hdtodd@eagle.wesleyan.edu (12/10/90)
Are there any rumors out there as to when/if SVR4 will be released for Towers?
bminnebo@vela.acs.oakland.edu (Brian P. Minnebo) (12/10/90)
In article <1990Dec9.150131.36596@eagle.wesleyan.edu> hdtodd@eagle.wesleyan.edu writes: > Are there any rumors out there as to when/if SVR4 will be released for >Towers? I asked my local SE a similar question, I wanted to know if the tower would ever have an Version of UNIX that had streams based tty drivers. He said he would check, came back a few days later and said that it would probably never happen. However, he did mention the new line... -- Brian Minnebo (Chrysler Motors MIS Tech Services) Oakland University: bminnebo@vela.acs.oakland.edu Chrysler Corp: ... uunet!umich!sharkey!cfctech!techsys!brian
nolan@iti.org (Mike Nolan) (12/10/90)
According to the latest info from NCR, SVR4 will *NOT* be ported to the Tower, but only to the System 3000 intel-based (80386 and 80486) platforms. This sharply contradicts what NCR was saying as recently as last May, when they were saying that SVR4 would be ported to the Tower, but only to the models using the 68030 or 68040 processor. If you're on a smaller Tower, you're SOL in either event. Of course, if AT&T succeeds in its takeover offer, that could easily change! Michael Nolan UUCP tssi!nolan (or possibly sparky!dsndata!tssi!nolan) Internet: nolan@helios.unl.edu
doug@letni.UUCP (Doug Davis) (12/11/90)
In article <nolan.660795858@helios> nolan@iti.org (Mike Nolan) writes: >According to the latest info from NCR, SVR4 will *NOT* be ported to the >Tower, but only to the System 3000 intel-based (80386 and 80486) platforms. >This sharply contradicts what NCR was saying as recently as last May, when >they were saying that SVR4 would be ported to the Tower, but only to the >models using the 68030 or 68040 processor. If you're on a smaller Tower, >you're SOL in either event. Perhaps I could shed a little light on this, as a developer of NCR tower clones (yes, not the peecee ones) we committed to SVR4 back in December of last year. The M68k release (from at&t) was expected sometime in October. To this date we haven't seen it, nor does it look like it will be any time real soon. Giving that the '86 code has been available for a while it is quite likely that NCR is only talking about SVR4 for '86 machines, since that code is in a known state. Yes, you can port the '86 code to the M68k platform, but that is non-trivial and most M68k vendors would rather start thier unixs from the base port. (Besides the fact nowdays you get to pay at&t twice for having a source license for more than one processer) doug __ Doug Davis/4409 Sarazen/Mesquite Texas, 75150/214-270-9226 {texsun|lawnet|smu}!letni!doug doug@letni.lonestar.org "Be seeing you..."
nolan@iti.org (Mike Nolan) (12/11/90)
doug@letni.UUCP (Doug Davis) writes: >Perhaps I could shed a little light on this, as a developer of >NCR tower clones (yes, not the peecee ones) we committed to SVR4 >back in December of last year. The M68k release (from at&t) was expected >sometime in October. To this date we haven't seen it, nor does it >look like it will be any time real soon. Giving that the '86 code >has been available for a while it is quite likely that NCR is only >talking about SVR4 for '86 machines, since that code is in a known state. According to the 'usually reliable' inside sources, NCR had already succeeded in porting enough of SVR4 to the 68k machines that they had working prototypes of their recently announced Cooperation software package running last spring. It is my understanding that the decision NOT to release SVR4 on 68k machines is a marketing decision, not an engineering one. (By contrast, the decision not to release the 88k systems was based on engineering results. According to at least one source, NCR built about 50 of the 88k machines, but the 80486 machines ran rings around them in performance comparisons.) Michael Nolan UUCP: tssi!nolan (sometimes sparky!dsndata!tssi!nolan is more successful) Internet: nolan@helios.unl.edu
TEMNGT23@ysub.ysu.edu (Lou Anschuetz) (12/11/90)
In article <nolan.660852526@helios>, nolan@iti.org (Mike Nolan) says: >It is my understanding that the decision NOT to release SVR4 on 68k machines >is a marketing decision, not an engineering one. (By contrast, the decision ^^^^^^^^^ What - they don't WANT user support for their product? This sounds like the not-very-good-decision to use Wollongong TCP/IP.... Say it ain't so Michael. >Michael Nolan >UUCP: tssi!nolan (sometimes sparky!dsndata!tssi!nolan is more successful)
borcher@cs-col.Columbia.NCR.COM (Tom Borcher) (12/12/90)
In article <1990Dec9.150131.36596@eagle.wesleyan.edu> hdtodd@eagle.wesleyan.edu writes: > > Are there any rumors out there as to when/if SVR4 will be released for >Towers? yes, lot's of them
haug@almira.uucp (Brian R Haug) (12/12/90)
In article <4172@vela.acs.oakland.edu> bminnebo@vela.acs.oakland.edu (Brian P. Minnebo) writes: >In article <1990Dec9.150131.36596@eagle.wesleyan.edu> >I asked my local SE a similar question, I wanted to know if the tower >would ever have an Version of UNIX that had streams based tty drivers. >He said he would check, came back a few days later and said that it >would probably never happen. However, he did mention the new line... Strictly speaking this information is incorrect. The TOWER32/700 OS Releases use a streams based tty subsystem. Your SE is probably correct for the TOWER32/8xx, TOWER32/[46]xx systems. I don't know whether the TOWER32/200 has a streams based tty subsystem or not. Share and Enjoy! Brian
rey@safn2.UUCP (rey) (12/12/90)
In article <90344.175456TEMNGT23@ysub.ysu.edu>, TEMNGT23@ysub.ysu.edu (Lou Anschuetz) writes: > This sounds like the not-very-good-decision to use Wollongong TCP/IP.... I am sure networking is a complex topic not easily addressed in a short article; but I have 2 Towers and would like to network them. I think networking sounds like the best way to grow computer power. I need better than that thought however; to justify spending $10k to get there. I have been reading on the subject; but can find no one who knows anything about networking willing to educate (sell) me to the point where I could make a proposal to my company. I guess I could buy a Sun for $10k. The vendors I have heard of are Banyan, Ftp, Lachman, Wollongong, Excellan. I sure would like to see a post or reference discussing the merits of the various products. Tower Net has been around for years. What is it? I remember Greg Hackney posting some info on NFS on Towers. Had to do with a v2 to v3 upgrade. Please do not refer me to my NCR Salesperson; The ones who know the product cannot fool with orders of quantity 1. -- Reynolds McClatchey (Southern Aluminum Finishing Co, Atlanta, GA, USA) Architectural Aluminum. Custom Fabrication. Paint, Powder Coating, Anodizing. uunet!safn2!rey MCImail 414-2935
TEMNGT23@ysub.ysu.edu (Lou Anschuetz) (12/12/90)
In article <773@safn2.UUCP>, rey@safn2.UUCP (rey) says: > >In article <90344.175456TEMNGT23@ysub.ysu.edu>, TEMNGT23@ysub.ysu.edu (Lou >Anschuetz) writes: >> This sounds like the not-very-good-decision to use Wollongong TCP/IP.... > >The vendors I have heard of are Banyan, Ftp, Lachman, Wollongong, Excellan. >I sure would like to see a post or reference discussing the merits >of the various products. Tower Net has been around for years. > (lots o' stuff deleted) Let me clarify my earlier point a bit. NCR now supplies a truly excellent Ethernet board by Excelan. NCR also used to get their software from Excelan, but for some unknown reason now uses software from Wollongong. The problem is that Wollongong software is host based rather than using the intelligent ethernet card from Excelan. This means that some things execute VERY SLOWLY (ie: a "netstat -r" on my tower 32/700 with 1-2 users can take over 30 MINUTES! CODAR indicates this is correct.) Also, a lot of basic functions, like the ability to send mail to sites with MX records, don't exist. The Wollongong software is sort of barebones ethernet. Since we are increasing our dependency on the ethernet (and doing away with HPSIO board connections via a terminal server), we are seriously considering replacing our ethernet components as well. At the present time, the best option appears to be "Sea Change Corporation" in Ontario. They have a product called TCP/IP+ for the tower family that uses an intelligent ethernet controller. Their product comes from CMC, which is also a fairly large supplier of network components. Note, however, that we have not tried this product yet, so cannot really comment on functionality, only on the literature. I am not affiliated in any way with NCR or Sea Change, just an average user of the tower, which I must say we otherwise like very much!. It is just a mystery to us how some of these decisions get made..... Lou Anschuetz temngt23@ysub.ysu.edu root@yfn.ysu.edu (at home on the tower)
wescott@Columbia.NCR.COM (Mike Wescott) (12/13/90)
In article <nolan.660852526@helios> nolan@iti.org (Mike Nolan) writes: > According to the 'usually reliable' inside sources, NCR had already succeeded > in porting enough of SVR4 to the 68k machines that they had working prototypes > of their recently announced Cooperation software package running last spring. I think that overstates the condition of that development effort, especially given the quality of the SysVr4 sources at that time. > It is my understanding that the decision NOT to release SVR4 on 68k machines > is a marketing decision, not an engineering one. Actually, it was a bit of both. Given a scarcity of engineering resources an a need to get to market early with a new product. > (By contrast, the decision > not to release the 88k systems was based on engineering results. According > to at least one source, NCR built about 50 of the 88k machines, but the 80486 > machines ran rings around them in performance comparisons.) Nope. -- -Mike Wescott mike.wescott@ncrcae.Columbia.NCR.COM
nigel@cnw01.storesys.coles.oz.au (Nigel Harwood) (12/13/90)
In article <4172@vela.acs.oakland.edu>, bminnebo@vela.acs.oakland.edu (Brian P. Minnebo) writes: > > I asked my local SE a similar question (re: SVR4), I wanted to know if the tower > would ever have an Version of UNIX that had streams based tty drivers. > He said he would check, came back a few days later and said that it > would probably never happen. However, he did mention the new line... SVR3 is streams based. -- <<<<<<<<<<<<<<<<<<<<<<<<< Nigel Harwood >>>>>>>>>>>>>>>>>>>>>>>>>>> << Post: Coles Myer Ltd, PO Box 2000 Tooronga 3146, Australia >> << Phone: +61 3 829 6090 E-mail: nigel@cnw01.storesys.coles.oz.au >> <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
root@mtiame.mtia.oz (Stephen Boucher) (12/13/90)
In article <773@safn2.UUCP> rey@safn2.UUCP (rey) writes: >I am sure networking is a complex topic not easily addressed in a short >article; but I have 2 Towers and would like to network them. I think >[...] >The vendors I have heard of are Banyan, Ftp, Lachman, Wollongong, Excellan. >I sure would like to see a post or reference discussing the merits >of the various products. Tower Net has been around for years. What is it? >[...] We use an ethernet board from CMC with their TCP+ package which runs the lower levels of the TCP/IP suite on the board to take the load of The Tower CPU. We run X11R4 to X-terminals from three distinct vendors through it with no problems at all (other than X11 memory requirements). I can't say first hand how well the package performs between multiple Towers as we only have one Tower per site at present, but it would likely be as good. Our supplier for this (and other quality products for Towers) is: Sea Change Corp. 6695 Millcreek Drive, Unit 8 Mississauga, Ontario, Canada L5N 5R8 Tel: (416) 542 9484 Fax: (416) 542 9479 The company president John Alsop is on the net at: uunet!seachg!jalsop Disclaimer: Sea Change don't pay us, we pay them. Jesper Peterson (regardless of what the header and .sig might say) -- -------------------------------------------------------------------------------- ACSnet: root@mtiame.mtia.oz "If you love something,set it free, UUCP: ...!uunet!munnari!mtiame.oz!root if you don't, PHONE: (03) 699-1022 buy it a John Saul novel"
syackey@secola.Columbia.NCR.COM (Steve Yackey) (12/14/90)
>The problem is that >Wollongong software is host based rather than using the intelligent >ethernet card from Excelan. people who wanted their host to behave as a gateway/router, wanted more connections (RAM) or wanted higher performance might consider host based a solution rather than a problem.
bminnebo@vela.acs.oakland.edu (Brian P. Minnebo) (12/17/90)
In article <1144@cnw01.storesys.coles.oz.au> nigel@cnw01.storesys.coles.oz.au (Nigel Harwood) writes: >SVR3 is streams based. On the NCR Tower 600/400, the tty device driver is not streams based, although streams are there for other things. -- Brian Minnebo (Chrysler Motors MIS Tech Services) Oakland University: bminnebo@vela.acs.oakland.edu Chrysler Corp: ... uunet!umich!sharkey!cfctech!techsys!brian