mishkin@jrst.apollo.hp.com (Nathaniel Mishkin) (04/04/91)
I hate to be cynical, but I just don't know how to react to the
discussion about the DM and about the lack of DOMAIN on the 9000/700.
It's really nice to hear about how people like the DM and DOMAIN but
I'm afraid it's just very late in the game. The world decided that
standard is better than better (and that even better in the form of
superset functionality is suspect). The world also decided that
cheaper (and better yet, free) is better than better. The rest -- you
must be standard and we're not going to let you make enough money (at
least in software) to do "better" as well as "standard" -- pretty much
follows on.
With the DCE, we're trying to carry forward as much "better" as we
can. As to the DM (with its odd mixture of the incredibly useful and
the incredibly clunky), I can only hope that after a while, given the
fact that X is so widespread and that the usability of what's out
there on top of it is so bad, someone will build something that has
the "incredibly useful" parts of the DM.
Not HP's opinions; just mine after a long haul.
--
-- Nat Mishkin
Cooperative Object Computing Division / East
Hewlett-Packard Company
mishkin@apollo.hp.comnazgul@alphalpha.com (Kee Hinckley) (04/05/91)
In article <MISHKIN.91Apr3231301@jrst.apollo.hp.com> mishkin@jrst.apollo.hp.com (Nathaniel Mishkin) writes: >I hate to be cynical, but I just don't know how to react to the >discussion about the DM and about the lack of DOMAIN on the 9000/700. >It's really nice to hear about how people like the DM and DOMAIN but >I'm afraid it's just very late in the game. The world decided that >standard is better than better (and that even better in the form of >superset functionality is suspect). The world also decided that >cheaper (and better yet, free) is better than better. The rest -- you >must be standard and we're not going to let you make enough money (at >least in software) to do "better" as well as "standard" -- pretty much >follows on. I actually agree with this, but with two caveats. First of all, switching to a standard does not mean that you can't improve on it. Apollo's DM has a good 6 to 10 years more history behind it than an X window manager. Improving Mwm with DM ideas would be well worth it, and HP is in an ideal position to influence that standard. Secondly I have to say I could care less about SR11 right now. My 16 meg DN3500 is dying trying to page sendmail, X and a Motif application against each other. I'd much rather run OSF/1, which I assume is no *worse*, and might even be better (lose a good file system, gain some performance). Unfortunately HP has chosen not to support OSF/1 on the DN[34]500 class machines. Instead I'm supposed to upgrade those machines. And of course my DN3000 is hopeless, but I guess I'll live with that. I don't know what that upgrade costs (can someone say? maybe my argument here is all wet), but if it gets anywhere close to $2000 I start thinking about buying something else. We just got a second hand Sun IPC* for $5K and it's a *lot* faster than my 3500 will ever be. Maybe when HP releases a low-end PA I'll look at that, but right now I'm real worried about bottom line, and the cheapest, easiest, and most useful upgrade I could make to a DN3500 is to run OSF/1. Maybe I'm wrong, and OSF/1 will be a dog and I really do need that CPU upgrade, but I have trouble believing that. >the incredibly clunky), I can only hope that after a while, given the >fact that X is so widespread and that the usability of what's out >there on top of it is so bad, someone will build something that has >the "incredibly useful" parts of the DM. The only "someone" that is qualified (discounting Alfalfa, and we don't have the money to finance the time it would take) is Apollo. I should note that Apollo also has the greatest incentive. You've got an awful lot of people in R&D who are refusing to switch to X right now - and that means you're creating yet another situation (Unix being the first) where Apollo isn't using the stuff it ships. -kee Yes I'm ashamed to admit it but it's true - I'm typing this on a Sun. NFS is a crock and slower than mollasses, all the terminal emulators are flakey, the machine hangs about once every two days. The debugger was written in the stone-age, the X server is almost as buggy as Apollo's share-mode X (more so actually, but smaller bugs). The SystemV merge is so screwed up that it's almost impossible to port software to it; but it's fast, it's cheap, and it's JLRU. You too can increase your productivity by setting the clock back 10 years. It's a sad world. -- Alfalfa Software, Inc. | Poste: The EMail for Unix nazgul@alfalfa.com | Send Anything... Anywhere 617/646-7703 (voice/fax) | info@alfalfa.com I'm not sure which upsets me more: that people are so unwilling to accept responsibility for their own actions, or that they are so eager to regulate everyone else's.
pphillip@cs.ubc.ca (Peter Phillips) (04/06/91)
In article <1991Apr5.023527.16647@alphalpha.com> nazgul@alphalpha.com (Kee Hinckley) writes: >In article <MISHKIN.91Apr3231301@jrst.apollo.hp.com> mishkin@jrst.apollo.hp.com (Nathaniel Mishkin) writes: >>the incredibly clunky), I can only hope that after a while, given the >>fact that X is so widespread and that the usability of what's out >>there on top of it is so bad, someone will build something that has >>the "incredibly useful" parts of the DM. >The only "someone" that is qualified (discounting Alfalfa, and we >don't have the money to finance the time it would take) is Apollo. >I should note that Apollo also has the greatest incentive. You've >got an awful lot of people in R&D who are refusing to switch to X >right now - and that means you're creating yet another situation (Unix >being the first) where Apollo isn't using the stuff it ships. Well, I understand that a lot of old time Apollo users would like to see DM stuff ported. As a new Apollo user interested in UNIX compatiblity, I'd rather see Apollo devote some of their diminishing resources to getting the bugs out of the current UNIX emulation under Domain/OS before they go and re-write the "Display Manager" for X. I'd rather have some psuedo-ttys that work properly than a window manager for X which likes to put a window on top of the current window and then beeps at you because a there is output in a window it just obscured :-). > > -kee > >Yes I'm ashamed to admit it but it's true - I'm typing this on a >Sun. NFS is a crock and slower than mollasses, all the terminal >emulators are flakey, the machine hangs about once every two days. >The debugger was written in the stone-age, the X server is almost >as buggy as Apollo's share-mode X (more so actually, but smaller >bugs). The SystemV merge is so screwed up that it's almost impossible >to port software to it; but it's fast, it's cheap, and it's JLRU. >You too can increase your productivity by setting the clock back >10 years. It's a sad world. Are Suns really that bad? Try getting "gdb" and the X11R4 release from MIT. If you find bugs, report them. As for NFS, it has at least two advantages over the Apollo file system: a) It is stateless. A curse and a blessing, I must admit, but at least diskless machines don't crash if the server does. On the other hand, you do have to live with "NFS server not responding..." occasionally. b) It seems to be portable. A lot of systems have some NFS compatibility, probably because of the public specification of the protocol and its relative independence on a particular underlying file system. Mere details. The *real* reason anyone wants to use UNIX or its many variations is because a lot of vendors try to support it. If you stick to one operating system supported by one vendor, whether it be MVS, VMS, Domain/OS, MS-DOS, TOPS-20, Macintosh, MTS, etc. then you (as an individual, a small business or a large corporation) are completely subject to the whims of that vendor. One day, Domain/OS is well and fine, the next day, its future is uncertain. With UNIX one does have the option of going to another vendor. You don't like Suns? Go and buy some DECstations, or MIPS machines, or IBM R/6000s, HP-UX, NeXT, PCs running SCO, etc. UNIX has many faults but in one way it is 10 years ahead of other operating systems: it isn't quite as proprietary. It is possible for any vendor to offer a version of UNIX because AT&T will provide the source code (for a substantial fee) and because the marketplace accepts the possibility of different implementations of the (sort of) same operating system. Come to think of it, the same reasons apply to X. For completeness, I should probably make some sort of motor vehicle analogy but I can't think of an apt one at the moment. -- Peter Phillips, pphillip@cs.ubc.ca | If an airplane crashes on the US/Canada {alberta,uunet}!ubc-cs!pphillip | border, where are the survivors buried?
nazgul@alphalpha.com (Kee Hinckley) (04/07/91)
In article <1991Apr6.072832.4012@cs.ubc.ca> pphillip@cs.ubc.ca (Peter Phillips) writes: >Well, I understand that a lot of old time Apollo users would like to >see DM stuff ported. As a new Apollo user interested in UNIX >compatiblity, I'd rather see Apollo devote some of their diminishing >resources to getting the bugs out of the current UNIX emulation under >Domain/OS before they go and re-write the "Display Manager" for X. Why? Why not have them spend the resources making OSF/1 better? I see no point at all in fixing Unix bugs in Domain/OS at this point in time. It's a dead-end. If you care about Unix then you'd rather see the Domain/OS features you like ported to OSF/1. Also, I'm not talking about putting the DM on X for the sake of old diehards. I'm talking about improving X by putting 10 years worth of graphical window-system experience to good use. >I'd rather have some psuedo-ttys that work properly than a window >manager for X which likes to put a window on top of the current window >and then beeps at you because a there is output in a window it just >obscured :-). How do you tell there's output in an X window if it's obscured? >Are Suns really that bad? Try getting "gdb" and the X11R4 release This is pretty funny! "Are Suns really that bad? Try running software that Sun didn't write." You've pretty much answered your own question :-). My impression in general is that Apollo's Unix has some fundamental flaws and some major bugs. Everyone else's Unix is basically all there, but has lots and lots of tiny little bugs. >from MIT. If you find bugs, report them. As for NFS, it has at least BTW. Where do I report them to? I haven't tried reading the docs yet. >two advantages over the Apollo file system: I don't particularly care to get into a features comparison. All I'm saying is that the Apollo file system is much more reliable, stable and fast. I have no illusions about having it anywhere else. I just hope the OSF/AFS learns from the mistakes of both Apollo and Sun's file systems. >implementations of the (sort of) same operating system. Come to think >of it, the same reasons apply to X. I've been using Unix for 10 years. I know the advantages (and disadvantages) by now. >For completeness, I should probably make some sort of motor vehicle >analogy but I can't think of an apt one at the moment. Unix is an old MG; fun, but not very practical. Apollo's job is to produce a Miata. A car with the pizzaz of an MG; but including all the modern conveniences that we've come to expect. Hey, what do you expect after being awake for two days? :-) -- Alfalfa Software, Inc. | Poste: The EMail for Unix nazgul@alfalfa.com | Send Anything... Anywhere 617/646-7703 (voice/fax) | info@alfalfa.com I'm not sure which upsets me more: that people are so unwilling to accept responsibility for their own actions, or that they are so eager to regulate everyone else's.
rees@dabo.citi.umich.edu (Jim Rees) (04/08/91)
In article <1991Apr6.072832.4012@cs.ubc.ca>, pphillip@cs.ubc.ca (Peter Phillips) writes:
I'd rather have some psuedo-ttys that work properly than a window
manager for X which likes to put a window on top of the current window
and then beeps at you because a there is output in a window it just
obscured :-).
I don't think anyone is defending the DM as a window manager. I use tvtwm
myself. The parts of the DM that I like are the fast, server-based
universal editor (you can do this with emacs but it's slow) and the pads.
As for NFS, it has at least
two advantages over the Apollo file system:
a) It is stateless. A curse and a blessing, I must admit, but at least
diskless machines don't crash if the server does. On the other hand,
you do have to live with "NFS server not responding..." occasionally.
I haven't heard anyone defend stateless file systems in a while. Even the
Sun file system guys are a little embarassed about statelessness now. I
think the future of file systems lies in large local caches. Your diskless
node won't crash, or even hang, if you've got everything cached locally. But
that's in the future. For right now, I prefer correctness over convenience,
so I stay away from NFS.
For completeness, I should probably make some sort of motor vehicle
analogy but I can't think of an apt one at the moment.
When I started running X, I went out and bought a '72 Buick that gets eight
miles to the gallon. It seemed like the appropriate thing to do.
By the way, I started writing a DM-pad emulator for X. It sort of almost
works. The problem is that it's hard to tell when the program at the other
end of the pty has eaten all its input.nazgul@alphalpha.com (Kee Hinckley) (04/09/91)
In article <50d9ecab.cb12@dabo.citi.umich.edu> rees@citi.umich.edu (Jim Rees) writes: >I don't think anyone is defending the DM as a window manager. I use tvtwm >myself. The parts of the DM that I like are the fast, server-based >universal editor (you can do this with emacs but it's slow) and the pads. I'll defend parts. The keydef capability is unmatched by any window manager other than gwm (actually, that might be a better choice than Mwm for building a psuedo-DM under X, but I'd rather see the improvements in Mwm). The ability to do everything from the keyboard is a major plus. Futhermore there are some features of the DM that are implemented poorly in other window managers. Mwm in particular could definitely use the DM's concept of Pop and NextWindow. I want a window manager where I can have back my keypad keydefs for laying out windows in a grid. On the other hand, I *don't* want single-cursor/ pointer-focus back (although if you're going to do pointer focus, I think single-cursor is better than two). > For completeness, I should probably make some sort of motor vehicle > analogy but I can't think of an apt one at the moment. > >When I started running X, I went out and bought a '72 Buick that gets eight >miles to the gallon. It seemed like the appropriate thing to do. I used to say that X set everything back a generation. My DN3500 runs like a DN300 using the DM. Now having used OL and Motif I think they set it back another generation. I get apps that run at about DN400 speed. (Maybe that's why the recent HP/Apollo was called a 400? :-) Ironicly, I think the platform independence of X is its major benefit. I could do without the remote display code. -- Alfalfa Software, Inc. | Poste: The EMail for Unix nazgul@alfalfa.com | Send Anything... Anywhere 617/646-7703 (voice/fax) | info@alfalfa.com I'm not sure which upsets me more: that people are so unwilling to accept responsibility for their own actions, or that they are so eager to regulate everyone else's.
rfh3273@galileo.rtn.ca.boeing.com (Dick Harrigill) (04/10/91)
In article <MISHKIN.91Apr3231301@jrst.apollo.hp.com> mishkin@jrst.apollo.hp.com (Nathaniel Mishkin) writes: >It's really nice to hear about how people like the DM and DOMAIN but >I'm afraid it's just very late in the game. Unless I've been dreaming, Apollo users have, since day 1, loudly proclaimed that they do NOT wish to lose functionality and have to constantly re-port software. It's really nice to hear about how HP listens to their customers, but I'm afraid it's just very untrue. Every ADUS conference since the merger has had users, sys admins & developers pleading (in writing) not to abandon us. >The world decided that >standard is better than better (and that even better in the form of >superset functionality is suspect). The world also decided that >cheaper (and better yet, free) is better than better. The rest -- you >must be standard and we're not going to let you make enough money (at >least in software) to do "better" as well as "standard" -- pretty much >follows on. What "world" do you live in? In the "real world" our objective is NOT to purchase the latest/greatest toys: it is to perform tasks as cost effective as possible to maximize profits. "Free" (as you desribe) is not necessarily better because it is not necessarily free (i.e. void of cost). With every "free" change comes millions of dollars in other costs including new hardware, software, training, loss of productivity during learning, possible loss of productivity due to loss of functionality, testing, verifying, trouble shooting, installing, coordinating - all of which cost $$$$. HP can give us, at no cost, the latest/greatest hardware, but if it doesn't run Domain/OS, have DM functionality AND run existing software, it will cost us dearly. Well, it's baseball season again and HP reminds me so much of the Seattle Mariners, a team with super potential that constantly gives its best players away and is thus a constant loser. Every other team in baseball seems to have a superstar former Mariner. Pretty soon, every other vendor will be selling products like hotcakes because they incorporate former Apollo features. -- Dick Harrigill, an independent voice from: Boeing Commercial Airplanes M/S 9R-49 PO BOX 3707 Renton Avionics/Flight Systems Seattle, WA 91824 Computing Support (206) 393-9539 rfh3273@galileo.rtn.ca.boeing.com
burdick@hpspdra.HP.COM (Matt Burdick) (04/11/91)
> >The world decided that standard is better than better > What "world" do you live in? In the "real world" our objective is NOT... The world in which Apollo was bought up by another company because it couldn't play the standards game well enough. -matt -- Matt Burdick | Hewlett-Packard burdick@hpspd.spd.hp.com | Intelligent Networks Operation
rand@HWCAE.CFSAT.HONEYWELL.COM (Douglas K. Rand) (04/15/91)
>>>>> On 9 Apr 91, nazgul@alphalpha.com (Kee Hinckley) in [Re: The Hardy Few] said: > I'll defend parts. The keydef capability is unmatched by any window manager > other than gwm [...] > [...] > I want a window manager where I can have back my keypad keydefs for laying > out windows in a grid. Since you mention GWM.... Thats what I'm using now. And I've got it where it pretty much does the things that I liked about the DM. I've got keys defined to move windows around in a grid. And the such. The only thing I haven't really figured out yet is the DM command ti. I think that GWM is what the DM should have grown up to be. -- Douglas Keenan Rand Honeywell -- Air Transport Systems Division Phone: +1 602 436 2814 US Snail: P.O. Box 21111 Phoenix AZ 85036 Internet: rand@ssdc.honeywell.com UUCP: ...!uunet!asuvax!apciphx!hwcae!rand
dave@jato.jpl.nasa.gov (Dave Hayes) (04/15/91)
rand@HWCAE.CFSAT.HONEYWELL.COM (Douglas K. Rand) writes: >Since you mention GWM.... Thats what I'm using now. And I've got it >where it pretty much does the things that I liked about the DM. I've >got keys defined to move windows around in a grid. And the such. The >only thing I haven't really figured out yet is the DM command ti. >I think that GWM is what the DM should have grown up to be. WHAT is GWM? Please...tell me....I'm DYIN here on this friggin' SUN they relegated me to! Replies by email, please. Thanks! -- Dave Hayes - dave@elxr.jpl.nasa.gov - ames!elroy!dxh When you have been your own teacher for a time, you may be ready to find someone else who can teach you.
wilsonj@gtephx.UUCP (Jay Wilson) (04/17/91)
I have been watching this little discussion for awhile now and up until now I have found little to disagree with. Well I have a problem with the comments that follow: > Are Suns really that bad? Try getting "gdb" and the X11R4 release > from MIT. If you find bugs, report them. As for NFS, it has at least > two advantages over the Apollo file system: > > a) It is stateless. A curse and a blessing, I must admit, but at least > diskless machines don't crash if the server does. On the other hand, > you do have to live with "NFS server not responding..." occasionally. > > b) It seems to be portable. A lot of systems have some NFS compatibility, > probably because of the public specification of the protocol and its > relative independence on a particular underlying file system. 1) Suns are really that bad. Over the last year we have lost 3 sys_admins to a new company here in town building a Sun network of 300 workstations. I keep in contact with these guys and they wish they were back supporting Apollos. 2) NFS has more problems than "NFS server not responding". Five years ago when we decided to go with Apollo workstations instead of Sun workstations, the Apollo won out because of ease of connectivity. Originally we were going to have a network of 500 workstations and everyone had to have the ability to talk to everyone else. Well we tested a small network of Suns and a small network of Apollos. Based on the amount of time needed to across mount all of the Sun systems, it would take 500 workstations 1 1/2 months to get everyone talking to everyone else if we took a power hit. The Apollos on the other hand would take about 10 mins. Additionally the management aspect of NS is much easier than NFS. With NFS you have to update each workstation's tables from that workstation. With NS you update using ctnode (from any workstation) and all of the other nodes will be able to find the node (if the address is not hashed already). We have a cron job on every node that refreshes the hash table twice aday. This is still better than going to every workstation and removing a mount point. As a side note we at one point had 900+ workstations and 8 sys_admins (112+:1). Over the last year we have scaled back to 583 workstations and 6 sys_admins (97:1). Lets see Sun sys_admins support that many workstations. ---------- Jay Wilson (wilsonj@gtephx) UUCP: {ncar!noao!asuvax | uunet!zardoz!hrc | att}!gtephx!wilsonj AG Communication Systems, Phoenix, AZ voice (602) 581-4496 fax (602) 581-4967 As usual, the comments above are solely mine and not necessarily that of AG Communication Systems.
jaf@Inference.COM (Jose Fernandez) (04/18/91)
Your posting seems to me to be unfairly critical of certain things that I
think may have been done a different way.
Originally we were going to have a network of 500 workstations and
everyone had to have the ability to talk to everyone else. Well we tested
a small network of Suns and a small network of Apollos. Based on the
amount of time needed to across mount all of the Sun systems, it would
take 500 workstations 1 1/2 months to get everyone talking to everyone
else if we took a power hit. The Apollos on the other hand would take
about 10 mins.
Automounting will significantly reduce the cross-mount time. Unfortunately, I
have no objective measurement of how well either an APOLLO or SUN node would
perform when a very large number of other nodes attempts to look into one of
its filesystems.
Additionally the management aspect of NS is much easier than NFS. With
NFS you have to update each workstation's tables from that workstation.
With NS you update using ctnode (from any workstation) and all of the
other nodes will be able to find the node (if the address is not hashed
already). We have a cron job on every node that refreshes the hash table
twice aday. This is still better than going to every workstation and
removing a mount point.
Huh? Shouldn't you be using NIS (Network Information Service, formerly known
as the Yellow Pages)?
As a side note we at one point had 900+ workstations and 8 sys_admins
(112+:1). Over the last year we have scaled back to 583 workstations and
6 sys_admins (97:1). Lets see Sun sys_admins support that many
workstations.
Well, if that's a challenge... Put up a bond for the salary price of 8 SA's,
plus SUN SW & HW maintenance... Just kidding. Maybe not. That's a hell of a
lot of money and would certainly be incentive...nazgul@alphalpha.com (Kee Hinckley) (04/19/91)
In article <335@texas.gtephx.UUCP> wilsonj@gtephx.UUCP (Jay Wilson) writes: >As a side note we at one point had 900+ workstations and 8 sys_admins (112+:1). >Over the last year we have scaled back to 583 workstations and 6 sys_admins >(97:1). Lets see Sun sys_admins support that many workstations. That's in line with what I've heard from another site. Going from Apollo networks to NFS/TCP networks they find they are going from a 100/1 sysadmin ratio to a 50/1 ratio. Twice as many sysadmins for the same number of machines. -- Alfalfa Software, Inc. | Poste: The EMail for Unix nazgul@alfalfa.com | Send Anything... Anywhere 617/646-7703 (voice/fax) | info@alfalfa.com I'm not sure which upsets me more: that people are so unwilling to accept responsibility for their own actions, or that they are so eager to regulate everyone else's.
nazgul@alphalpha.com (Kee Hinckley) (04/19/91)
In article <1991Apr18.171957.17924@alphalpha.com> nazgul@alphalpha.com (Kee Hinckley) writes: >ratio to a 50/1 ratio. Twice as many sysadmins for the same number of >machines. (Here I go, following up to myself :-) This was one of those things that Apollo was never able to figure out how to sell. When you get to the checkoff list you see "networked file system" and you check it for both Sun and Apollo. You don't discover things like scalability until long after you've committed one way or the other and by then it's far too late. -- Alfalfa Software, Inc. | Poste: The EMail for Unix nazgul@alfalfa.com | Send Anything... Anywhere 617/646-7703 (voice/fax) | info@alfalfa.com I'm not sure which upsets me more: that people are so unwilling to accept responsibility for their own actions, or that they are so eager to regulate everyone else's.