tj@gpu.utcs.utoronto (tj) (03/15/87)
I sent another article that started as a printer review and ended up
with me setting fire to the curtains in my computer room when the
flames got too high!!! Decided to address the issue here.
We decided to use PostScript Printers at U of Toronto, good choice so far.
We use them attached to Unix and via network we send MVS, VM/CMS and Unix
code to the devices. (We have a Linotype Model 100 1250 dpi typesetter,
a TI 2115 15 ppm Laser Printer and an Apple LAserWriter)
We also share the devices with micros, both IBM/PC's and Mac's.
It is in the area of PC's that I have tho most complaints about drivers
but none of the areas are free from blame.
Final Word II: Gold Star... works well, handles Xon/Xoff. Reliable. Even
handles the error messages that come back from device.
Microsoft Word 3.0: Silver Star... works well, you MUST issue the mode
command to make sure the port is the right baud etc BEFORE
you go into MicroSoft Word. Another "feature" is that it sends
down an initialization file ONCE to the device. If you turn off
the printer in mid job, or switch to the Linotype then the
device isn't re-initialized and the job vanishes. The way to
get it to re initialize is to go out of Word and back in OR
switch to a different device then switch back to APPLASER.
DOES NOT handle error messages. %%[Paper tray empty]%% causes
the job to flush from there on. No support for the IBM extended
character set. Access to symbol set but whats on teh screen is
not near whats on the paper!
WordPerfect 4.2: No Star yet. Word Perfect has done the poorest job of any
company I have seen. We got 4.2 from the computer shop and
tried it. Their test page didn't work anywhere near correctly.
Newer version works they said. (Yes virginia there are different
versions of 4.2 from the devices supported point of view)
6 wks later our 4.2 network version arrives. Same story.
I have heard on the net from someone who has used it from
4.2 and had success. There is a naming change between the
initialization files on the working 4.2 and the failing 4.2
so it is easy to tell whether you have the bad or the good
4.2 (damn I can't remember which was which, will post)
OTHER PROBLEMS.....
WordPerfect DOES NOT handle the Xon/Xoff. They give you
a file to send to your LaserWriter that sets it to DTR
handshaking mode and a file that sets it back to factory
settings. Two problems. 1) the printer may not have been
at factory settings (we used even parity not none) and 2)
Early Laserwriters DID NOT support DTR handshaking therefore
you are out of luck if you have an old one. The product
sucks so far... lets hope they are listening. (Note to
wordperfect: you call me when its done, we are busy writing
a Microsoft Word course to teach at U of Toronto since
WordPerfect doesn't deliver what they say... anybody want
teaching notes for WordPerfect real cheap!)
P.C. Write: Silver Star. They came real close. One command messes up a bit.
Some conflict in the fonts and the Landscape selection. Not
generally used, may be fixed by now. Supports (I think)
Xon/Xoff (another person tested PC Write, not me) and has
pretty good support for accented chars etc. For an inexpensive
package they deserve a gold star. Unlike WordPerfect you get
at LEAST what you pay for.
Nota Bene 2: (II?) Silver star. The first driver they sent had some problems
and at first the second one did too, then suddenly stasrted
to work. So far so good. Does not have Xon/Xoff built in so
they supply a TSR Xon/Xoff handler. Accent char support is just
OK as I understand. Sort of relative, since Nota Bene is
real popular with people that use foreign langs, it might be
considered poor accent support, but by others standards I guess
it is OK.
Macintosh Stuff: A fiasco is happening here. A single user with but one
software package may never experience problems. Network users
using different versions of the finder and system and
the initialization file (LaserPrep) will fight. Each machine
sends "their" initialization stuff and there ARE
incompatibilities. This is a GIANT can of worms and only a
couple have come out so far. The worst is yet to come.
Unix stuff: There are problems here too. Let me relate... Accounting is done
here at U of Toronto. (imagine wanting to bill people!!!)
There is some software that comes with the Adobe transcript
package. In general it is good, but not perfect. you can no
longer count lines of data and divide by 55 to determine
pages to charge for. We opted for page AND time charges since
antisocial users were sending escher (sp) drawings taking
15 mins per page and we thought they should pay! We charge
CPU time on teh printer! 68000 is worth $.20 per minute.
This applies to LaserWriter, Omnilaser, and Linotype.
We had to modify the driver to ask the printer for usertime
as well as page count. (oh yes, the driver asks the pagecount
at the start and end of the job and determines pages from that
instead os estimating!) (more on this) There were problems
with the usertime thing... the PostScript usertime is useful
but (Adobe LISTEN) has a bug (I think it is) in that when the
printer stops because the paper tray is empty or missing or the
Linotype roll of film ends etc, usertime keeps incrementing.
Manual says usertime is the amount of time spent interpretting
users PostScript code, not waiting fro the operator to fill
paper tray! Anyway we had to send code to the printer that
replaced showpage and copypage so that it looks at usertime
that elapses BETWEEN showpages/copypages and adds a finite
amount of time for each showpage/copypage so even if it
hangs there is no increase. (it works). Pagecount.....
good for most purposes except that the Linotype has varying
size pages. Our header page is .5 inches long and users send
usually from 8.5 inch to 17 inch pages but 51 inch or more is
possible. We have to charge for inches. no inchcount PostScript
operator. We send down code to the printer that replaces
showpage/copypage and records the pageparams page width in
inches and keeps track of them. We also replace the pagecount
operator to tell us how many 8.5 inch pages a user has printed
(.5 inch header plus one 8.5 inch page would normally return
pagecount of 2 but with ours we get 1.050!) This works.
Linotype..... I want the designers shot and killed. The device
has no means of reporting back to the PostScript Raster Image
Processor (rasterizer... RIP) that the paper casette is full
so it will cheerfullt put a 150 ft roll of paper into a casette
designed for 12 feet. Also doesn't report that the paper casette
is missing. In either case output ends up on the floor! Really
bad for photographic paper before processing unless you live in
total darkness!! Send the Linotype code (modeled after some
contributed by Adobe thank you) that counts pages (soon to
be changed to inches as per previous probs) and stops after
12 pages and via a dumb terminal hooked to the alternate
serial port and via an undocumented postscript function
in serverdict or statusdict called altout sends a message to
this port telling op to process output. Now all these problems
made changes to the transcript software necessary. Pagecounts
were now float valuse, we had an extra paramater, usertime etc.
It is doable, we still have a few bugs, but it works pretty
well. Certainly wasn't drop in place stuff.
Feel free to call for more info
Terry Jones
416-978-4924
(software developers feel free to call and apologize and send working stuff!)
(Especially WordPerfect, in the dog house... almost given up on you!!!)