dbjag@io.UUCP (David Benjamin x4050) (01/12/89)
Does anyone have a method of transferring files over a *high-speed* modem connection between an Amiga and a Sun (4.2 BSD UNIX)? The connection is through a pair of US Robotics Courier HST modems connected at 9600 baud (ARQ). I currently use Kermit for file transfers, but because of its constant error checking it clogs the hell out of the high-speed connection protocol employed by the modems. It is actually faster to reconnect at 2400 baud to transfer using kermit than to use 9600 baud. I'd like to be able to transfer binary at as fast a rate as possible, relying on the modems for error checking. Is there anyone out there with a similar situation who *IS ACTUALLY USING* a method for high-speed binary file transfer from Amiga to a UNIX machine? I'd appreciate any advice, direction or war stories. Thanks. -- - Dave Benjamin - - Interleaf - - ...!eddie.mit.EDU!ileaf!dbjag -
dave@csd4.milw.wisc.edu (David A Rasmussen) (01/12/89)
From article <891@io.UUCP>, by dbjag@io.UUCP (David Benjamin x4050): > Does anyone have a method of transferring files over a *high-speed* modem > I currently use Kermit for file transfers, but because of its constant error > Thanks. > I am using azcomm as I type this. I must not have the right emulation, but it does transfer stuff fast with zmodem. I usually use vt1002.8 or whatever. -- Dave Rasmussen c/o Computing Services Division @ U of WI - Milwaukee Internet: dave@csd4.milw.wisc.edu Uucp: uwvax!uwmcsd1!uwmcsd4!dave {o,o} Some girl with psychic powers asked me, "Tbone what's your sign?" \U/ I blinked and answered "Neon!"
papa@pollux.usc.edu (Marco Papa) (01/12/89)
In article <891@io.UUCP> dbjag@io.UUCP (David Benjamin x4050) writes: >Does anyone have a method of transferring files over a *high-speed* modem >connection between an Amiga and a Sun (4.2 BSD UNIX)? The connection is >through a pair of US Robotics Courier HST modems connected at 9600 baud (ARQ). ^^^^^^^^^^^^^^^^^^^^^^^ ^^^^^^^^^ >I currently use Kermit for file transfers, but because of its constant error >checking it clogs the hell out of the high-speed connection protocol employed >by the modems. It is actually faster to reconnect at 2400 baud to transfer >using kermit than to use 9600 baud. This is of course "commercial", BUT I've been using the US Robotics COURIER HST at 9600 baud with A-Talk III for quite a while. The current version of A-Talk III will be able to to send and receive with ZMODEM protocol at well over 1000 char/sec. I've experienced even 1200 char/sec on some long ARC files. >I'd like to be able to transfer binary at as fast a rate as possible, >relying on the modems for error checking. The fastest A-Talk III/Courier HST speed can be obtained when using RTS/CTS flow control, HST flow control on BOTH transmit and receive, no compression (use ARC files), and A-Talk III baud rate at 19,200. The proper settings on the HST are set with: AT&K0&H1&B1&R2 On A-Talk III select RTS/CTS and 16K input buffer from Preferences, and select BAUD 19200. >Is there anyone out there with a similar situation who *IS ACTUALLY USING* ^^^^^^^^^^^^^^^^^^ >a method for high-speed binary file transfer from Amiga to a UNIX machine? >I'd appreciate any advice, direction or war stories. Yes, I am. Note that the above speeds have been obtained with Channel 1, a Mass BBS that supports the Courier HST. Some UNIXes have problems receiving data as fast as A-Talk III can send, especially when not using RTS/CTS [i which case X-o/X-off should be used for ZMODEM]. There is no problem when the UNIX box is sending files, though. The next release of A-Talk III will contain a specific modem entry for the Courier HST, that will select the above settings automatically (poof!). -- Marco Papa 'Doc' -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= uucp:...!pollux!papa BIX:papa ARPAnet:pollux!papa@oberon.usc.edu "There's Alpha, Beta, Gamma and Diga!" -- Leo Schwab [quoting Rick Unland] -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
hummel@m.cs.uiuc.edu (01/12/89)
Written 12:11 pm Jan 11, 1989 by dbjag@io.UUCP in comp.sys.amiga: /* ---------- "high-speed file transfer" ---------- */ > Is there anyone out there with a similar situation who *IS ACTUALLY USING* > a method for high-speed binary file transfer from Amiga to a UNIX machine? > I'd appreciate any advice, direction or war stories. /* End of text from m.cs.uiuc.edu:comp.sys.amiga */ Yes. As administrator of the Amiga archive at uxe, I have used various methods to upload approximately 100MB of material. Until last week, this has all been at standard serial line speeds (1200-19.2K baud). Far and away the best solution has been Matt Dillon's DNET software. I have used it to upload to VAXen running BSD 4.3 and Multimaxen under a 4.2 derivative. Transfers with the VAXen have been flawless, while intermittently hung sessions with the Multimaxen have actually uncovered UMAX kernel bugs! Some particularly strong features of DNET are its efficient sliding-window protocol and low-overhead packet structure, multiple concurrent sessions, and preservation of directory heirarchies. I have also tried Kermit under vt100. As you pointed out, it is slow and has other problems as well. The most roundabout approach tried was to compress the disks and copy them to 720K MS-DOS format using DOS2DOS, then tftp'ing the compressed files from a PC on Ethernet to the archive's staging area on an IBM 3081. It worked, but was a major hassle. < Lionel ---------- Lionel Hummel 404 W. High St., #6 hummel@cs.uiuc.edu Urbana, IL 61801 {pur-ee,uunet}!uiucdcs!hummel (H) (217)344-5303 Dept. of Computer Science (W) (217)333-7408 University of Illinois, Urbana-Champaign
jwright@atanasoff.cs.iastate.edu (Jim Wright) (01/12/89)
In article <891@io.UUCP> dbjag@io.UUCP (David Benjamin x4050) writes: >Does anyone have a method of transferring files over a *high-speed* modem >connection between an Amiga and a Sun (4.2 BSD UNIX)? The connection is >through a pair of US Robotics Courier HST modems connected at 9600 baud (ARQ). >I'd like to be able to transfer binary at as fast a rate as possible, >relying on the modems for error checking. >Is there anyone out there with a similar situation who *IS ACTUALLY USING* >a method for high-speed binary file transfer from Amiga to a UNIX machine? I have my 2000 hardwired to my Apollo 4000 at 19200 bps. My startup sequence always starts up Handshake, a shareware communications package. I have found it to be the only program reliable at these high speeds. (I can't claim to have tried them all.) It is a very good emulation of VT100, which is a plus on a Unix system. It supports YMODEM-Batch. It does not currently support ZMODEM. For your situation, YMODEM-G would be ideal. You could let those sexy modems do all your error detection/correction for you, while the protocol simply validates the files and informs you of problems. I have not seen any implementations of YMODEM-G for either the Amiga or Unix. Next best would be ZMODEM. It's capable of much greater throughput than any of the other error detecting protocols. While ZMODEM is finally starting to appear on the Amiga, I have found none to work well at high speeds. They either crash the Amiga or hang the serial line on the Apollo. I can verify the ZMODEM on the Unix, since it works flawlessly when connected to PCs using Procomm+. Third in line would be YMODEM-Batch. Multiple file transfers are indispensible. This is a relatively efficient protocol, so you should get good throughput. Kermit is good if you go through a protocol converter or a communications switch. The best that can be said about it is that it works on just about everything (which really is a pretty good thing to say!). It's not fast, but it's better than nothing. Hope this helps. Let us (Net) know if you find something better.
louie@trantor.umd.edu (Louis A. Mamakos) (01/12/89)
You might make sure that you are using larger block sizes when transfering files using KERMIT. The default size is very small; I usually use 1K block sizes which improves the throughput vastly. Louis A. Mamakos WA3YMH Internet: louie@TRANTOR.UMD.EDU University of Maryland, Computer Science Center - Systems Programming
papa@pollux.usc.edu (Marco Papa) (01/13/89)
In article <646@atanasoff.cs.iastate.edu> jwright@atanasoff.cs.iastate.edu (Jim Wright) writes: |In article <891@io.UUCP| dbjag@io.UUCP (David Benjamin x4050) writes: ||I'd like to be able to transfer binary at as fast a rate as possible, ||relying on the modems for error checking. | ||Is there anyone out there with a similar situation who *IS ACTUALLY USING* ||a method for high-speed binary file transfer from Amiga to a UNIX machine? | |For your situation, YMODEM-G would be ideal. You could let those sexy |modems do all your error detection/correction for you, while the protocol |simply validates the files and informs you of problems. I have not seen |any implementations of YMODEM-G for either the Amiga or Unix. Well, you didn't look around enough. ZMODEM and YMODEM-g for UNIX have been available "with sources" for over a year from Chuck Forsberg's TeleGodzilla UNIX-based BBS (503)-621-3746. Chuck is the author of ZMODEM. As far as the Amiga is concerned, A-Talk III Release 1.0e includes Ymodem-G. Note that Ymodem-G does require and error-correcting modem to work properly. |Next best would be ZMODEM. It's capable of much greater throughput |than any of the other error detecting protocols. While ZMODEM is finally |starting to appear on the Amiga, I have found none to work well at high |speeds. They either crash the Amiga or hang the serial line on the |Apollo. I can verify the ZMODEM on the Unix, since it works flawlessly |when connected to PCs using Procomm+. I know of 3 implementations of ZMODEM for the Amiga that work just fine at high speeds: Online! has it (I don't know which version), ZCOMM [a modified version of DJ James Comm 134 program] has it, and A-Talk III has it. There is also a version of Chuck's sz/rz programs that can be used as external protocols with other pd/shareware programs such as Access!. |Third in line would be YMODEM-Batch. Multiple file transfers are |indispensible. This is a relatively efficient protocol, so you should |get good throughput. A-Talk III has that one, too. |Kermit is good if you go through a protocol converter or a communications |switch. The best that can be said about it is that it works on just |about everything (which really is a pretty good thing to say!). It's |not fast, but it's better than nothing. Kermit has been around on the Amiga since the benning. Practically any commercial comm program supports it. And don't forget WXMODEM (windowed-Xmodem). With People Link you can get throughput that approaches ZMODEM. As far as supporting protocols, right now the Amiga is definitely not second to any of the other machines, being IBM PCs, Macs or any other. -- Marco Papa 'Doc' -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= uucp:...!pollux!papa BIX:papa ARPAnet:pollux!papa@oberon.usc.edu "There's Alpha, Beta, Gamma and Diga!" -- Leo Schwab [quoting Rick Unland] -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
daveh@cbmvax.UUCP (Dave Haynie) (01/13/89)
in article <891@io.UUCP>, dbjag@io.UUCP (David Benjamin x4050) says: > Keywords: Amiga sun modem > Does anyone have a method of transferring files over a *high-speed* modem > connection between an Amiga and a Sun (4.2 BSD UNIX)? Standard Kermit is a pig. Don't use it over a high speed modem, or really, any other time you can avoid it. > The connection is through a pair of US Robotics Courier HST modems > connected at 9600 baud (ARQ). My usual connection is a direct line between a UNIX VAX and my Amiga at 19,200 baud. > I currently use Kermit for file transfers, but because of its constant error > checking it clogs the hell out of the high-speed connection protocol employed > by the modems. It is actually faster to reconnect at 2400 baud to transfer > using kermit than to use 9600 baud. > I'd like to be able to transfer binary at as fast a rate as possible, > relying on the modems for error checking. For fast transfers, you'd do much better going to ZModem. I get extremely fast transfers using ZMODEM, probably better than 4x that of kermit, at least (never actually measured it). If you can count on the error check/correct of your modem link, supposedly YMODEM-G will go even faster. The ZMODEM sources for both Amiga and UNIX are pretty readily available. The commercially available ATalk3 program (distributed by OXXI) supports ZMODEM and YMODEM-G, though I haven't seen any publically distributed YMODEM-G sources, I suspect it's pubically documented somewhere. > -- > - Dave Benjamin - > - Interleaf - > - ...!eddie.mit.EDU!ileaf!dbjag - -- Dave Haynie "The 32 Bit Guy" Commodore-Amiga "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: D-DAVE H BIX: hazy Amiga -- It's not just a job, it's an obsession
mp1u+@andrew.cmu.edu (Michael Portuesi) (01/13/89)
dbjag@io.UUCP (David Benjamin x4050) writes: > Is there anyone out there with a similar situation who *IS ACTUALLY USING* > a method for high-speed binary file transfer from Amiga to a UNIX machine? > I'd appreciate any advice, direction or war stories. By far the best way to transfer files between an Amiga and an Unix box is Matt Dillon's DNET. The "putfiles" command offers exceptional throughput (far, far better than Kermit, even with long packets), can handle transferring entire directory trees, and the DNet protocol includes its own flow control so you don't have to worry about choking your communications line. Plus you don't have to worry about binaries getting munged during the transfer...it works perfectly every time. DNet is available for anonymous FTP from ucbvax in pub/amiga, or you can find it on Fish Disk 145. -- Michael Portuesi / Information Technology Center / Carnegie Mellon University INET: mp1u+@andrew.cmu.edu / BITNET: mp1u+@andrew UUCP: ...harvard!andrew.cmu.edu!mp1u+ "I'm very sorry, Master, but that WAS the backup system" -- Slave
scotth@harlie.SGI.COM (Scott Henry) (01/13/89)
From article <IXnFBoy00VsfQBwYZP@andrew.cmu.edu>, by mp1u+@andrew.cmu.edu (Michael Portuesi): > dbjag@io.UUCP (David Benjamin x4050) writes: >> Is there anyone out there with a similar situation who *IS ACTUALLY USING* >> a method for high-speed binary file transfer from Amiga to a UNIX machine? >> I'd appreciate any advice, direction or war stories. > > By far the best way to transfer files between an Amiga and an Unix box > is Matt Dillon's DNET. The "putfiles" command offers exceptional > throughput (far, far better than Kermit, even with long packets), can > handle transferring entire directory trees, and the DNet protocol > includes its own flow control so you don't have to worry about choking > your communications line. Plus you don't have to worry about binaries > getting munged during the transfer...it works perfectly every time. > > DNet is available for anonymous FTP from ucbvax in pub/amiga, or you > can find it on Fish Disk 145. > > -- > Michael Portuesi / Information Technology Center / Carnegie Mellon University > INET: mp1u+@andrew.cmu.edu / BITNET: mp1u+@andrew > UUCP: ...harvard!andrew.cmu.edu!mp1u+ > > "I'm very sorry, Master, but that WAS the backup system" -- Slave Unfortunately, your statement is not entirely correct. Dnet is the best way to transfer files between an Amiga and a BSD Unix box. The unix end DOES NOT compile on a SysV box, and probably not on a Xenix box, either. I know, I spent some time trying do move it to SysV, but gave up (I didn't have the time for a major re-write of the signalling stuff). Scott Henry <scotth@sgi.com> -- Scott Henry
jwright@atanasoff.cs.iastate.edu (Jim Wright) (01/13/89)
In article <14625@oberon.USC.EDU> papa@pollux.usc.edu (Marco Papa) writes: >Well, you didn't look around enough. I said this was based on my personal experience. And I know its limited. > ZMODEM and YMODEM-g for UNIX have been >available "with sources" for over a year from Chuck Forsberg's TeleGodzilla >UNIX-based BBS (503)-621-3746. Chuck is the author of ZMODEM. Yup, that's where I got my Unix version of sz/rz. Thanks Chuck. I didn't know about YMODEM-G, since I haven't had need for it. Thanks Marco for the info. >I know of 3 implementations of ZMODEM for the Amiga that work just fine at >high speeds: [...] That hasn't been my experience. One thing that bothers me a bit is that the versions *I* tried don't like to live with RAM:. They always see it as being full, and either abort the download or hang the system. I know that this can be overcome by some esoteric combination of switches, but the last thing I want to do is memorize umpteen parameters for ZMODEM. >As far as supporting protocols, right now the Amiga is definitely not >second to any of the other machines, being IBM PCs, Macs or any other. I'd say the IBM-PC communications area is definitely much more mature than the Amiga, but things are picking up quickly. Dunno 'bout Macs. I'd give your comm program a look (*YES* I caught the hint :-) but I can't see hooking up a modem to the *one*and*only* serial port on my Amiga. (Hint, hint, cash waiting for good serial card. :-)
kent@swrinde.swri.edu (Kent D. Polk) (01/13/89)
In article <IXnFBoy00VsfQBwYZP@andrew.cmu.edu> mp1u+@andrew.cmu.edu (Michael Portuesi) writes: >dbjag@io.UUCP (David Benjamin x4050) writes: >> Is there anyone out there with a similar situation who *IS ACTUALLY USING* >> a method for high-speed binary file transfer from Amiga to a UNIX machine? >> I'd appreciate any advice, direction or war stories. > >By far the best way to transfer files between an Amiga and an Unix box >is Matt Dillon's DNET. The "putfiles" command offers exceptional >throughput (far, far better than Kermit, even with long packets), can >handle transferring entire directory trees, and the DNet protocol >includes its own flow control so you don't have to worry about choking >your communications line. Plus you don't have to worry about binaries >getting munged during the transfer...it works perfectly every time. > >DNet is available for anonymous FTP from ucbvax in pub/amiga, or you >can find it on Fish Disk 145. > >-- >Michael Portuesi / Information Technology Center / Carnegie Mellon University >INET: mp1u+@andrew.cmu.edu / BITNET: mp1u+@andrew >UUCP: ...harvard!andrew.cmu.edu!mp1u+ > >"I'm very sorry, Master, but that WAS the backup system" -- Slave I agree entirely. Matt Dillons Dnet is the greatest thing since OP amps! Also you can have several amiga processes sending files to the Sun simultaneously (only one process at a time from the Sun to the Amiga though). In addition, you get multiple windowed terminal sessions that are resizable (dnet 'fterms'). So, the way it works is you start up dnet locally, call the Sun at 9600 baud, login, type "dnet", you get a windowed 'fterm' on the amiga, & a csh on the Sun. You start up more windows by typing "fterm" on the Amiga. Fabulous. Be sure to use Interlaced hi-res & overscan so you can get 50 row terminal sessions. Ah yes, no transmission noise while in a terminal session also. Disadvantages? : Not a remotely mounted filesystem. :^) Does not emulate a terminal - uses Amiga ansi (I have a termcap I wrote if you need one). - no Tektronics graphics (you might write a Tek graphics fterm :^) ). I use it all day long here at work - Amiga 500 hooked to Sun fileserver. =============================================================================== Kent Polk - Southwest Research Institute |> Frogsoundz: Ultrasonic {cs.utexas.edu, gatech!petro, sun!texsun}!swrinde!kent|> waveforms time-dilated ___/\___/\___/\___/\___/\___/\___/\___/\___/\___/\____|> & played on my Amiga. ===============================================================================
rick@sbcs.sunysb.edu (Rick Spanbauer) (01/13/89)
In article <7200073@m.cs.uiuc.edu>, hummel@m.cs.uiuc.edu writes: > > Written 12:11 pm Jan 11, 1989 by dbjag@io.UUCP in comp.sys.amiga: > /* ---------- "high-speed file transfer" ---------- */ > > Is there anyone out there with a similar situation who *IS ACTUALLY USING* > > a method for high-speed binary file transfer from Amiga to a UNIX machine? One method of high speed transfer is to use an Ameristar Ethernet board and NFS to provide transparent file services between a host with NFS server and the Amiga. Of course this isn't as cheap as grabbing some PD software to do kermit things around. > Far and away the best solution has been Matt Dillon's DNET software. I have Best is subjective :-) People who have our stuff might be inclined to differ with your opinion! > Lionel Hummel 404 W. High St., #6 Rick Spanbauer Ameristar
atheybey@PTT.LCS.MIT.EDU (01/13/89)
If you are connecting to some flavor of BSD Unix, DNet is wonderful. The ability to have multiple open connections is a *big* win. If I am in emacs and want to do something in the shell, I just pop to my CLI, type "run fterm", and get another window with a Unix shell. Of course, I could suspend the emacs, but then I would have to redraw its screen when I resumed it (at 1200 baud, this is a big deal). You can even continue to work on the remote computer (for example, reading USENET :-) while downloading files. Of course, things are then kinda slow, but it still works and is useful. That plug for DNet aside... I have tried DNet with the USRobitics HST modems, and have a few reservations. (I ended up deciding that I couldn't afford an HST modem, and so did not look for a solution too vigorously.) The HST modems have a *unidirectional* 9600 baud channel. The two modems negotiate which direction gets the high speed channel, based on the number of characters in the modems' buffers. The other direction is 300 baud. The echoing of characters (when typing to a Unix shell) was actually *slower* than at 1200 baud. I think that this is because DNet sends a packet of several characters when you type one character. The sending modem then thinks that it has large amount to send, and asks for the high speed channel. When the echo of the typed character comes back, it is also in a packet, and so the modem on the Unix end asks for the high speed channel. I think that the two modems spend a lot of time swapping the high speed channel when you type. For sending large amounts of data (I cat'ed and putfile'd the same file) the modems achieved about 3800 bps. I can't say whether this number is limited by the interaction of DNet with the modems, or by the sending ability of the remote computer. I (stupidly) did not time sending a file without using DNet. The setup was Amiga<-->some form of terminal concentrator<-->Ethernet<-->uVax II. I do know that other people dialing up to the terminal concentrator with the HST modems and terminals do achieve rates much closer to 9600 bps (at least fast enough that a VT100 can't handle it without flow control :-). If one could permamently assign the high speed channel to the Unix-->Amiga direction, I think that DNet would have performed better (except for uploads, of course). I don't know if it is possible with this modem. --------------------------------------------------------------------------- Andrew Heybey, atheybey@ptt.lcs.mit.edu Room 509, 545 Technology Square Advanced Network Architecture Group Cambridge, MA 02139 MIT Laboratory for Computer Science (617) 253-6011 ---------------------------------------------------------------------------
papa@pollux.usc.edu (Marco Papa) (01/14/89)
In article <653@atanasoff.cs.iastate.edu> jwright@atanasoff.cs.iastate.edu (Jim Wright) writes: |Yup, that's where I got my Unix version of sz/rz. Thanks Chuck. |I didn't know about YMODEM-G, since I haven't had need for it. Thanks |Marco for the info. The Chuck Forsberg's sources include implementation of 1/2 of Ymodem-g (I forgot, I think it was receive). It took me 30 min to modify the other one. |That hasn't been my experience. One thing that bothers me a bit is that |the versions *I* tried don't like to live with RAM:. They always see it |as being full, and either abort the download or hang the system. I know |that this can be overcome by some esoteric combination of switches, but |the last thing I want to do is memorize umpteen parameters for ZMODEM. I really don't know what is the problem with RAM: you are mentioning, BUT I "exclusively" use RAM: with upload/downloads and A-Talk III because it gives me the maximum speed. And I use the "standard" 1.3 RAM: device. I don't use ANY esoteric combination of switches: just RAM 1.3 straight off the WB disk. Of course, I cannot comment for "other" programs, just for my own. |I'd give your comm program a look (*YES* I caught the hint :-) but I |can't see hooking up a modem to the *one*and*only* serial port on |my Amiga. (Hint, hint, cash waiting for good serial card. :-) Just wait for an announcement SOON (hint!, hint!). -- Marco Papa 'Doc' -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= uucp:...!pollux!papa BIX:papa ARPAnet:pollux!papa@oberon.usc.edu "There's Alpha, Beta, Gamma and Diga!" -- Leo Schwab [quoting Rick Unland] -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
papa@pollux.usc.edu (Marco Papa) (01/14/89)
In article <8901131551.AA12559@PTT.LCS.MIT.EDU> atheybey@PTT.LCS.MIT.EDU writes: |I have tried DNet with the USRobitics HST modems, and have a few |reservations. (I ended up deciding that I couldn't afford an HST |modem, and so did not look for a solution too vigorously.) The HST |modems have a *unidirectional* 9600 baud channel. The two modems |negotiate which direction gets the high speed channel, based on the |number of characters in the modems' buffers. The other direction is |300 baud. Yes, that's a problem with the HST and simlilar "uniditectionsl" high-speed modems. It is not a problem at all with a "normal" terminal program. |I do know that other |people dialing up to the terminal concentrator with the HST modems and |terminals do achieve rates much closer to 9600 bps (at least fast |enough that a VT100 can't handle it without flow control :-). Way, way too fast for a "real" vt100. The HST will allow transfers of compressed files with speeds of 1500 char/sec, and transfers of text files of up to 1800 char/sec or more depending on the file. The "unidirectionality" of the HST modem works in its favour with a "normal" communication program. Characters typed are not bufferred. Protocols like Zmodem and Ymodem-g use NO acknowlegment, so one channel is not used unless there is an error. |If one could permamently assign the high speed channel to the |Unix--|Amiga direction, I think that DNet would have performed better |(except for uploads, of course). I don't know if it is possible with |this modem. I'm not sure it can be done, since it is the modem that "dinamycally" reassigns the high speed channel to either send or receive. Surely I could not find an AT-command that allows me to set that. -- Marco Papa 'Doc' -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= uucp:...!pollux!papa BIX:papa ARPAnet:pollux!papa@oberon.usc.edu "There's Alpha, Beta, Gamma and Diga!" -- Leo Schwab [quoting Rick Unland] -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
paolucci@snll-arpagw.UUCP (Sam Paolucci) (01/14/89)
In article <24863@sgi.SGI.COM> scotth@harlie.SGI.COM (Scott Henry) writes: ->From article <IXnFBoy00VsfQBwYZP@andrew.cmu.edu>, by mp1u+@andrew.cmu.edu (Michael Portuesi): ->> dbjag@io.UUCP (David Benjamin x4050) writes: ->>> Is there anyone out there with a similar situation who *IS ACTUALLY USING* ->>> a method for high-speed binary file transfer from Amiga to a UNIX machine? ->>> I'd appreciate any advice, direction or war stories. ->> ->> By far the best way to transfer files between an Amiga and an Unix box ->> is Matt Dillon's DNET. The "putfiles" command offers exceptional ->> throughput (far, far better than Kermit, even with long packets), can ->> handle transferring entire directory trees, and the DNet protocol ->> includes its own flow control so you don't have to worry about choking ->> your communications line. Plus you don't have to worry about binaries ->> getting munged during the transfer...it works perfectly every time. ->> ->> DNet is available for anonymous FTP from ucbvax in pub/amiga, or you ->> can find it on Fish Disk 145. ->> ->> -- ->> Michael Portuesi / Information Technology Center / Carnegie Mellon University ->> INET: mp1u+@andrew.cmu.edu / BITNET: mp1u+@andrew ->> UUCP: ...harvard!andrew.cmu.edu!mp1u+ ->> ->> "I'm very sorry, Master, but that WAS the backup system" -- Slave -> ->Unfortunately, your statement is not entirely correct. Dnet is the best way ->to transfer files between an Amiga and a BSD Unix box. The unix end DOES NOT ->compile on a SysV box, and probably not on a Xenix box, either. I know, I ->spent some time trying do move it to SysV, but gave up (I didn't have the ->time for a major re-write of the signalling stuff). -> -> Scott Henry <scotth@sgi.com> ->-- -> Scott Henry Also, I have not been able to get it working correctly in an ULTRIX machine which is BSD 4.2 compatible. -- -+= SAM =+- "the best things in life are free" ARPA: paolucci@snll-arpagw.llnl.gov
aaron@madnix.UUCP (Aaron Avery) (01/14/89)
In article <653@atanasoff.cs.iastate.edu> jwright@atanasoff.cs.iastate.edu (Jim Wright) writes: >I'd give your comm program a look (*YES* I caught the hint :-) but I >can't see hooking up a modem to the *one*and*only* serial port on >my Amiga. (Hint, hint, cash waiting for good serial card. :-) Marco's program seems to do absolutely _everything_ you might want. I have only a modem to hook up to my serial port at home, but I do sympathize with you. I wouldn't, however, like to have too many copies of A-Talk III in memory at once (on multiple ports) as that thing is really *big*. Now that Marco has hinted at you about A-Talk III, I can tell you that ASDG is now shipping our two port serial module for the Twin-X multi-purpose I/O board. It probably won't be the cheapest way to add just one port to your amiga, but it's the only one you can get now! Twin-X is only available for the A2000, and if you're interested, you can ask Marco how well it works, as he's got a few of our ports hooked up right now. -- Aaron Avery, ASDG Inc. "A mime is a terrible thing to waste." -- Robin Williams UUCP: {harvard|rutgers|ucbvax}!uwvax!nicmad!madnix!aaron ARPA: madnix!aaron@cs.wisc.edu
karl@sugar.uu.net (Karl Lehenbauer) (01/14/89)
In article <891@io.UUCP| dbjag@io.UUCP (David Benjamin x4050) writes: > I'd like to be able to transfer binary at as fast a rate as possible, > relying on the modems for error checking. > > Is there anyone out there with a similar situation who *IS ACTUALLY USING* > a method for high-speed binary file transfer from Amiga to a UNIX machine? I use uupc on my Amiga to talk to uucp on my Unix system at 9600 baud. The nice thing about that is you don't have to attend the file transfers and if a transfer screws up or dies, the originator (Unix or the Amiga) still knows what files haven't been sent and will resume on the next connection. Being a windowed protocol, uucp will achieve higher speeds than XMODEM, YMODEM and non-sliding-windows kermit (lots of kermits seem to have problems with binary data or have to quote 8-bit chars in a two-for-one arrangement -ick) Also, I have successfully done XMODEM and Kermit transfers between the Unix system (386) and the Amiga at 19200 baud. Finally, even with error-correcting modems, I'm not sure that it's a good idea to not do end-to-end error checking in software because data can still glitch along the path that's not between the modems. -- -- uunet!sugar!karl | "We've been following your progress with considerable -- karl@sugar.uu.net | interest, not to say contempt." -- Zaphod Beeblebrox IV -- Usenet BBS (713) 438-5018
wen@husc4.HARVARD.EDU (A. Wen) (01/15/89)
In article <32@snll-arpagw.UUCP> paolucci@snll-arpagw.UUCP (Sam Paolucci) writes: >Also, I have not been able to get [DNET] working correctly in an ULTRIX machine >which is BSD 4.2 compatible. >-- > -+= SAM =+- I've compiled and run DNET on an Ultrix 2.3 machine with no problems. -- A. Wen wen@husc4.HARVARD.EDU wen@husc4.BITNET {seismo!harvard!husc4}
peter@sugar.uu.net (Peter da Silva) (01/15/89)
In article <24863@sgi.SGI.COM>, scotth@harlie.SGI.COM (Scott Henry) writes: > From article <IXnFBoy00VsfQBwYZP@andrew.cmu.edu>, by mp1u+@andrew.cmu.edu (Michael Portuesi): > > dbjag@io.UUCP (David Benjamin x4050) writes: > >> Is there anyone out there with a similar situation who *IS ACTUALLY USING* > >> a method for high-speed binary file transfer from Amiga to a UNIX machine? > >> I'd appreciate any advice, direction or war stories. UUPC has got to be the most portable, fastest, and for multiple files most convenient way. I wish I could say the same of GNUUUCP, but it doesn't do windows. -- Peter "Have you hugged your wolf today" da Silva `-_-' Hackercorp. ...texbell!sugar!peter, or peter@sugar.uu.net 'U`
hummel@s.cs.uiuc.edu (01/16/89)
Written 9:32 am Jan 13, 1989 by rick@sbcs.sunysb.edu in comp.sys.amiga: In article <7200073@m.cs.uiuc.edu>, hummel@m.cs.uiuc.edu writes: >> Far and away the best solution has been Matt Dillon's DNET software. I have > Best is subjective :-) People who have our stuff might be inclined to > differ with your opinion! /* End of text from comp.sys.amiga */ Best isn't subjective; it's context sensitive :-) In the context that I was writing, I'll stand by my statement that >> at standard serial line speeds (1200-19.2K baud). DNET is the best there is. Taking my statement (and, I believe, the original question) out of context (which you did), and disregarding cost, I'll agree that Ameristar produces a better solution. As a matter of fact, the "up until last week" in my response was a qualifier intended because I had just gotten the opportunity to use Northwestern University's Ethernet'd Amiga lab to post another update to the archive. I had mentioned so when I posted an announcement of the update, but I didn't dwell on it this time. Since you took my words out of context to plug your product, I really have few qualms about airing a complaint about your NFS software: It fails to preserve the case of files. Now, I know that there is some question here because UNIX is case sensitive and AmigaDOS isn't. But if the local filesystem PRESERVES case then a remote one should too. In any case, you did a fine job on your software, but NFS is not an answer to the original poster's question. My only regret: Mentioning the use of an MS-DOS machine. It was irrelevant. You would have had nothing to apologize about if you had selected that as your straw man rather than DNET. (Too bad you missed the RIGHT opening!-) < Lionel ---------- Lionel Hummel 404 W. High St., #6 hummel@cs.uiuc.edu Urbana, IL 61801 {seismo,pur-ee,convex}!uiucdcs!hummel (H) (217)344-5303 Dept. of Computer Science (W) (217)333-7408 University of Illinois, Urbana-Champaign
hummel@s.cs.uiuc.edu (01/16/89)
Eep! I haven't heard from anyone yet, but before someone notices and gets the wrong idea, let me correct myself... Written 3:58 pm Jan 15, 1989 by hummel@s.cs.uiuc.edu in comp.sys.amiga: > Since you took my words out of context to plug your product, I really have few > qualms about airing a complaint about your NFS software: It fails to preserve > the case of files. Now, I know that there is some question here because UNIX ^^^^^ I meant _file names_, not files! It's just that I times some type than form faster sentences I coherent. :-) > is case sensitive and AmigaDOS isn't. But if the local filesystem PRESERVES > case then a remote one should too. In any case, you did a fine job on your > software, but NFS is not an answer to the original poster's question. /* End of text from s.cs.uiuc.edu:comp.sys.amiga */ < Lionel Inverse Mode ---------- Lionel Hummel 404 W. High St., #6 hummel@cs.uiuc.edu Urbana, IL 61801 {seismo,pur-ee,convex}!uiucdcs!hummel (H) (217)344-5303 Dept. of Computer Science (W) (217)333-7408 University of Illinois, Urbana-Champaign
rminnich@super.ORG (Ronald G Minnich) (01/17/89)
In article <212200006@s.cs.uiuc.edu> hummel@s.cs.uiuc.edu writes: > >Written 9:32 am Jan 13, 1989 by rick@sbcs.sunysb.edu in comp.sys.amiga: >In article <7200073@m.cs.uiuc.edu>, hummel@m.cs.uiuc.edu writes: >Best isn't subjective; it's context sensitive :-) In the context that I was >writing, I'll stand by my statement that >>> at standard serial line speeds (1200-19.2K baud). >DNET is the best there is. well, then there are two best's, i guess. The Karn tcp/ip on the amiga was kinda nice. I could ftp to swan.ulowell.edu directly from ronsamiga.udel.edu; DNET is point-point only. If you have a Cisco box you can plug your amiga right in and be on the network, all for free. At 19.2 Kb i saw about the same throughput and utilization (99%, more or less) for DNET and amigatcp. A 19.2 kb link is nice, considering that the original arpanet links were 56 kb! The single biggest problem was that i ran out of time to get the last bug out of internet.device, so you did not have the 'run a program for another window' that dnet has. But you could still have multiple sessions right from the original Karn code. I am going to try to submit this to comp.sources.amiga again (it got lost in the Black Hole at purdue last march) and see if anyone fixes the bug. ron P.S. Yes, i use DNET now and did then; it's great. Ran fine under Ultrix {1,2}.x.
paolucci@snll-arpagw.UUCP (Sam Paolucci) (01/18/89)
In article <212200006@s.cs.uiuc.edu> hummel@s.cs.uiuc.edu writes:
->
->Since you took my words out of context to plug your product, I really have few
->qualms about airing a complaint about your NFS software: It fails to preserve
->the case of files. Now, I know that there is some question here because UNIX
->is case sensitive and AmigaDOS isn't. But if the local filesystem PRESERVES
->case then a remote one should too. In any case, you did a fine job on your
->software, but NFS is not an answer to the original poster's question.
If they retained case sensitivity of file names then you would be forced to
type "NewCLI", "EndCLI", etc. How many Amiga users invoke commands using
the exact case of the file names which are saved in? I rest my case.
I think that Ameristar took the only sensible choice considering that the
Amiga ignores case sensitivity presently.
->Lionel Hummel 404 W. High St., #6
->hummel@cs.uiuc.edu Urbana, IL 61801
--
-+= SAM =+-
"the best things in life are free"
ARPA: paolucci@snll-arpagw.llnl.gov
cmcmanis%pepper@Sun.COM (Chuck McManis) (01/19/89)
In article <36@snll-arpagw.UUCP> (Sam Paolucci) writes: >I think that Ameristar took the only sensible choice considering that the >Amiga ignores case sensitivity presently. You are correct and it is a wart on the NFS protocol rather than the Ameristar implementation of it. One of the main uses the Amiga 2000 here at Sun gets is to look at Unix <--> NonUnix interactions on the NFS and other protocols. That's why there is a Version 3 of NFS in the works to handle such things as different usages of case. Consider the problem our poor cousin has with filenames that are restricted to 11 characters (one of them being a '.'). This is the most challenging area of the new protocol work. But enough of all that. Yes, Ameristart makes the best NFS product for the Amiga (and the worst too I suppose since it seems to be the only one) and if you want to share files between an Amiga and a larger machine it is the only way to fly. --Chuck McManis uucp: {anywhere}!sun!cmcmanis BIX: cmcmanis ARPAnet: cmcmanis@sun.com These opinions are my own and no one elses, but you knew that didn't you.
acs@pccuts.pcc.amdahl.com (Tony Sumrall) (01/19/89)
In article <36@snll-arpagw.UUCP> paolucci@snll-arpagw.UUCP (Sam Paolucci) writes: >In article <212200006@s.cs.uiuc.edu> hummel@s.cs.uiuc.edu writes: >-> >->Since you took my words out of context to plug your product, I really have few >->qualms about airing a complaint about your NFS software: It fails to preserve >->the case of files. Now, I know that there is some question here because UNIX >->is case sensitive and AmigaDOS isn't. But if the local filesystem PRESERVES >->case then a remote one should too. In any case, you did a fine job on your >->software, but NFS is not an answer to the original poster's question. > >If they retained case sensitivity of file names then you would be forced to >type "NewCLI", "EndCLI", etc. How many Amiga users invoke commands using >the exact case of the file names which are saved in? I rest my case. Whoa up there! The Amiga doesn't care about case when it's *processing* files but: * it allows the user to create filenames containing both upper and lower case * it preserves the filename during copy and rename operations. I create filenames so that I can have some idea of their contents. SomethingAboutMyLatestProject is easier to read than somethingaboutmylatestproject (I shy away from spaces so I don't have to quote the filename). Do you still think your case is rested :-) > -+= SAM =+- > ARPA: paolucci@snll-arpagw.llnl.gov -- Tony Sumrall acs@pccuts.pcc.amdahl.com <=> amdahl!pccuts!acs [ Opinions expressed herein are the author's and should not be construed to reflect the views of Amdahl Corp. ]
paolucci@snll-arpagw.UUCP (Sam Paolucci) (01/20/89)
In article <743@pccuts.pcc.amdahl.com> acs@pccuts.pcc.amdahl.com (Tony Sumrall) writes: ->In article <36@snll-arpagw.UUCP> paolucci@snll-arpagw.UUCP (Sam Paolucci) writes: ->>If they retained case sensitivity of file names then you would be forced to ->>type "NewCLI", "EndCLI", etc. How many Amiga users invoke commands using ->>the exact case of the file names which are saved in? I rest my case. -> ->Whoa up there! The Amiga doesn't care about case when it's *processing* ->files but: -> -> * it allows the user to create filenames containing both upper and -> lower case -> * it preserves the filename during copy and rename operations. -> ->I create filenames so that I can have some idea of their contents. ->SomethingAboutMyLatestProject is easier to read than ->somethingaboutmylatestproject (I shy away from spaces so I don't have to ->quote the filename). Do you still think your case is rested :-) I guess you missed the point entirely. I know that the Amiga does not care about case sensitivity, but a Sun does. So if a Sun is your NFS file server and you ask for "newcli" while the file on the Sun is called NewCLI, the Sun will tell the Amiga "don't got it". That is the reason why Ameristar downcases all the file names when you copy them to your server. Note that the case is not modified if you copy the files to your server using FTP. However, then to invoke a new CLI you would be forced to type "NewCLI". ->> -+= SAM =+- ->> ARPA: paolucci@snll-arpagw.llnl.gov ->-- ->Tony Sumrall acs@pccuts.pcc.amdahl.com <=> amdahl!pccuts!acs -- -+= SAM =+- "the best things in life are free" ARPA: paolucci@snll-arpagw.llnl.gov
bader+@andrew.cmu.edu (Miles Bader) (01/21/89)
paolucci@snll-arpagw.UUCP (Sam Paolucci) writes: > ... > I guess you missed the point entirely. I know that the Amiga does not > care about case sensitivity, but a Sun does. So if a Sun is your > NFS file server and you ask for "newcli" while the file on the Sun > is called NewCLI, the Sun will tell the Amiga "don't got it". That > is the reason why Ameristar downcases all the file names when you copy > them to your server. Note that the case is not modified if you copy > the files to your server using FTP. However, then to invoke a new CLI > you would be forced to type "NewCLI". Actions like this (trashing case) should be left to the user. There's very little more irritating than a program that purports to know better than the user what he wants. -Miles
mikeeb@sco.COM (Michael E. Bays) (01/21/89)
does anyone know of any PD or non-PD programs that will let me read a MS-DOS 3 1/2 disk on the Amiga drive?? I want to be able to get simple files off the mddos disk and place them in ram: and then put them on the amiga disk HELP!!! michael
acs@pccuts.pcc.amdahl.com (Tony Sumrall) (01/21/89)
In article <38@snll-arpagw.UUCP> paolucci@snll-arpagw.UUCP (Sam Paolucci) replies to my reply: >I guess you missed the point entirely. I know that the Amiga does not >care about case sensitivity, but a Sun does. Gotcha! OK, I'm embarrassed and I hope your case is resting well. Thanks for being kind (I could've been flamed to a cinder :-). > -+= SAM =+- > ARPA: paolucci@snll-arpagw.llnl.gov -- Tony Sumrall acs@pccuts.pcc.amdahl.com <=> amdahl!pccuts!acs [ Opinions expressed herein are the author's and should not be construed to reflect the views of Amdahl Corp. ]
paolucci@snll-arpagw.UUCP (Sam Paolucci) (01/22/89)
In article <kXpwHly00Uka42MyYR@andrew.cmu.edu> bader+@andrew.cmu.edu (Miles Bader) writes: ->paolucci@snll-arpagw.UUCP (Sam Paolucci) writes: ->> ... ->> I guess you missed the point entirely. I know that the Amiga does not ->> care about case sensitivity, but a Sun does. So if a Sun is your ->> NFS file server and you ask for "newcli" while the file on the Sun ->> is called NewCLI, the Sun will tell the Amiga "don't got it". That ->> is the reason why Ameristar downcases all the file names when you copy ->> them to your server. Note that the case is not modified if you copy ->> the files to your server using FTP. However, then to invoke a new CLI ->> you would be forced to type "NewCLI". -> ->Actions like this (trashing case) should be left to the user. There's ->very little more irritating than a program that purports to know ->better than the user what he wants. But if they left the case alone then you probably would be very upset that typing "newcli" or "endcli" would not work when the files where residing on an NFS server. >-Miles -- -+= SAM =+- "the best things in life are free" ARPA: paolucci@snll-arpagw.llnl.gov
jim@b11.INGR.COM (Jim Levie ) (01/23/89)
In article <38@snll-arpagw.UUCP> paolucci@snll-arpagw.UUCP (Sam Paolucci) writes: >->In article <36@snll-arpagw.UUCP> paolucci@snll-arpagw.UUCP (Sam Paolucci) writes: >->>If they retained case sensitivity of file names then you would be forced to >->>type "NewCLI", "EndCLI", etc. How many Amiga users invoke commands using >->>the exact case of the file names which are saved in? I rest my case. >-> >->Whoa up there! The Amiga doesn't care about case when it's *processing* >->files but: >-> >-> * it allows the user to create filenames containing both upper and >-> lower case >-> * it preserves the filename during copy and rename operations. > . > . >I guess you missed the point entirely. I know that the Amiga does not >care about case sensitivity, but a Sun does. So if a Sun is your >NFS file server and you ask for "newcli" while the file on the Sun >is called NewCLI, the Sun will tell the Amiga "don't got it". That >is the reason why Ameristar downcases all the file names when you copy >them to your server. Note that the case is not modified if you copy >the files to your server using FTP. However, then to invoke a new CLI >you would be forced to type "NewCLI". > In my opinion this is a case where two wrongs do not make a right... The Amiga SHOULD be case sensitive at the DOS/EXEC level, if you don't want to type "NewCLI" just rename it to "newcli". Workbench and/or the CLI could be made to ignore or honor case also, or better yet let the user select how they want this handled in "Preferences". Ameristar's downcasing of the file names when accessing an NFS partition is only workable if the directories are only being used by Amigas. Just try using NFS to work on a directory containing stuff that is used both on the Amiga and Unix, and that contains mixed case file names. As long as the Amiga only reads the files everything is fine, but if you edit any of the files... crunch, now all the Makefiles are broken. Or, compress a bunch of files on the Amiga to an NFS partition, then try to uncompress them under Unix. I could go on and on citing problems that I see all the time that are caused by this. -- =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ Jim Levie REMTECH Inc Huntsville, Al The opinions expressed above are just that. Ph. (205) 536-8581 email: uunet!ingr!b11!jim
paolucci@snll-arpagw.UUCP (Sam Paolucci) (01/23/89)
In article <359@b11.INGR.COM> jim@b11.UUCP (Jim Levie ) writes: ->In article <38@snll-arpagw.UUCP> paolucci@snll-arpagw.UUCP (Sam Paolucci) writes: ->>I guess you missed the point entirely. I know that the Amiga does not ->>care about case sensitivity, but a Sun does. So if a Sun is your ->>NFS file server and you ask for "newcli" while the file on the Sun ->>is called NewCLI, the Sun will tell the Amiga "don't got it". That ->>is the reason why Ameristar downcases all the file names when you copy ->>them to your server. Note that the case is not modified if you copy ->>the files to your server using FTP. However, then to invoke a new CLI ->>you would be forced to type "NewCLI". ->> -> ->In my opinion this is a case where two wrongs do not make a right... The ->Amiga SHOULD be case sensitive at the DOS/EXEC level, if you don't want to ->type "NewCLI" just rename it to "newcli". Workbench and/or the CLI could be ->made to ignore or honor case also, or better yet let the user select how they ->want this handled in "Preferences". -> -> Ameristar's downcasing of the file names when accessing an NFS partition ->is only workable if the directories are only being used by Amigas. Just ->try using NFS to work on a directory containing stuff that is used both on ->the Amiga and Unix, and that contains mixed case file names. As long as ->the Amiga only reads the files everything is fine, but if you edit any of ->the files... crunch, now all the Makefiles are broken. Or, compress a ->bunch of files on the Amiga to an NFS partition, then try to uncompress ->them under Unix. I could go on and on citing problems that I see all the ->time that are caused by this. I agree with you entirely. The point is that the Amiga SHOULD be case sensitive at the DOS/EXEC level, but currently is not. Until that happens we are left to make do with the current situation. Many times I've had to log in the Unix machine to rename files so that some of my Makefiles would work. My original point was in reply to a post which criticized Ameristar for making the decision to downcase all file names. I don't think that Ameristar would have done that if at the DOS/EXEC level case sensitivity would have been preserved. Thus I feel that people are barking up the wrong tree in criticizing Ameristar for this. -> Jim Levie REMTECH Inc Huntsville, Al -> The opinions expressed above are just that. -> Ph. (205) 536-8581 email: uunet!ingr!b11!jim -- -+= SAM =+- "the best things in life are free" ARPA: paolucci@snll-arpagw.llnl.gov
cg@myrias.UUCP (Chris Gray) (01/25/89)
In article <44@snll-arpagw.UUCP> paolucci@snll-arpagw.UUCP (Sam Paolucci) writes: >I agree with you entirely. The point is that the Amiga SHOULD be case >sensitive at the DOS/EXEC level, but currently is not. Until that >happens we are left to make do with the current situation. Many times >I've had to log in the Unix machine to rename files so that some of my >Makefiles would work. NO NO NO NO NO!!!! Please do NOT make DOS case sensitive. I consider the way AmigaDOS handles file names to be a great improvement over other systems (e.g. UNIX). It allows me to have names that show up as meaningful on a directory listing, and at the same time allows me to not bother with the shift key when typing in command names. I LIKE this. A lot!! If anything is done in this area, I would suggest a switch or option somewhere (perhaps when the CLI/Shell is started up) that controls the sensitivity. Something I could probably live with is making DOS itself case sensitive (I'm not terribly worried about capitalizing file names inside programs), but have the CLI/Shell accept either case on command lines. -- Chris Gray Myrias Research, Edmonton +1 403 428 1616 {uunet!mnetor,ubc-vision,watmath,vax135}!alberta!myrias!cg
healy@nrl-cmf.UUCP (Liam Healy) (01/26/89)
Sender: Reply-To: healy@cmsun.nrl.navy.mil (Liam Healy) Followup-To: Distribution: Organization: NRL Connection Machine Facility, Washington, DC Keywords: Just a point to clarify terminology: there is a difference between case-sensitive and case-preserving, UNIX is case-sensitive and case-preserving, VMS is neither (everything mapped to uppercase), AmigaDOS is not case-sensitive, but is case- preserving. That is, on the Amiga a program given a name, e.g., FooBar, will keep that combination of upper and lower case, but may be referred to with any case combination (e.g., fOObAR). Liam Healy, Naval Research Lab healy@cmsun.nrl.navy.mil
bader+@andrew.cmu.edu (Miles Bader) (01/26/89)
paolucci@snll-arpagw.UUCP (Sam Paolucci) writes: > I agree with you entirely. The point is that the Amiga SHOULD be case > sensitive at the DOS/EXEC level, but currently is not. Why should it? -Miles
paolucci@snll-arpagw.UUCP (Sam Paolucci) (01/26/89)
In article <gXrVpay00Uka04vOgJ@andrew.cmu.edu> bader+@andrew.cmu.edu (Miles Bader) writes: ->paolucci@snll-arpagw.UUCP (Sam Paolucci) writes: ->> I agree with you entirely. The point is that the Amiga SHOULD be case ->> sensitive at the DOS/EXEC level, but currently is not. -> ->Why should it? When I said that the Amiga should be case sensitive, it was only in the context of eliminating the problem related with NFS. Of course if your Amiga is not networked and not using NFS, there would be no need to have case sensitivity. However, I think that case sensitivity would help in a network environment, particularly when most of the other hosts on the network talk Unix. ->-Miles -- -+= SAM =+- "the best things in life are free" ARPA: paolucci@snll-arpagw.llnl.gov
cjp@antique.UUCP (Charles Poirier) (02/01/89)
Summary: References: <891@io.UUCP> <212200006@s.cs.uiuc.edu> <36@snll-arpagw.UUCP> <743@pccuts.pcc.amdahl.com> <38@snll-arpagw.UUCP> <359@b11.INGR.COM> <44@snll-arpagw.UUCP> <798@myrias.UUCP> Reply-To: vax135!cjp (Charles Poirier) In article <798@myrias.UUCP> cg@myrias.UUCP writes: >Please do NOT make DOS case sensitive. I consider the way AmigaDOS handles >file names to be a great improvement over other systems (e.g. UNIX). It >allows me to have names that show up as meaningful on a directory listing, >and at the same time allows me to not bother with the shift key when typing >in command names. I LIKE this. A lot!! Definitely. Please don't compromise this *feature*. >Something I could probably live with is making DOS itself case >sensitive ... but have the CLI/Shell accept either case on command lines. No, not unless you can convince it to not do a full directory scan on each PATH directory for a wrong-case command. That would be a disaster for the large number of CLI commands that are by default capitalized. Cheers, -- Charles Poirier (decvax,ucbvax,mcnc,attmail)!vax135!cjp "Docking complete... Docking complete... Docking complete..."