blarson@blars (04/20/91)
In article <1991Apr18.152749.8762@cbnewsk.att.com> dennis@cbnewsk.att.com (dennis.seipel) writes: >I am developing a DCE and want to know an acceptable number of character >that a DCE may transmit after the DTE sends XOFF flow control. 0.9 (finish the character you already started sending.) >Is there >a standard for this or can anyone refer me to a study of what terminal >manufactures typically design for? Prime has been violating this principal of least astonishment for years, it mainly cause problems on serial printers. (HP printers were one of the worst to use with a Prime.) Prime does allow the system andministrator to adjust the maximum number of characters sent after an xoff is received, for one you limit the line to 10 CPS (sent at 19200 BPS of course). In other words, most equiptment will let you send some characters after it sends the xoff. The communications latency and processing time reduce the number that can in practice be sent by the sending side. For reliable communications, you should try not to send any additional characters after getting an xoff. 16 characers will, in practice, be accepted by the majority of terminals and printers. Check the manuals of the equiptment in question if you care, but don't expect to find this info, it's rarly there. > Dennis Seipel > dennis@inuxy.att.com Ah, the same company that designed an ISDN async switch that wouldn't work with several brands of computers and terminals and had a poorly thought out user interface. Maybe someday if you fix these problems we can use our 5ESS for terminal connections. (I know that AT&T is a huge company, and you probably have nothing to do with this.) -- blarson@usc.edu C news and rn for os9/68k! -- Bob Larson (blars) blarson@usc.edu usc!blarson Hiding differences does not make them go away. Accepting differences makes them unimportant.
cepek@vixvax.mgi.com (04/22/91)
In article <189@blars>, blarson@blars writes: > In article <1991Apr18.152749.8762@cbnewsk.att.com> > dennis@cbnewsk.att.com (dennis.seipel) writes: >>I am developing a DCE and want to know an acceptable number of character >>that a DCE may transmit after the DTE sends XOFF flow control. : >>Is there >>a standard for this or can anyone refer me to a study of what terminal >>manufactures typically design for? : > In other words, most equiptment will let you send some characters > after it sends the xoff. : > Check the manuals of the equiptment in question if you care, but don't > expect to find this info, it's rarly there. Sure enough, I went to check my manuals, and couldn't find it! :-| I recall that for either the WYSE 85 terminal (VT220 comp.) or the DataSouth 180(+) printer there was some statement about being rather tolerant during this form of handshaking. Something about sending XOFF when the buffer was x% full, then another when the buffer was really full. I even think that x was user selectable. BUT, I would NOT suggest relying on "tolerant" equipment on the other end. I bring this up to make people aware of this variation on the theme. The recommendations from blarson seem sound to me. ,---------------------------------,___________________________ / Mike Cepek, Programmer/Analyst / Internet: CEPEK@MGI.COM / \ Management Graphics, Inc. \ Voice: +1 612/851-6112 \ "Engage." / 1401 East 79th Street / Operator: +1 612/854-1220 / \ Minneapolis, MN 55425 USA \ Fax: +1 612/854-6913 \ `--------------------------------/___________________________/