jd3l+@andrew.cmu.edu (Jean-Marc Debaud) (02/19/90)
Keith Perkins (n245bq@tamunix.tamu.edu) writes: >Personally, I'm going to wait for the new NeXT to come out before I make any >judgement. The IBM looks good, but I like the NeXT for other things besides >speed. I like the BSD compatable OS, the DSP chip, the music capability, and >the optical drive - all of which I belive the IBM machines lack. If the new >NeXT is up to date in terms of speed (both optical drive and processor), then >it should be comparable to these new IBMs. All this is quite true. But as for the optical drive, I think the tech. is not up to speed. Why would you need to put a hard disk is the cube otherwise ? True, this is new tech. and is nice but it doesn't work well. So there is nothing to be too exited about. You want to wait for the next NeXT. My memory is failing me here, but I think we may not see it for a while. And with what ? Software version 0.69 ? There is clearly a problem of delays with NeXT. I also strongly doubt that the 68040 is fast enough to match even remotely the entry level Power Station. I have been told that AIX of IBM s... (is so bad) Why is this ? Are you sure it is not compatible with other Unix like BSD ? I agree that NeXT sound chip is nice and IBM lacks it but I am sure they can remedy to this especially with the power they boost. Jean-Marc De Baud Carnegie Mellon jd3l+@andrew.cmu.edu
n245bq@tamunix (Keith Perkins) (02/19/90)
In article <QZriSt600VAEA8Mkhk@andrew.cmu.edu> write: >Keith Perkins (n245bq@tamunix.tamu.edu) writes: > >All this is quite true. But as for the optical drive, I think the tech. >is not up to speed. Why would you need to put a hard disk is the cube >otherwise ? True, this is new tech. and is nice but it doesn't work >well. So there is nothing to be too exited about. Agreed, but it seems that the low speeds of the OD might one day surpass the highest speeds of the current Hard Drives (see the end of the Mac World OD article for more info). >You want to wait for the next NeXT. My memory is failing me here, but I >think we may not see it for a while. And with what ? Software version >0.69 ? There is clearly a problem of delays with NeXT. I also strongly >doubt that the 68040 is fast enough to match even remotely the entry >level Power Station. Actually, the software will probably be at least 1.0a if not 2.0, and while rumors abound there has been no official announcement for a release date, so the new NeXT cannot be delayed. The 68040 can run at 20 MIPS and 3.5 MFLOPS which currently beats both the SPARC chips and the 80486 by a fairly large margin (see Feb Byte, page 96A for more info on the 68040). >I have been told that AIX of IBM s... (is so bad) Why is this ? Are you >sure it is not compatible with other Unix like BSD ? The IBM AIX is a combination/hybrid of BSD 4.3 and System 5. This means that sometimes programs are compatible, and somtimes they're not. There was a review out in either MIPS or Byte which looked at AIX ver 3, and they were seriously impressed. I think that compatiblity could be it's only negative factor, and that with IBM backing AIX, it could go on to rival other UNIX standards. >I agree that NeXT sound chip is nice and IBM lacks it but I am sure they >can remedy to this especially with the power they boost. > > Jean-Marc De Baud > Carnegie Mellon > jd3l+@andrew.cmu.edu Keith Perkins Texas A&M University n245bq@tamunix.tamu.edu
ron@woan.austin.ibm.com (Ronald S. Woan/2100000) (02/19/90)
In article <QZriSt600VAEA8Mkhk@andrew.cmu.edu>, jd3l+@andrew.cmu.edu (J Debaud) writes: |>Keith Perkins (n245bq@tamunix.tamu.edu) writes: |>I have been told that AIX of IBM s... (is so bad) Why is this ? Are you |>sure it is not compatible with other Unix like BSD ? Working for IBM, I am obviously a "little" biased, but having worked on Suns and Apollos most of my life, I'd like to point out AIX 3 is more BSD-like than all previous AIX versions. If I remember correctly, it supports 99.44% of the BSD system calls (a few signals still missing, I believe) while being completely SysVR2 compliant. This is much better than most other so-called "BSD compliant" OSs. Unlike earlier versions of AIX for the IBM RT. AIX 3 also supports full job control (ctrl-Z in the csh or ksh). |>I agree that NeXT sound chip is nice and IBM lacks it but I am sure they |>can remedy to this especially with the power they boost. Remeber you got SCSI support and a micro-channel bus, so adding the DSP will be no problem. If you wait for the new NeXt boxes, pray they use something faster than the 68040. Then again, if they do, you won't be binary compatible, and might as well have switched to IBM. Another plus side of the IBM is the hypertext based documentation (it is really quite nice, basically cross-linked man pages with a Motif interface, bookmarks, notepads). Ron +-----All Views Expressed Are My Own And Are Not Necessarily Shared By------+ +------------------------------My Employer----------------------------------+ + Ronald S. Woan (IBM VNET)WOAN AT AUSTIN, (AUSTIN)ron@woan.austin.ibm.com + + outside of IBM @cs.utexas.edu:romp!auschs!woan.austin.ibm.com!ron + + last resort woan@peyote.cactus.org +
johnm@well.sf.ca.us (john markoff) (02/20/90)
just a note to read the IBM prices with some care. The IBM Powerstation 320 is not a standalone machine. It does not have the capacity to hold all of AIX 3. So figure the cost of a Powerserver into the equation....
dandb@k.gp.cs.cmu.edu (Dean Rubine) (02/20/90)
Keith Perkins (n245bq@tamunix.tamu.edu) (I think it was him) writes: >I agree that NeXT sound chip is nice and IBM lacks it but I am sure they >can remedy to this especially with the power they boost. In <1550@awdprime.UUCP> ron@woan.austin.ibm.com (Ronald S. Woan) replies: >Remeber you got SCSI support and a micro-channel bus, so adding the >DSP will be no problem. To me, the great thing about the NeXT was that it came with all sorts of stuff as standard equipment. As a developer, I'll be much more tempted to write applications which use the NeXT sound chip, since I know there will be one on every machine. This is a marked difference from the typical situation, where I can buy add-on DSP hardware for, say, my IBM or Apple. Since only a few people will buy such hardware, it is much less appealing for software developers to write programs which utilize that hardware, as the potential market is so much smaller. Thus any such software written will likely cost a lot, and/or be of lower quality. -- ARPA: Dean.Rubine@CS.CMU.EDU PHONE: 412-268-2613 [ Free if you call from work ] US MAIL: Computer Science Dept / Carnegie Mellon U / Pittsburgh PA 15213 DISCLAIMER: My employer wishes I would stop posting and do some work.
ron@woan.austin.ibm.com (Ronald S. Woan) (02/20/90)
In article <8059@pt.cs.cmu.edu>, dandb@k.gp.cs.cmu.edu (Dean Rubine) writes: |>In <1550@awdprime.UUCP> ron@woan.austin.ibm.com (Ronald S. Woan) replies: |>>Remeber you got SCSI support and a micro-channel bus, so adding the |>>DSP will be no problem. |> |>To me, the great thing about the NeXT was that it came with all |>sorts of stuff as standard equipment. As a developer, I'll be much |>more tempted to write applications which use the NeXT sound chip, |>since I know there will be one on every machine. I have to agree with you here. Raising the lowest common denominator does indeed make it easier on the developers to use things suchs as the DSP. Unfortunately, there are few main stream applications that require the DSP these days. You might say in that regard that we are not a leader in pushing this type of technology, but you would also have to see our flip side of why should we force people to pay for a DSP when chances are that they won't need it. I had to say voice mail in this day and age can't really cut it as a great mainstream application though it is good for annotations due the the size of digitized sound files. Also consider the DSP as just a quick 24(?)-bit floating point processor and compare our MFLOPS rating (32-bit). When DSP applications become mainstream, as have intensive floating point applications, you will see DSP standard in everyones' machines. As for the 68040, still not as fast as our box (Interger or Floating point). Hopefully, IBM (I have no idea what our exact plans are in this regard except that we have made the announcement that we are commited to it) will be one of the leaders in bringing out the OSF/1/2 OS with a MACH kernel and (golly gee) AIX interface. Ron +-----All Views Expressed Are My Own And Are Not Necessarily Shared By------+ +------------------------------My Employer----------------------------------+ + Ronald S. Woan (IBM VNET)WOAN AT AUSTIN, (AUSTIN)ron@woan.austin.ibm.com + + outside of IBM @cs.utexas.edu:ibmchs!auschs!woan.austin.ibm.com!ron + + last resort woan@peyote.cactus.org +
wb1j+@andrew.cmu.edu (William M. Bumgarner) (02/20/90)
ron@woan.austin.ibm.com (Ronald S. Woan) writes: > In article <8059@pt.cs.cmu.edu>, dandb@k.gp.cs.cmu.edu (Dean Rubine) writes: > |>In <1550@awdprime.UUCP> ron@woan.austin.ibm.com (Ronald S. Woan) replies: > |>>Remeber you got SCSI support and a micro-channel bus, so adding the > |>>DSP will be no problem. > |> > |>To me, the great thing about the NeXT was that it came with all > |>sorts of stuff as standard equipment. As a developer, I'll be much > |>more tempted to write applications which use the NeXT sound chip, > |>since I know there will be one on every machine. > > I have to agree with you here. Raising the lowest common denominator > does indeed make it easier on the developers to use things suchs as > the DSP. Unfortunately, there are few main stream applications that > require the DSP these days. You might say in that regard that we are Of course there aren't any mainstream APPs that require a DSP-- up until the NeXT, there wasn't a gurantee that a machine would have a DSP (and if it did, it was $1000+ on top of system price). You can't have a mainstream app that requires a piece of hardware before the hardware is mainstream also... As far as software is concerned, NeXT is in its infancy-- wait until those applications that are written specifically for the NeXT are released... at that point in time, a proper evaluation of the hardware and system advantages can be made. b.bumgarner | Disclaimer: All opinions expressed are my own. wb1j+@andrew.cmu.edu | I officially don't represent anyone unless I NeXT Campus Consultant | explicity say I am doing so. So there. <Thpppt!> "I ride tandem with the random/Things don't run the way I planned them..."
bbc@legia.rice.edu (Benjamin Chase) (02/21/90)
ron@woan.austin.ibm.com (Ronald S. Woan/2100000) writes: >|>Keith Perkins (n245bq@tamunix.tamu.edu) writes: >|>I have been told that AIX of IBM s... (is so bad) Why is this ? Are you >|>sure it is not compatible with other Unix like BSD ? >Unlike earlier versions of AIX for the IBM RT. AIX 3 also supports >full job control (ctrl-Z in the csh or ksh). Well, thank the Lord. Does your version 3 csh get "$#argv" (notation for "number of arguments passed to the shell") correct, or does it still think that the '#' is the beginning of a comment? I hate to think how many shell scripts use that one piece of notation. "Standard" and "compliant" are very squishy terms when applied to Unix, and even more so its _utilities_. Also, you have to ask yourself why you wanted to have a fully compliant Unix. Because it gave you warm fuzzies? Because it sounds good on marketing sheets? I don't know if NFS is standard BSD Unix, but I'd be very annoyed at a Unix that didn't have it or something as useful. "How about IBM's "Remote Virtual Disk" or whatever they called it?" Don't make me laugh. I could tell that cheap NFS knock-off was a loser just by reading the excellent IBM documentation that came with it. If I had grown accustomed to Mach (perhaps someday I'll have the opportunity...), perhaps I would be unwilling to settle for less than that, too. I wonder if it's "BSD compliant", or if CMU et al had to warp the BSD interface a little to make it better than BSD unix? > If you wait for the new NeXt boxes, pray they use something faster > than the 68040. Then again, if they do, you won't be binary > compatible, and might as well have switched to IBM. Let's look at this more closely. If you bought a NeXT, and are now thinking about buying the (hypothetical) new NeXT: new processor, possibly from a new processor family still has a DSP (same chip?) still has an Optical Disk? probably new versions of the operating system bundled applications, probably new versions definitely new versions of the compilers still uses NeXTStep, probably new version Now lets look at IBM's offerings. Suppose you bought an RT, which might be the case if you're in academia. If we're comparing NeXTs and IBMs, I argue this is a fair restriction, since it is NeXT's (original) market. Maybe IBM even put an RT on your desk for free during this past year. So, what if you get a new Power Series? new processor, totally new family IBM is apparently abandoning the ROMP? new operating system version AIX 2.2.1 is the highest supported version of AIX on your old RT. AIX 3.whatever is the first version for the new IBM boxes. new compilers The America chip set has a totally new instruction format; I'm pretty sure IBM has the only compilers for it. new window system? I really don't know the story here, but I know that historically IBM's support of a particular window system for its hardware has been less than coherent. Sort of like Sun. Sort of unlike NeXT. What really is the difference here? You might as well have _not_ switched to the IBM. Or, if you've got an old IBM, you might as well switch to the new NeXT, since your old IBM and new IBM are about as "binary incompatible" as can be. The two lists above are very inconclusive. I argue that NeXT is much more committed to keeping much of the ancillary parts of the system the same, because that (perceived) guaranteed "minimal" configuration is what will preserve their niche market. ("Minimal" has the wrong connotation, since Jobs' notion of "minimal" appears fairly ample compared to many manufacturers'.) On the next NeXT, you'll probably still have a fairly large set of the hardware toys found on the first NeXT. This encourages sophisticated software that expects and exploits the presence of these hardware toys. Also, in NeXT, you've got a company that is clearly interested in that product line, and the window system that runs on it. In recent history, I see no such coherencey, commitment, and "minimal" configuration from IBM for its academic products. I hate for this posting to be such a diatribe, but you asked for it, Ronald. -- Ben Chase <bbc@rice.edu>, Rice University, Houston, Texas
davef@jessica.Stanford.EDU (David Finkelstein) (02/21/90)
In article <1550@awdprime.UUCP> @cs.utexas.edu:ibmchs!auschs!woan.austin.ibm.com!ron writes: >Another plus side of the IBM is the hypertext based documentation (it is >really quite nice, basically cross-linked man pages with a Motif >interface, bookmarks, notepads). Except it doesn't run under NextStep, and you can't use it to access any other texts (which you can do with the Digital Librarian) and you can't author in it. You can't create your own documents with it. True they may add that functionality in the future, but when they do they better make it so that your original documents didn't have to be written in Interleaf (the hypertext based documentation was first written in Interleaf and brought over). But it does have a lot of potential, and I think IBM understands this University need. David Finkelstein Academic Information Resources Stanford University davef@jessica.stanford.edu
dar@telesoft.com (David Reisner) (02/21/90)
In article <4274@helios.TAMU.EDU>, n245bq@tamunix (Keith Perkins) writes: > ...The 68040 can run at 20 MIPS and 3.5 MFLOPS (All I want is a machine that will do >= 10 MIPS with fast i/o and a separate display processor. (And to think that I used to think that having my own 11/780 would be enough!)) > The IBM AIX is a combination/hybrid of BSD 4.3 and System 5. This means > that sometimes programs are compatible, and somtimes they're not. There > was a review out in either MIPS or Byte which looked at AIX ver 3, and they > were seriously impressed. I think that compatiblity could be its only negative > factor, and that with IBM backing AIX, it could go on to rival other UNIX > standards. Advertisements indicate that IBM's POWER uses OSF/Motif (which appears to be much prettier than OpenLook). My understanding is that OSF has also chosen Mach as their Unix kernel (good news for NeXT). Perhaps (maybe? hopefully?) IBM will move to (an IBM pissed in version of) Mach. -David {uunet,ucsd}!telesoft!dar, dar@sdcsvax.ucsd.edu