gors@well.UUCP (Gordon Stewart) (08/04/89)
I received an e-mail message from an IBMer asking me to mention a couple
of examples of AIX's peculiar commands --
those that deviate from Sys V by having non-standard names, or by having
(horror of horrors!) multi-character options! A clear violation
of the Sys V Standards ("improvements", if you ask IBM) ....
I would respond via e-mail, but my reply bounced. Check your address!
Two examples off the top of my head --
'print' (whatever happened to 'pr'?) has several multi-char options,
including 'dd' (drop dead) and 'ca' (for 'cancel')
'tctl' takes options 'rewind' 'retension' 'reset' etc....
Trust me when I say that these aren't the only ones!
Such "improvements" we can all live without!! It's still not "user-
friendly"! Assume, for a change, that even naive users have a modicum
of intelligence, and can figger out command options (or run 'man') --
but what happens when the new user writes a nifty shell program, moves
it to the /usr/local/bin directory, and finds that when he/she types
the command, he/she gets a "file not found" message -- still gotta do a
rehash! That's the sort of thing that could stand improvement --
instead of pissing in the drink to improve the flavor!
Michael Sierchio
"The opinions expressed herein aren't necessarily"
--
{apple, pacbell, hplabs, ucbvax}!well!gors
gors@well.sf.ca.us
(Doolan) | (Meyer) | (Sierchio) | (Stewart)karish@forel.stanford.edu (Chuck Karish) (08/05/89)
In article <13000@well.UUCP> gors@well.UUCP (Gordon Stewart) wrote: >I received an e-mail message from an IBMer asking me to mention a couple >of examples of AIX's peculiar commands -- >those that deviate from Sys V by having non-standard names, or by having >(horror of horrors!) multi-character options! > >Two examples off the top of my head -- > >'print' (whatever happened to 'pr'?) `pr' is there. I use it all the time. > has several multi-char options, > including 'dd' (drop dead) and 'ca' (for 'cancel') I don't much care for these, either. >'tctl' takes options 'rewind' 'retension' 'reset' etc.... These are quite similar to the options to the BSD `mt' command, which performs the same functions. This usage reflects prior art. >Trust me when I say that these aren't the only ones! Why should we? If we're going to put up with this ill-tempered babble, you might at least follow through and substantiate your criticism. >Such "improvements" we can all live without!! Neither of these are the gratuitous re-namings of existing SysV utilities you seemed to be complaining about in the original note. The `print' command is not in the SVID. On BSD systems, it's a trivial shell script that calls `pr', then `lpr'. Earlier versions of AIX came with the FTP utility named `xftp' and the TELNET utility named `tn'. The BSD `rsh' utility was called `remsh', to avoid conflict with the SysV restricted shell (`rsh'). As of 2.2.1, all three have the BSD names, and the restricted shell is called `Rsh'. >It's still not "user-friendly"! >Assume, for a change, that even naive users have a modicum >of intelligence, and can figger out command options (or run 'man') -- >but what happens when the new user writes a nifty shell program, moves >it to the /usr/local/bin directory, and finds that when he/she types >the command, he/she gets a "file not found" message -- still gotta do a >rehash! You're taking IBM to task for not re-writing the C shell? What should they call it, after they fix everything you don't like? They provided two alternate user interfaces with AIX (neither of which I use when I can avoid them) that are, arguably, more user-friendly than the C shell. >Michael Sierchio >"The opinions expressed herein aren't necessarily" I quite agree. Chuck Karish {decwrl,hpda}!mindcrf!karish (415) 493-9000 karish@forel.stanford.edu
buck@siswat.UUCP (A. Lester Buck) (08/05/89)
In article <13000@well.UUCP>, gors@well.UUCP (Gordon Stewart) writes: I am not a big fan of AIX, but these points are nitpicking. > Two examples off the top of my head -- > > 'print' (whatever happened to 'pr'?) has several multi-char options, > including 'dd' (drop dead) and 'ca' (for 'cancel') At least it still means to really print something. The Korn shell appropriated print to mean "new echo", which I consider horrific. One does wonder, though, how one uses the real AIX print command in the Korn shell... > 'tctl' takes options 'rewind' 'retension' 'reset' etc.... The vanilla SVR2 port on the NSC ICM-3216 has a command "mt", with essentially the same syntax and keywords. I consider this an improvement over picking one of the above keywords for the -r, and forever after trying to remember what the others were. > Michael Sierchio > {apple, pacbell, hplabs, ucbvax}!well!gors > gors@well.sf.ca.us If you want to complain about AIX, why not pick some real issues, like how IBM has no plans to sell source licenses EVER, or how OSF/1 will have a completely new driver interface "under the hood" that looks like VRM, so we all get to rewrite our drivers in the near future. -- A. Lester Buck ...!texbell!moray!siswat!buck
clarke@acheron.uucp (Ed Clarke/10240000) (08/05/89)
From article <4267@portia.Stanford.EDU>, by karish@forel.stanford.edu (Chuck Karish): > 2.2.1, all three have the BSD names, and the restricted shell is > called `Rsh'. > >>It's still not "user-friendly"! Thank you! I've been looking for 'Rsh' for a week now with little luck. I thought it was gone forever. If you DO want something that drives me wild, the missing '-i' flag on grep is a good candidate. I've replaced the standard IBM version with GNUgrep ( it's faster anyway ). Disclamer: these are my opinions, all mine and you can't have them. -- Ed Clarke acheron!clarke
gors@well.UUCP (Gordon Stewart) (08/06/89)
Apologies for occupying bandwidth with my (Michael Sierchio's) banter
-- my irritation is not so much with AIX per se, but the
misrepresentations of it by IBMers to me personally -- And it is vexing
to have to take weeks (literally) to solve some of the problems we've
been having with Product Support and Defect Support.
Netdom has been MUCH more helpful (cf. my Baseband Adapter queries)
than IBM -- All told, AIX represents IBM's hurried geste to catch the
ship as it was leaving the dock.
And if less time had been spent making putative improvements, it would
be a more robust and (dare I say?) useful product.
--
{apple, pacbell, hplabs, ucbvax}!well!gors
gors@well.sf.ca.us
(Doolan) | (Meyer) | (Sierchio) | (Stewart)nghiem@walt.cc.utexas.edu (Alex Nghiem) (08/07/89)
[flames about knocking AIX peculiarities] I think personal attacks against people who are upset about AIX are unwarranted. It takes a lot of time to document every single instance that AIX is peculiar, but if IBM keeps track of postings on the net long enough it should be clear what those incompatibilites are. Also, it should not be hard for IBM to do a comparitive analysis of AIX funtionality/performance with other BSD/SysV systems. The market place is flooded with documentation for the systems IBM is marketing the RT/AIX against. Also, there is a lot to be said about the bottom line--the RT has simply not sold as well as workstations produced by other vendors. It should not be to hard to determine why. The RT has not performed to users' satisfaction for lots of reasons, and word does get out through the grape wine. The net is part of that vine. It should be obvious that applications that users generally use on workstations simply do not run with out substantial rewrite on the RT. Signal incompatibilities needs to be looked into. Since scripts for other systems do not work, obviously a testing/comparison of the man pages for competing systems against AIX is warranted. Enough said--let's get back to business. (:->)
dyer@spdcc.COM (Steve Dyer) (08/07/89)
Let's also emphasize that if IBM had shipped every RT with 4.3BSD,
the marketplace would be a whole lot more favorable towards the RT.
OK, so there'd be some bitching about the C compiler, but there would
also be a whole lot more pressure coming from IBM to Metaware to get
a compiler which is bug-free. 4.3BSD for the RT is mostly a "no-surprises"
environment which integrates well into a network of Suns and VAXes
running Berkeley UNIX.
--
Steve Dyer
dyer@ursa-major.spdcc.com aka {ima,harvard,rayssd,linus,m2c}!spdcc!dyer
dyer@arktouros.mit.edubengsig@oracle.nl (Bjorn Engsig) (08/07/89)
Article <13000@well.UUCP> by gors@well.UUCP (Gordon Stewart) says: | |instead of pissing in the drink to improve the flavor! Is it necessary to use this language just because the subject has to do with IBM? | |Michael Sierchio Are you the same person as Gordon Stewart? | To add to the subject of multiletter options, can you mention anything more ugly than the 'stty everything' needed on the Sun to give *almost* every tty setting. It's not only IBM doing that, and I don't find your examples that bad. If you have some examples as bad as 'stty everything', please let's all know about it. -- Bjorn Engsig, ORACLE Europe \ / "Hofstadter's Law: It always takes Path: mcvax!orcenl!bengsig X longer than you expect, even if you Domain: bengsig@oracle.nl / \ take into account Hofstadter's Law"
jon@shaffer.UUCP (Jon Doran/60000) (08/09/89)
Gee, I should have included the referenced article ... too late. In responce to the individual who offered two examples of non-standard commands in AIX (print and tctl): "pr" is still around, it never sent anything to the printer before, it is just a filter. "print" is a bit weird, but I've never seen too much standardization in this department. I remember UCB changing the print command every release... Anyway this is a valid point, but if it bothers you make a shell script; that is what I did. "tctl" is the streaming tape control program. This is hardware dependant, and has not been standardized to the best of my knowledge. I remember several different vendors with similar programs, none of which have used the same name. Unix has several hardware dependant programs: 4014, 300, 450 and other similar stuff in /usr/bin. I'm sure these caused similar uproar upon introduction. In summary, the people who work on AIX development have fought hard to keep the names and behavior of Unix utilities the same as on other versions of Unix. I'm sure there are going to be minor differences show up, it would be a miracle if they didn't. But IBM is actively seeking to avoid this. If any of the folks using AIX have any other examples, please share them with the rest of us. Only this way can they be corrected. Jon Doran IBM AWD, Austin TX
jon@shaffer.UUCP (Jon Doran/60000) (08/09/89)
In article <16655@ut-emx.UUCP>, nghiem@walt.cc.utexas.edu (Alex Nghiem) writes: > Also, there is a lot to be said about the bottom line--the RT ^^^^^^ > has simply not sold as well as workstations produced by other ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > vendors. It should not be to hard to determine why. The RT ^^^^^^^^ OK, so IBM is not winning any popularity contests. And I agree that there is plenty of room for improvement... But IBM is ranked 4th in workstation sales. I'm not sure if this is by unit or by sales. 1. HP (includes Apollo Div) 2. Sun 3. ??? 4. IBM ? SGI ? DEC It rather surprised me, but says something. Jon Doran IBM AWD, Austin TX
abstine@image.soe.clarkson.edu (Arthur Stine) (09/06/89)
From article <2534@shaffer.UUCP>, by jon@shaffer.UUCP (Jon Doran/60000): > In article <16655@ut-emx.UUCP>, nghiem@walt.cc.utexas.edu (Alex Nghiem) writes: > >> Also, there is a lot to be said about the bottom line--the RT > ^^^^^^ >> has simply not sold as well as workstations produced by other > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ >> vendors. It should not be to hard to determine why. The RT > ^^^^^^^^ > > OK, so IBM is not winning any popularity contests. And I agree > that there is plenty of room for improvement... But IBM is > ranked 4th in workstation sales. I'm not sure if this is by > unit or by sales. > 1. HP (includes Apollo Div) > 2. Sun > 3. ??? > 4. IBM > the ??? would be DEC. The percentage of installed systems that I've seen mentioned in recent trade-rags in something like: 1 HP/Apollo = > 30% 2 Sun ~ 30% 3 DEC ~ 23% 4 IBM ~ 2% and then all the others. -- Art Stine Sr Network Engineer Clarkson U ABStine@CLVMS.Clarkson.Edu
phil@ingr.com (Phil Johnson) (09/22/89)
In article <1989Sep5.193948.14894@sun.soe.clarkson.edu> abstine@image.soe.clarkson.edu (Arthur Stine) writes: >From article <2534@shaffer.UUCP>, by jon@shaffer.UUCP (Jon Doran/60000): >> In article <16655@ut-emx.UUCP>, nghiem@walt.cc.utexas.edu (Alex Nghiem) writes: -- stuff deleted -- >> OK, so IBM is not winning any popularity contests. And I agree >> that there is plenty of room for improvement... But IBM is >> ranked 4th in workstation sales. I'm not sure if this is by >> unit or by sales. >> 1. HP (includes Apollo Div) >> 2. Sun >> 3. ??? >> 4. IBM >> >the ??? would be DEC. The percentage of installed systems that I've seen >mentioned in recent trade-rags in something like: > >1 HP/Apollo = > 30% >2 Sun ~ 30% >3 DEC ~ 23% >4 IBM ~ 2% > The current ranking of U.S. workstation vendors for the 1988-89 analysis period is as follows: NOTE - 1) These rankings are restricted to Workstation products. 2) Nnumbers in parenthesies denote the ranking, for a specific category, that differs from the general ranking (Installed based). Installed Units Revenues Base Shipped (Equipment/services) 1. Sun 128,000 33.1% 69,000 37.2% $1,462 M 32.7% 2. Apollo 83,010 21.5% (3) 29,010 15.6% (4) $ 654 M 14.6% 3. DEC 72,500 18.8% (2) 41,600 22.4% $ 993 M 20.9% 4. HP 54,200 14.0% 24,275 13.1% (3) $ 669 M 15.0% 5. Intergraph 17,727 4.1% 9,895 5.3% $ 329 M 7.4% 6. SGI 12,105 3.1% 4,555 2.5% $ 199 M 4.5% 7. IBM 8,500 2.2% 3,900 2.1% $ 82 M 1.8% -- Philip E. Johnson UUCP: usenet!ingr!b3!sys_7a!phil MY words, VOICE: (205) 772-2497 MY opinion!