zs04+@andrew.cmu.edu (Zachary T. Smith) (07/21/89)
Several people have requested information describing ENIX, a
low-cost version of Unix for the 386, which I mentioned in a few
previous postings.
What follows is a quick, commented rundown of what you get
(or what of it that I remember) for the relatively low price of $610.
Although I now run ENIX on my PC, I should note that I've had limited
time to play with it, and that although I don't mind spending large
sums of money for the kind of machine and software I like, I'm not
a software guru, so I can't answer complicated questions about it.
ENIX is a derivative of AT&T's system V.3.2 for the Intel 80386.
For about $610 (that includes shipping) you get all of the following
as part of the 2-user system with development support:
o UNIX itself
o TCP/IP support (kind of mandatory that you install it; works
even if you don't have a networking card).
o Streams (networking support, I think they call it)
o 2 kilobyte file system support
o a reasonable C compiler, lint, lex, etc.
o a usable, fast 640x480 monochrome X server (VGA).
o some X clients, demo stuff mostly, also xdm.
o twm and uwm (X11 window managers) and sample .twmrc and
.uwmrc files. no awm or gwm, though.
o a unuseably slow 16-color X server (VGA). No sign of super-VGA
support under X.
o both color and mono EGA X server and a big-screen mono X server
(I haven't tried these, I guess they work)
o extended-terminal support (terminfo stuff, I think).
o ftp, rsh, rcp, uucp support (remote file system support, etc.)
o mail support.
o Bourne and C shells.
o system administration and protection stuff.
o nroff (troff may be there too, I haven't checked).
I'm sure there are some things I've forgotten, but you get the idea.
A few things they don't give you include:
o like I said above- a useable color X server.
o a debugger (dbx is not there).
o manual pages, or even the man command.
o bdftosnf: the utility you need to add new fonts to X.
o emacs
o MSDOS disk read/write (you can access the drives, there's just no
support for DOS).
o ability to run MSDOS tasks under Unix. This is special option.
o A real C shell (there's no pushd, popd, or jobs, from what
I can tell). Note, though, that a 'ps' command serves the same
function as the C shell's 'jobs' command, and that aliases and
history -are- included.
o A real 'mv' command (it doesn't relocate directories; there's
a shell script called mvdir to do this).
System speed is reasonable, but not fantastic, on a 20 Mhz 386 with
4 megs of zero wait state RAM (specifically, a Gateway 2000 machine). I'd
say it's about as fast as a VAX 11/780.
A good review of ENIX as well as most of the other major 386 Unix
systems can be found in the Febuary 89 issue of MIPS. Or just ask ENIX for
their broshure, which includes a copy of the review.
Their number is (415) 683-ENIX.
-Zach T. Smith (zs04@andrew.cmu.edu)dave@micropen (David F. Carlson) (07/25/89)
In article <4YlaPoO00WB5QCWGBM@andrew.cmu.edu>, zs04+@andrew.cmu.edu (Zachary T. Smith) writes: > > Several people have requested information describing ENIX, a > low-cost version of Unix for the 386, which I mentioned in a few > previous postings. > Due to lawyers, the product has been renamed ESIX. (ENIX is too close to that variant starting with an "X") > o Bourne and C shells. Of course, a very usable and preferable ksh. Erroneous missing items: > o a debugger (dbx is not there). sdb is included and *is* a debugger. Adb is not included. > o MSDOS disk read/write (you can access the drives, there's just no > support for DOS). mcopy, mdel, mdir, mmd, mrd, mren and mtype will copy, del, make directory, remove directory, rename a file and display a file, respectively, on an MS-DOS disk under ESIX. Current versions will not work on the DOS partition of a hard disk. Floppy only. > > o A real C shell (there's no pushd, popd, or jobs, from what > I can tell). Note, though, that a 'ps' command serves the same > function as the C shell's 'jobs' command, and that aliases and > history -are- included. Use ksh. Jobs is BSD-ism. This is System V remember. > > o A real 'mv' command (it doesn't relocate directories; there's > a shell script called mvdir to do this). See /usr/lib/mv_dir. Works for me. > -Zach T. Smith (zs04@andrew.cmu.edu) What current ESIX is missing (although next rev is schedule for 15 AUG) is SCSI drive support, multiple serial port support, couple of console bugs, and a doggy slow floppy driver (no cacheing). Equivalant set of code from SCO Xenix would price out at $2875! So $610 looks to be quite a bargain. -- David F. Carlson, Micropen, Inc. micropen!dave@ee.rochester.edu "The faster I go, the behinder I get." --Lewis Carroll
goer@sophist.uucp (Richard Goerwitz) (07/25/89)
In article <803@micropen> dave@micropen (David F. Carlson) writes: >Due to lawyers, the product has been renamed ESIX. This kind of thing, while a bit off-the-cuff, still bothers me. Remember: Lawyers don't sue people. People sue people. -Richard L. Goerwitz goer@sophist.uchicago.edu rutgers!oddjob!gide!sophist!goer
madd@bu-cs.BU.EDU (Jim Frost) (08/08/89)
In article <803@micropen> dave@micropen (David F. Carlson) writes: |In article <4YlaPoO00WB5QCWGBM@andrew.cmu.edu>, zs04+@andrew.cmu.edu (Zachary T. Smith) writes: |Erroneous missing items: | |> o a debugger (dbx is not there). | |sdb is included and *is* a debugger. Adb is not included. We tried it and found it lacking. It's better than nothing, but not much. When bringing up GNU emacs, a bug in the optimizer causes fns.c to compile incorrectly. This exists in all variants of AT&T 386 UNIX which use the AT&T compiler. Under Interactive we were able to track down the problem (mostly -- a helpful ISC employee aided us, too). Under ENIX (*was* called ENIX then) the debugger just didn't work. |> o A real 'mv' command (it doesn't relocate directories; there's |> a shell script called mvdir to do this). | |See /usr/lib/mv_dir. Works for me. Should be called automagically by mv. /usr/lib isn't in my path, nor do I believe it should be. |Equivalant set of code from SCO Xenix would price out at $2875! So $610 |looks to be quite a bargain. Yes ... and no. For the price there is nothing else, but I'd recommend Interactive even though a really workable system will cost about $900-$1000 (a lot less if you don't care about X windows or networking; more if you need NFS). We didn't get ENIX up and running on a network so no comments on that. We did try running X windows and found their version abysmally slow and missing a variety of necessary X utilities (xset comes to mind; if not that, something equally important). Looks like it was a quick-and-dirty port and they just didn't bother including anything that didn't work right off the bat. In contrast, the Interactive X was remarkably fast (beat the Sun 386i version and was running on a much slower processor, under SysV, and in less memory). Interactive even went so far as to include X utilities off the net which were not in the X11R3 release. We never bothered to try the XENIX X since it was R2. As a development system, I recommend Interactive over both ENIX and XENIX. I expect ENIX will grow into something more usable, but in its current state it's pretty tough to develop on. XENIX is, well, XENIX, with all that implies in terms of strangenesses, compiler problems, etc. One last thing to note: XENIX will run beautifully in 4Mb; the others really want 6Mb or more. But for the price difference you can get quite a lot of memory. jim frost software tool & die madd@std.com
davidsen@sungod.crd.ge.com (ody) (08/09/89)
In article <36041@bu-cs.BU.EDU> madd@bu-it.bu.edu (Jim Frost) writes: | As a development system, I recommend Interactive over both ENIX and | XENIX. I expect ENIX will grow into something more usable, but in its | current state it's pretty tough to develop on. XENIX is, well, XENIX, | with all that implies in terms of strangenesses, compiler problems, | etc. It depends on what your target is... if you need to run in UNIX and MS-DOS the Xenix cross compiler will save you a lot of time. You can use one make file to roll both versions, and Xenix/286, too. That's important if you resell (or give away) your software. And the new development set has the AT&T compilers included. The MSC compiler also supports a lot of ANSI stuff not in the IN/ix compiler (as of June). I don't think you can broadly say one is "the best," because they all have advantages. | | One last thing to note: XENIX will run beautifully in 4Mb; the others | really want 6Mb or more. But for the price difference you can get | quite a lot of memory. The street price is about $300 more for Xenix. That will just about buy the 2 extra MB. Before I bounght Xenix I got evaluation copies of Xenix, IN/ix, and MicroPort. MP wouldn't run serial ports without double panic, and both systems with the AT&T compiler failed on some simple programs (the compiler generated asm source using registers not in the 386). I chose Xenix because it was the most solid I could get. If I could wait until October I would look at Open Desktop. With TCP (including a good SLIP), X11R3, DOSmerge, NFS, and a database for <$1k list, it seems like a really good deal. I don't know what the development kit will hold, but it should include the 3.2 stuff, which is MSC and AT&T both. I also assume it will have the fast filesystem from 3.2 (any comments, SCO?) which is good. I am happier giving facts and letting people decide what fits them. Obviously I have been using SCO for many years, but I also have used MP and tried IN/ix, and they are both a good fit for some people. The recent issue of _MIPS_ on disk controllers and UNIX on 386's is relevant to a decision, also. bill davidsen (davidsen@crdos1.crd.GE.COM) {uunet | philabs}!crdgw1!crdos1!davidsen "Stupidity, like virtue, is its own reward" -me
palowoda@megatest.UUCP (Bob Palowoda) (08/10/89)
From article <1539@crdgw1.crd.ge.com>, by davidsen@sungod.crd.ge.com (ody): > In article <36041@bu-cs.BU.EDU> madd@bu-it.bu.edu (Jim Frost) writes: > > | As a development system, I recommend Interactive over both ENIX and > | XENIX. I expect ENIX will grow into something more usable, but in its > | current state it's pretty tough to develop on. ^^^^^^^^^^^^^^^^^^^ Please elaborate. XENIX is, well, XENIX, > | with all that implies in terms of strangenesses, compiler problems, > | etc. No argument here. > > It depends on what your target is... if you need to run in UNIX and > MS-DOS the Xenix cross compiler will save you a lot of time. You can use > one make file to roll both versions, and Xenix/286, too. That's > important if you resell (or give away) your software. And the new With alot of the dos software that makes extensive use of graphics libs that are not available on xenix I don't think this is a good route to go. Another alternative is to use Merge, VPIX, SimulTask and if you can run both compiliers. > | > | One last thing to note: XENIX will run beautifully in 4Mb; the others > | really want 6Mb or more. But for the price difference you can get > | quite a lot of memory. Depends which way the market prices are going. Right now ram is dropping and the price of a UNIX OS has be on the increase. and both systems with the AT&T compiler failed on some > simple programs (the compiler generated asm source using registers not > in the 386). Example? > > If I could wait until October I would look at Open Desktop. With TCP > (including a good SLIP), X11R3, DOSmerge, NFS, and a database for <$1k Be very carefull here. This is the price for a one user efficient. You will notice to move up to the next level (a multi machine enviornment with networking) that will enable two or more persons on the network to use the software is going to cost approx $1500.00 MORE. Than there is one more price package that will kick it above that for devolpment. More or less there price structure looks to be just a bit higher than ISC's. The plus is the database I guess. The ESIX product is about half this price. > > I am happier giving facts and letting people decide what fits them. > Obviously I have been using SCO for many years, but I also have used MP > and tried IN/ix, and they are both a good fit for some people. That reminds me. What's with Dell Computers with it's UNIX and Xwindow OS it's offering? It claims to have the Motiff user interface etc. Is this really SCO's UNIX? Anybody check up on it? > > The recent issue of _MIPS_ on disk controllers and UNIX on 386's is > relevant to a decision, also. Check out the new Cacheing controller by Mylex! 15Mb/sec? How did they obtain this number? ---Bob -- Bob Palowoda *Home of Fiver BBS* login: bbs Work: {sun,decwrl,pyramid}!megatest!palowoda Home: {sun}ys2!fiver!palowoda (A XBBS System) 2-lines BBS: (415)623-8809 2400/1200 (415)623-8806 1200/2400/9600/19200
fyl@fylz.UUCP (Phil Hughes) (08/11/89)
In article <6972@megatest.UUCP>, palowoda@megatest.UUCP (Bob Palowoda) writes: > From article <1539@crdgw1.crd.ge.com>, by davidsen@sungod.crd.ge.com (ody): > > In article <36041@bu-cs.BU.EDU> madd@bu-it.bu.edu (Jim Frost) writes: > > It depends on what your target is... if you need to run in UNIX and > > MS-DOS the Xenix cross compiler will save you a lot of time. You can use > > one make file to roll both versions, and Xenix/286, too. That's > > important if you resell (or give away) your software. And the new > With alot of the dos software that makes extensive use of graphics > libs that are not available on xenix I don't think this is a good > route to go. Another alternative is to use Merge, VPIX, SimulTask > and if you can run both compiliers. After trying a few options, I ended up with ENIX on a 386 as the "real" system and an XT with DOS 2.10 and a used copy of Turbo C 1.5 as the "compile and test it for DOS" platform. I found that VP/ix under Xenix would crash with my application that works fine on DOS. Also, as the product is designed to run on the minimum of DOS systems, having a "minimum" to actually test it on is helpful. The development and testing is done with the ENIX system and all the real source is maintained there as well as the distribution system. The only thing different is the Borland style make file. -- Phil Hughes -- FYL -- 8315 Lk City Wy NE -- Suite 207 -- Seattle, WA 98115 {amc-gw,uunet!pilchuck}!ssc!fylz!fyl