Sun-Spots-Request@RICE.EDU (William LeFebvre) (05/19/88)
SUN-SPOTS DIGEST Wednesday, 18 May 1988 Volume 6 : Issue 92 Today's Topics: Re: 4.0-notes on archive server Re: Looking for Sun APL Re: XY451 firmware Re: Sun 4/110 memory Re: gammontool (2) ho termcap entry & security question A SunView Interrupt Button (any suggestions) Sun OS 4.0 opinions wanted 875 Meg Disk yet? 3rd party SCSI disks for 386i ? Device Drivers on SunOS 4.0? Send contributions to: sun-spots@rice.edu Send subscription add/delete requests to: sun-spots-request@rice.edu Bitnet readers can subscribe directly with the CMS command: TELL LISTSERV AT RICE SUBSCRIBE SUNSPOTS My Full Name Recent backissues are available via anonymous FTP from "titan.rice.edu". For volume X, issue Y, "get sun-spots/vXnY". They are also accessible through the archive server: mail the request "send sun-spots vXnY" to "archive-server@rice.edu" or mail the word "help" to the same address for more information. ---------------------------------------------------------------------- Date: Mon, 16 May 88 10:06:43 EDT From: wyatt%cfa@harvard.harvard.edu (Bill Wyatt) Subject: Re: 4.0-notes on archive server The shar file containing the 4.0-notes.ms file (received from archive-server@rice.edu) actually had two copies in it, thus doubling the 28k file length. I suspect this is incorrect... [[ Something must have gone wrong en route to you. The copy in the archives is 28K long and (as far as I can tell) contains only one copy of the document. Or perhaps you requested it twice: once in the subject field and once in the body! The archive server does consider the subject field to be a candidate for a valid command. --wnl ]] ------------------------------ Date: Mon, 16 May 88 12:46:06 EDT From: mjab@think.com Subject: Re: Looking for Sun APL I don't know about any public domain APL's for the SUN APL's for the SUN that cost money include (in order of how much I know about them): Sharp APL/UX I.P. Sharp Associates 1200 First Federal Plaza, Rochester NY 14614 Attn: David G. Smith (716) 546-7270 this is a very full-featured APL with Nested Arrays, Event handling, Shared Variables, and advanced operators based on Iverson's "Dictionary of APL." The only real drawback is that the interface with the rest of the UNIX world is non-standard. This product is based on the one below. STSC APL Same as the above but without the Dictionary Operators and with a different event handling system and possibly without shared variables (I'm not sure). The nested arrays in the STSC product follow IBM's conventions of "Strand Notation." This would be a good choice if compatibility with IBM's APL-2 is a factor. Suffers from the same lack of a standard Unix interface. STSC Inc. 2115 East Jefferson St. Rockville, MD 20852 Attn: Stephanie Crognale (301) 984-5444 There are probably other Sun implementations that I don't know about such as Dyalog APL from England. If the Public Domain aspect is more important than the Sun aspect, I know of the following *Sharp APL/PC for IBM PC and close compatibles. An IBM 370 APL and an emulator on which to run it. Not speedy, but full featured. Available from on-line networks and from I.P. Sharp. *I-APL for Sinclair QL, IBM PC, BBC Micro, and (I think) Apple II. This is a new implementation meant for use in schools. It is available on electronic bulletin boards and from: Edward M. Cherlin 6611 Linville Drive Weed, CA 96094 (916) 938-4684 *APL-90 For Apple Macintosh SYNC 12, Place Hotel di Ville 42000 Saint-Etienne France Contact J.J. Giradot 011-33-77-32-65-62 ------------------------------ Date: Mon, 16 May 88 11:01:22 PDT From: brian@ucsd.edu (Brian Kantor) Subject: Re: XY451 firmware Some have reported problems with more than one Eagle drive giving intermittant "sequencer error" and other such messages. We ran into these errors with the XY450 controller. Upon calling Xylogics and finally finding the person who knew the answers, we found that we could buy the proper firmware for a moderate price - I think it was $50 or $100 or something like that. The explanation I remember was that the total number of bytes per sector (including sync, preamble, header, data preamble, and postamble) required by the controller was slightly more than the 600 bytes the Eagle provides. I think the chips they sent were a new ROM and a new sequencer PAL. Clearly you should call Xylogics and ASK. Brian Kantor UC San Diego ------------------------------ Date: Mon, 16 May 88 11:51 CDT From: David F. Dow <dow@mcc.com> Subject: Re: Sun 4/110 memory Postal-Address: 3500 W. Balcones Center Drive, Office #3.11108 Austin, TX 78759 Business-Phone: 338-3777 I'd like to clarify the 4/110 memory issue. Clearpoint does offer this memory. The following is from one of their information sheets. Other companies probably do (or will) offer this memory as well. >Clearpoint seems to be looking at about $500-$600/M purchased in >quantities of 8M. The price seems to be fluctuating a lot, but this is about the range I've seen. Minimum purchase is 16M. You can't use 8M by itself anyway. >There is some indication that you can mix types so you >could have 8M of 1M simms and fill your other slots with 256k simms >for a total of 12M. There are two banks of 16 SIMMS on the 4/110 CPU board. Each bank must contain like memory, but the two banks can have different types, or one bank can be left empty. Thus, here are the possible configurations: 8M = 32 X (1/4)MB SIMMS 16M = 16 X 1MB SIMMS 20M = 16 X (1/4)MB SIMMS + 16 X 1MB SIMMS 32M = 32 X 1MB SIMMS The board can be jumpered to configure 4MB with 1/4M SIMMS or 8MB with 1M SIMMS, but Sun says this is for diagnostic purposes only. Note that whichever upgrade path you choose, you will end up discarding some of the 1/4MB SIMMS. Perhaps you can order another 4/110 without any memory (fat chance). I'm hoping that Sun will offer some kind of upgrade path like they do with 260 memory, although Sun claims they will not. I am not affiliated in any way with Clearpoint. I asked these questions when we ordered our 4/110s and they were the only company that could answer them. Hope this helps. -- David F. Dow ------------------------------ Date: 16 May 88 13:40:51 GMT From: Rich Salz <rsalz@bbn.com> Subject: Re: gammontool (1) I don't have access to 3.4 or 3.5 sources, but assuming they haven't changed it from 3.2, the program plays fair. Writing a good cheater is very hard to do. N.B.: the folks at UCB released their backgammon game to the Usenet newsgroup comp.sources.games; it's archived on uunet.uu.net, among other places. (Don't ask me about it.) /rich $alz Please send comp.sources.unix-related mail to rsalz@uunet.uu.net. [[ A very simple way that it *could* cheat (I'm not saying that I think it actually *does* this) is to only allow you to roll unusable rolls when you have a man on the bar. It would be a little harder to get the program to give itself needed rolls, I guess. As for the UCB backgammon, the two games (gammontool and UCB backgammon) may have had identical origins, but afterplaying both I can definitely say that they do not use the same playing strategy! That leads me to believe that they are not the same game. --wnl ]] ------------------------------ Date: Mon, 16 May 88 10:59:03 CDT From: wca@emx.utexas.edu (William C. Anderson) Subject: Re: gammontool (2) Gammontool players: Surely not all of the SunSpots gammonfool players have missed the bugs in gammonfool which allow a decent player to pound him into submission, to wit: 1) Gammonfool evidently has poor recognition of his opponent's defensive position (this is probably inherited from the BSD backgammon program, which was recently "declassified" and posted to the net). In particular, gammonfool doesn't recognize a "6 prime" defensive position. In a game where you have a little "shape" and have set up a 6 prime pinning gammonfool in your home board, and where the 'fool has a pip count advantage, start doubling! The 'fool will return the doubles until the game value is 64. Then, roll home your 6 prime and take the 64 games (or 128, if you get some rolls) to the bank. 2) Gammonfool accepts stupid doubles in the endgame (just like BSD backgammon). For example, assume that you have the cube (or that it hasn't been offered) and that both you and the 'fool have men on your respective 1 and 2 spikes, AND that the roll is yours. You are sure to win. Double the 'fool and he accepts, since his only criterion for accept/reject is evidently pip count, and pip count is even! This "feature" alone should allow you to stay ahead of the 'fool in games. Of course, these bugs take a lot of the fun out of gammonfool. I am a decent player (OK, so I've read Magriel once, but it was a long time ago) but am by no means a great player. I haven't played the 'fool in quite some time (the thrill is gone), but my score file says that the game count is Myself: 887 and the 'Fool: 457. Oh, that I could do so well in Vegas! Somehow, I doubt that these bugs have a high enough priority to be "fixed in 4.0". William Anderson - University of Texas Computation Center - wca@emx.utexas.edu [[ Under what version of the operating system did you last play gammon[ft]ool? --wnl ]] ------------------------------ Date: Mon, 16 May 88 09:57:30 edt From: mlijews@nswc-wo.arpa Subject: ho termcap entry & security question Digging around in the man pages for kbd, I found that the Sun monitors do have a home sequence. The proper entry to add to /etc/termcap is: ho=\ESC[H Just thought someone might be interested. I also have a question about how people give access to their Sun's, which are normally kept locked via lockscreen, to visitors without giving away passwords. We recently had an episode here where a visitor, when he couldn't get onto the Sun due to the screen being locked and the owner being away for the day, did an L1-A on it! It just so happened that the owner had a multi-day calculation running at the time, which was promptly lost. The question: how do people let other people use their Sun, without giving up their password and still maintain a reasonable level of security? Mike Lijewski (mlijews@nswc-wo.arpa) Applied Math Branch NSWC Silver Spring, MD 20903 ------------------------------ Date: Mon, 16 May 88 10:47:39 EDT From: csrobe@icase.arpa (Charles S. Roberson) Subject: A SunView Interrupt Button (any suggestions) I have a SunView application that is performing a simulation which does some integration and (hopefully) eventually converges and stops. As the simulation is running it displays some graphical information about the progress of the integration. The actually performance of the simulation depends on many parameters and the input data -- as a result, sometimes we want to interrupt the routine doing the integration (e.g. we see that it is not going to converge or we realize that we munged a parameter) w/o aborting the SunView process. My question is this, is there some way that I can define an INTERRUPT button that when depressed takes control from the rogue routine and returns it to the window_manager? Any and all help will be greatly appreciated! Thanks, -chip Chip Roberson ARPANET: csrobe@icase.arpa 1105 London Company Way BITNET: $csrobe@wmmvs.bitnet Williamsburg, VA 23185 UUCP: ...!uunet!pyrdc!gmu90x!wmcs!csrobe ------------------------------ Date: 16 May 88 16:20:56 GMT From: woods@handies.ucar.edu (Greg Woods) Subject: Sun OS 4.0 opinions wanted We have to perform an upgrade to at least 3.5 on one of our local Sun systems (there is something in there that the graphics project needs that is not in earlier versions). Since we also do a lot of communicating on the Internet, we would really like to go to a resolver-based kernel (that causes ALL gethostby*() calls to query the name server instead of /etc/hosts). The only system Sun has that does this is 4.0. My boss claims to have a source inside Sun that says that 4.0 is unstable under high loads (a frequent condition around here). So, I am looking for people that have actual experience with 4.0. How does it perform under high loads? Is it full of bugs or relatively bug-free? (I understand it is an almost complete kernel rewrite rather than a simple upgrade). If you answer this, send it to both sun-spots@rice.edu (no doubt many readers of this group are also interested), and also please to me personally (woods@ncar.ucar.edu), as we need feedback quickly as a decision on which way to go here must be made in the next few weeks, and I'd really like to avoid either an unnecessary extra system upgrade or installing an unstable system (which we can't afford on the production machine in question). Thanks for your input. --Greg [[ Two things I'd like to point out. One: to my knowledge 4.0 is still in beta test. It is due out real soon now, but beta test sites are under an agreement to not say anything to the general public about their experiences. What I'm saying is that I doubt anyone can answer your question just yet. ---BULLETIN: after typing that last comment, I received notification from one of our readers that he has received a real copy of version 4.0. His message will appear in the next issue.--- Two: (and maybe you know this) the kernel itself has no say in how host names get translated into addresses. That's all done in the libraries. The name resolver package comes with a library that includes its own gethostby* calls. Unfortunately, it requires that you include a different netdb.h file. If you can get source for all your favorite network programs then you can recompile and link with the resolver library and it will use the resolver. Also, I believe that there is a way to get the yp-server equivalent of /etc/hosts to check the name resolver if it can't find the host in its database. As you say, this will all be much better in 4.0 --wnl ]] ------------------------------ Date: 16 May 88 17:14:23 GMT From: km@emory.UUCP (Ken Mandelberg) Subject: 875 Meg Disk yet? Has Sun started delivering systems with the new 875 Meg disk option yet? It was announced several months ago, but what I hear from our local people it is still in Beta test. The information I have is that Sun is in principle using both NEC 2363s and Hitachi DK815-10. The NEC is larger and is more available locally, and I am considering adding one obtaind from a third party. In the past Sun has been willing to maintain customer supplied drives as long as they coincide with drives they already maintain and satisfy certain conditions (correct revision, Sun cables, etc). In this case I am being told they won't take the NEC drive under contract. Has anypne else tried? Is anyone running the NEC drive on Sun yet? Ken Mandelberg | {decvax,sun!sunatl,gatech}!emory!km UUCP Emory University | km@emory BITNET Dept of Math and CS | km@emory.ARPA ARPA,CSNET Atlanta, GA 30322 | Phone: (404) 727-7963 ------------------------------ Date: Mon, 16 May 88 13:26:21 MDT From: lyndon@ncc.nexus.ca (Lyndon Nerenberg) Subject: 3rd party SCSI disks for 386i ? We have a 386i/250 on order and are wondering what to do about disks for this machine. Sun's prices for the large disk option seem a bit high. We currently run a pair of CDC Wren IV's on our 3/280S, and are very impressed with the performance. Does anyone know of any changes to the SCSI interface on the 386i that would prevent us from moving one of the Wrens over? (The 386i will have a Sun supported 90MB drive so installation won't be a problem. We would like to use the Wren for the user file system) Lyndon Nerenberg Nexus Computing Inc. {alberta,uunet}!ncc!lyndon || lyndon@ncc.Nexus.CA ------------------------------ Date: 16 May 88 20:28:43 GMT From: woods@handies.ucar.edu (Greg Woods) Subject: Device Drivers on SunOS 4.0? We have several local device drivers (for things like the CRAY UNIX Station and a hyperchannel network), some of which have even been locally written or locally hacked. Has anyone gotten any non-Sun-supported device drivers that they had been using under SunOS 3.* to work on 4.0? How hard was it to do? What kinds of changes did you have to make? --Greg (woods@ncar.ucar.edu) ------------------------------ End of SUN-Spots Digest ***********************