Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!wasatch!cs.utexas.edu!tut.cis.ohio-state.edu!ucbvax!SALT.ACC.COM!chris From: chris@SALT.ACC.COM (Chris VandenBerg) Newsgroups: comp.protocols.tcp-ip Subject: (none) Message-ID: <8904062252.AA26751@SALT.ACC.COM> Date: 6 Apr 89 22:52:24 GMT Sender: daemon@ucbvax.BERKELEY.EDU Organization: The Internet Lines: 63 Good afternoon all, I thought I'd take a moment or two to sumarize the responses that I received when I sent a query about TN3270 function- ality in a terminal server. In addition to all the "gee, that's a good idea" mail, I did get a few informative notes, mostly from people at the companies that will (hopefully) make these beasts. Here's what I received: >From: "Milo S. Medin" (NASA ARC NSI Project Office) >Encore is very close to having that availible. Next release I'm told... >Milo ******************************************************************* >One of the problems here is mapping the 327X stream into a ascii/ansi terminal >of some type (where some type is one of at least a hundred). Carrying all >this "curses" (forgive me) information around in memory is very wasteful. A >standard protocol might help vendors (and definitely users). This protocol >upon identification of terminal type to the terminal server would ask some >server (who has disk storage) what is the curses defnition for this terminal >to be mapped into. Sounds like an IETF Working group. >Another place to possibly look is how ISO VTP with forms mode does this, knowing >them they probably have left it up to a vendor's agreement. >Marty *************************************************************** >From: William Westfield >cisco plans to do tn3270 in our terminal server. The main reason it hans't >been done yet is that it is HARD. Not technically hard, you understand - >it is just that the models don't mesh well. If you look at the berkeley >tn3270, you find that the majority of the code deals with the user interface; >reading the users config file and figuring out which function keys get mapped >to which PF keys, and so on. This is swell when each user can have a long >".tn3270rc" or whatever file. But terminal servers don't have file systems, >or termcap databases, or users, or even (frequently) terminal types. >This makes things difficult. >Bill Westfield >cisco Systems. **************************************************************** >From swb@chumley.tn.cornell.edu Mon Feb 27 05:47:32 1989 >Xylogics bought Encore's "Annex" product, and the whole development >staff, including Jack O'Neill, the manager, who's a pretty good guy. >Jack had been planning on bringing out 3270 emulation very early this >year. It's been delayed, of course, by the move, but you should call >Jack at Xylogics. ***************************************************************** So, it would seem that folks are working on the product, people agree there's a market, but nothing's running yet. Thanks for all the replies, people really took some time to express opinions and ideas on how to make it all work. If things have changed since I sent this out please let me know. I've got a list of people who are interested(:*) Thanks again, Chris VandenBerg ACC chris@salt.acc.com 805-963-9431