J Greely) writes: >>We didn't buy their software; I just wrote a small program that opens >>the port and spits the job across. > >Since half a dozen people have asked me about this in half a dozen >hours, For all those who want something which opens a port and dumps data, the QMS host software for UNIX does exactly that, although it also checks for status packets from the printer. You can get it from gatekeeper.imagen.com, in the pub/cts/QMSoft_release directory. Any one of the .tar files will have the right stuff. It's there for QMS customers, but it's there. The module which does the work is called qef.c. QMS printers listen on port 35, so if you've got someone else's TCP port, you need to change the port number. Phil Weingart weingart@napa.imagen.com My employer accepts no responsibility for my opinions, and I can't say that I blame them. Article 73190 of comp.os.vms: Newsgroups: comp.lang.postscript,comp.os.vms,comp.os.msdos.misc Path: cs.utk.edu!gatech!howland.reston.ans.net!europa.eng.gtefsd.com!uunet!iphase.com!iex!peters From: peters@iex.UUCP (James A. Peters) Subject: Re: Brain-Dead Auto-Sensing Postscript Printers Message-ID: <1993Sep24.224106.1662@iex.uucp> Sender: usenet@iex.uucp (USENET news) Organization: IEX Corporation References: <21SEP199323170949@ariel.lerc.nasa.gov> <23SEP199310390402@almach.caltech.edu> Date: Fri, 24 Sep 1993 22:41:06 GMT Lines: 51 Xref: cs.utk.edu comp.lang.postscript:20067 comp.os.vms:73190 comp.os.msdos.misc:8017 In article jgreely@dinosaur.cis.ohio-state.edu (J Greely) writes: >In article <23SEP199310390402@almach.caltech.edu> shoppa@almach.caltech.edu > (Timothy D. Shoppa x4256) writes: >>We just got a HP LaserJet 4m, and its format-guessing is not quite perfect. >>When connected via a serial port to a VMS print queue, I can fool it by >>stacking a postscript job followed by a non-postscript job in the queue. > >Sounds like your spooler is broken. It should be appending a >Control-D to each job, so that the printer knows where each job ends. >Otherwise, it has to rely on the timeouts. > >>We are less than happy with the 4m's "network interface", however. We bought >>it thinking that we just hook it up (via thinwire or twisted pair) and >>it would look just like a lpd ("networked printer") device (this is what >>the HP salesperson had suggested, without guaranteeing!) > >Personally, I wouldn't *want* a printer that spoke lpd; I rather like >the HP4's approach of listening on TCP port 9100, and letting you >speak to it interactively with telnet (hint, hint). We didn't buy >their software; I just wrote a small program that opens the port and >spits the job across. > >>as long as you follow the text you want printed >>with 56000 non-printing characters to flush the buffer. > >You're kidding, right? Why not just close the port and reopen it? >Works for me... i am not real impressed with the hp 10baseT/bnc ethernet directjet interface either. i do not have an opinion on the bsd lpr interface software for it is an addon for an hp4si mx, even though you get 10M, postscript, netware mswindows drivers, appletalk and ethernet directjet interfaces as standard. attempting to order it turned out to be non-trivial. our shop is using lan- tronix eps print servers for our hp3si's, hpIId's, qms printers. having to administer yet another piece of software is distasteful. hp's lack of a timely response to a sales order for $150 made the decision to use another eps, list price $695, pretty simple. bottom line is if the ethernet had not come with the mx i would not ordered it. the appletalk interface is very nice and easy to use with the exception of support for the adobe written and hp marketed drivers. as for auto-sensing the hp4si has been flawless. in summary: hp4si good. hp appletalk directjet good. hp ethernet directjet not good. adobe/hp apple printer driver flakey. lantronix eps print server good. regards, james, peters@iex.com ps. i have an unused hp directjet ethernet interface for sale. Article 73214 of comp.os.vms: Newsgroups: comp.lang.postscript,comp.os.vms,comp.os.msdos.misc Path: cs.utk.edu!avdms8.msfc.nasa.gov!europa.eng.gtefsd.com!uunet!imagen!weingart From: weingart@imagen.com (Phil Weingart) Subject: Re: Brain-Dead Auto-Sensing Postscript Printers Message-ID: <1993Sep25.121424.20875@imagen.com> Sender: usenet@imagen.com Nntp-Posting-Host: napa Organization: QMS Inc., Imagen Division References: <1993Sep24.040105.18180@imagen.com> Date: Sat, 25 Sep 1993 12:14:24 GMT Lines: 24 Xref: cs.utk.edu comp.lang.postscript:20073 comp.os.vms:73214 comp.os.msdos.misc:8036 In article jgreely@dinosaur.cis.ohio-state.edu (J Greely) writes: >In article <1993Sep24.040105.18180@imagen.com> weingart@imagen.com > (Phil Weingart) writes: >>For all those who want something which opens a port and dumps data, the >>QMS host software for UNIX does exactly that, although it also checks >>for status packets from the printer. > > By the way, Phil, can you >restrict what hosts are able to reach the printer this way? No. If your QMS printer has a hard disk, you can store the equivalent of a hosts.equiv file on the printer and enable the feature to restrict hosts. There is a similar file/switch facility for restricting ftp sessions. You could also use UNIX features; on the host which spools to the printer, wipe out hosts.lpd and hosts.equiv, and require anyone who uses the printer to get a login on the spooling host. ------------------------------------------------------------------------------- Philip K. Weingart, Systems Engineer QMS Philadelphia philw@sales.imagen.com 735 Chesterbrook Blvd. Phone: 215-296-4900 Fax: 215-296-2191 Wayne, PA 19087 ---------------------------------------------------------------------------- Path: utkcs2!dublin.iona.ie!mailhost.nmt.edu!chaos.tcct.nmt.edu!warren Mailing-List: lprng@iona.com Date: Wed, 10 Sep 1997 16:50:09 -0600 (MDT) From: Warren Marts To: lprng@iona.com cc: garrett@qualcomm.com Subject: Re: [LPRng] HP 5Si left output bin In-Reply-To: <199709101537.IAA12885@yavin.qualcomm.com> Message-ID: > Do you know where I can get PPDs for the newer HP printers? I seem only to > find Windows .EXE files and Mac HQX files. (Which I know contain the PPDs, > but I want them in a format I can deal with natively on UNIX.) I've been > thinking of adding PPD parsing/option handling to ifhp/psfilter. I'd be happy to uncompress them on a PC and ftp them to our server for general FTP access, if you have a neither a Mac or DOS machine around or a Unix 'unzip'. As an aside, I've found the support area of www.adobe.com to be the best all around source for ppds and postscript drivers. They have a collection of PPD files for probably every Postscript device ever made, and you just download one file per manufacturer. I've found that they will have Windows Postscript drivers 2-6 months before the same version will appear on HP's or Microsoft's sites, although their Mac drivers tend to lag what's available at Apple. I wonder if this mean Adobe has passed Mac driver development to Apple, while continuing to do Windows drivers itself? -- Warren Marts -- Systems Programmer Socorro, NM, USA -- Computer Center, New Mexico Tech -----------------------------------------------------------------------------