george@osc.COM (George Baggott) (06/27/90)
Does anyone out there know anything about putting an Open Look look and feel on a motif application? In an ideal world I would be able to link against some Open Look version of Xm and my application would be ported to open look without my having to change a line of code. Does such a thing exist? I am aware of Solbourne's OI Toolkit which supports both Motif and Open Look from a single C++ toolkit. It looks neat, but I would prefer to avoid throwing away my motif and UIL code and starting from scratch. Any info would be greatly appreciated. I will post a summary of whatever replies I get... Thanks in advance, George
willis@ecoult.ncsu.edu (Bill Willis) (06/28/90)
In article <2986@osc.COM> george@osc.com (George Baggott) writes: >Does anyone out there know anything about putting an Open Look look and feel on >a motif application? In an ideal world I would be able to link against some >Open Look version of Xm and my application would be ported to open look without >my having to change a line of code. Does such a thing exist? I am aware of >Solbourne's OI Toolkit which supports both Motif and Open Look from a single >C++ toolkit. It looks neat, but I would prefer to avoid throwing away my motif >and UIL code and starting from scratch. Any info would be greatly appreciated. >I will post a summary of whatever replies I get... I think that the two are mutually exclusive. Each GUI is first a look and feel then a toolkit to impleme the required widgets. They are different enough that this is probably never going to happen... Stick with Motif... It looks like it will be around longer and touch more platforms... Bill Willis, Director willis@ecovsa.ncsu.edu Engineering Computer Operations willis%ecovsa@ncsuvx.ncsu.edu Box 7901, School of Engineering
mayer@hplabsz.HPL.HP.COM (Niels Mayer) (06/29/90)
In article <2986@osc.COM> george@osc.com (George Baggott) writes: >Does anyone out there know anything about putting an Open Look look and feel on >a motif application? In an ideal world I would be able to link against some >Open Look version of Xm and my application would be ported to open look without >my having to change a line of code. What about style guide compliance?? I can imagine a switchable look on the UI-components (widgets), but I think you'd need even higher level abstractions to capture differing layout styles, application-dialog-level interaction, etc. Such higher-level abstractions might be possible by employing UIMS (*) technologies to provide a more abstract description of application dialog. > I am aware of > Solbourne's OI Toolkit which supports both Motif and Open Look from a single > C++ toolkit. It looks neat, but I would prefer to avoid throwing away my motif > and UIL code and starting from scratch. To get style-guide level compliance, I think you'd have to start over again anyways. [footnote(*): I'm talking about "real UIMSs" the term has been weakened lately by marketeers calling direct manipulation builders UIMSs.] ------------------------------------------------------------------------------- Niels Mayer -- hplabs!mayer -- mayer@hplabs.hp.com Human-Computer Interaction Department Hewlett-Packard Laboratories Palo Alto, CA. *
patl@bodacia.Eng.Sun.COM (Pat Lashley) (06/29/90)
In article <1990Jun28.111050.17353@ncsuvx.ncsu.edu> willis@ecoult.ncsu.edu.UUCP (Bill Willis) writes: > > ... > >Stick with Motif... It looks like it will be around longer and touch >more platforms... A common mis-perception. Let me present the following counter-arguments: 1) Count the number of applications actually shipping. OpenLook wins by almost 2:1. Count the number of SunView applications total. Easily the second most widely used look and feel in the world. Think about how easy most of them are to port using xview.... 2) The XView OpenLook toolkit source is freely-redistributable; and included with the X11R4 source; you have to pay for Motif. Not just to develop using it, but a per-customer right-to use. I'm not about to spend the bucks for a Motif developer's licence for my Amiga at home when I will be able to get the OpenLook toolkits with X11. 3) Two OpenLook toolkits are provided with the SysV R.4 distribution and in the Sun OpenWindows package. (OpenWindows is available to Sun users for a nominal media charge. The right-to-use is included in your SunOS licence.) 4) OpenLook defines a 2D look which works much better on monochrome and low-res displays. (Have you ever seen Motif on monochrome? Fuzzy looking isn't it...) 5) OpenLook has a smoother, more-consistant user interface. (Ok, this one borders on opinion instead of pure fact. :-) -- X-Face: #FowkUVVz[9{ux;7z%!?7>\5DCdVqaja5uk!4Z~)5*f@-"n&||t35?wVN+UloPr-Q;iR\;t snA%,sJ:+$a[eV(aKz4\=`MIH#{`/#HW>TT6Hx=Xp06oj>ta|]bFa'1BiI5Wj_y7n,l)tFuEd(oE`V 3w'0..-`[}nX:VVJ&@Br$cCu|/iqA4VC}/APx:gge9-fj(@V*~W[L@KP@^AcXvel])1%zy[&c}t"\z :X,J8<1D%I;J>tY6EZ7lx,8R&JhgPyZ4Zz[3J`#N@zc&d<"V+&O*;gRd^)xC`34h8[!Vb+
thomson@cs.utah.edu (Rich Thomson) (06/29/90)
In article <138082@sun.Eng.Sun.COM> patl@bodacia.Eng.Sun.COM (Pat Lashley) writes: >4) OpenLook defines a 2D look which works much better on monochrome and > low-res displays. (Have you ever seen Motif on monochrome? Fuzzy > looking isn't it...) I don't find anything "fuzzy" about monochrome Motif. We use it at E&S all the time on our NCD xterms and it looks just fine. Of course it looks even better on an ESV ;-}. >X-Face: #FowkUVVz[9{ux;7z%!?7>\5DCdVqaja5uk!4Z~)5*f@-"n&||t35?wVN+UloPr-Q;iR\;t > snA%,sJ:+$a[eV(aKz4\=`MIH#{`/#HW>TT6Hx=Xp06oj>ta|]bFa'1BiI5Wj_y7n,l)tFuEd(oE`V > 3w'0..-`[}nX:VVJ&@Br$cCu|/iqA4VC}/APx:gge9-fj(@V*~W[L@KP@^AcXvel])1%zy[&c}t"\z > :X,J8<1D%I;J>tY6EZ7lx,8R&JhgPyZ4Zz[3J`#N@zc&d<"V+&O*;gRd^)xC`34h8[!Vb+ Just what is this thing? A bitmap of your face or something? -- Rich Rich Thomson thomson@cs.utah.edu {bellcore,hplabs,uunet}!utah-cs!thomson ``If everybody is thinking the same thing, is anybody thinking?'' --Bob Johnson
willis@ecoult.ncsu.edu (Bill Willis) (06/29/90)
I didn't mean to start something, but here goes... In article <138082@sun.Eng.Sun.COM>, patl@bodacia.Eng.Sun.COM (Pat Lashley) writes: > > 1) Count the number of applications actually shipping. OpenLook wins > by almost 2:1. Count the number of SunView applications total. > Easily the second most widely used look and feel in the world. > Think about how easy most of them are to port using xview.... I was counting X applications only and counting those that I know of under development. I do not believe that the 2:1 number holds. For another measure, and remember we need to pick what will be delivering next fall instead of now, count the number of time Motif is mentioned in your favorite, non-biased trade journal and count the time OpenLook is mentioned. This may seem trivial, but I think it represents something. > > 2) The XView OpenLook toolkit source is freely-redistributable; and > included with the X11R4 source; you have to pay for Motif. Not > just to develop using it, but a per-customer right-to use. I'm > not about to spend the bucks for a Motif developer's licence for > my Amiga at home when I will be able to get the OpenLook toolkits > with X11. I pay nothing for Motif (yes it is imbedded in the workstation price) unless I want source. The libraries come with most of the workstations that I order and it is supported by the vendor. I do not HAVE to get R4 or build it or mess with anything. Now, I know that Motif isn't very mature yet. In fact, it leaks like a sieve in many cases, but, it will improve, and we have applications running under Motif running on 7 platforms now and we only had to build it ourselves for the Sun systems. > > 3) Two OpenLook toolkits are provided with the SysV R.4 distribution and > in the Sun OpenWindows package. (OpenWindows is available to Sun > users for a nominal media charge. The right-to-use is included in > your SunOS licence.) Again, Motif toolkits are provided with nearly everything we are interested in. If it isn't provided now, it soon will be, or it is available for the asking. DG, Digital, HP/Apollo, IBM, Solbourne, SCO, etc.... The only significant player missing is Sun. > > 4) OpenLook defines a 2D look which works much better on monochrome and > low-res displays. (Have you ever seen Motif on monochrome? Fuzzy > looking isn't it...) We have defined monochrome resources which yield a perfectly acceptable appearance for Motif.... > > 5) OpenLook has a smoother, more-consistant user interface. (Ok, this one > borders on opinion instead of pure fact. :-) It doesn't border on opinion, it clearly is an opinion :-) BTW, many of the statements that I make here are also opinions. But, they come from a lot of experience in this market place. I will not argue technical superiority of either toolkit. We have used both and found significant problems with both. But from the marketplace and recent announcements, the race nearly is over and the leader is clear. I hear that UI is close to endorsing Motif as a toolkit. As I understand it they only have to see that it works under SysV R.4. It will... I would not be surprised to see many major customers requiring vendor supplied Motif libraries in the near future. Bill Willis, Director willis@ecovsa.ncsu.edu Engineering Computer Operations willis%ecovsa@ncsuvx.ncsu.edu Box 7901, School of Engineering North Carolina State University (919) 737-2458 We have met the enemy, and they are us --- Pogo
smith@canon.co.uk (Mark Smith) (06/29/90)
patl@bodacia.Eng.Sun.COM (Pat Lashley) writes: > 4) OpenLook defines a 2D look which works much better on monochrome and > low-res displays. (Have you ever seen Motif on monochrome? Fuzzy > looking isn't it...) > > 5) OpenLook has a smoother, more-consistant user interface. (Ok, this one > borders on opinion instead of pure fact. :-) *** you guessed it - yet another flame on *** Open Look as it appears on SunOS could be a really nice looking interface if you guys at Sun would get a decent graphic designer in to finish the job. I really respect the technical achievement behind making the merged server work, but it's just not good enough to have Bill or Scott whip up a trash can icon over lunch. I could scream every time somebody at Sun responds to the criticism that Motif just looks more pleasing by saying "well, scientific studies have shown..." Aesthetics matters, and I see nothing that suggests anyone at Sun is really making sure that the final product has that extra bit of polish. And, yes, I have seen OW2. (It's a shame that some of the work done on the 386i didn't migrate to other Sun machines - look at the graphical quality of the Mailtool icons compared to the SPARC SunOS versions.) *** flame off *** Sorry, this isn't really X stuff. I'll go away now. ==================================================================== Mark Smith Canon Research Centre Europe smith@canon.co.uk 19 Frederick Sanger Road ..ukc!uos-ee!canon!smith Guildford Surrey UK -------------------------------------------------------------------- "We always planned to ease ourselves into pure research anyway..." -- David Cronenberg's _Dead Ringers_
ben@hpcvlx.cv.hp.com (Benjamin Ellsworth) (06/30/90)
> 1) Count the number of applications actually shipping. OpenLook wins > by almost 2:1. Count the number of SunView applications total. > Easily the second most widely used look and feel in the world. By some estimates, there are more lines of program written in COBOL currently running in the the world than all other languages (execpt maybe Fortran) combined. Do we therefore conclude that COBOL is the language of choice for new program development? ----------------------------------------------------------------------- Benjamin Ellsworth | ben@cv.hp.com | INTERNET Hewlett-Packard Company | {backbone}!hplabs!hp-pcd!ben | UUCP 1000 N.E. Circle | (USA) (503) 750-4980 | FAX Corvallis, OR 97330 | (USA) (503) 757-2000 | VOICE ----------------------------------------------------------------------- All relevant disclaimers apply. -----------------------------------------------------------------------
murray@flatirons.central.sun.COM (Murray Stein [Rocky Mtn. Regional MTS]) (06/30/90)
+----- Begin Included Message -----
+From: news@ncsuvx.ncsu.edu (Bill Willis)
+Subject: Re: Motif -> Open Look look & feel
+Message-Id: <1990Jun29.033504.5003@ncsuvx.ncsu.edu>
+To: xpert@expo.lcs.mit.edu
+Status: R
+I didn't mean to start something, but here goes...
+In article <138082@sun.Eng.Sun.COM>, patl@bodacia.Eng.Sun.COM (Pat
+Lashley) writes:
+>
+> 1) Count the number of applications actually shipping. OpenLook wins
+> by almost 2:1. Count the number of SunView applications total.
+> Easily the second most widely used look and feel in the world.
+> Think about how easy most of them are to port using xview....
+I was counting X applications only and counting those that I know of
+under development. I do not believe that the 2:1 number holds. For
+another measure, and remember we need to pick what will be delivering
+next fall instead of now, count the number of time Motif is mentioned in
+your favorite, non-biased trade journal and count the time OpenLook is
+mentioned. This may seem trivial, but I think it represents something.
Nope. 2:1 is right. And we can prove it. I do agree that the real measure will
be the applications delivered in the coming 6-12 months. BTW, OPEN LOOK has something
like 300 applications in development.
As for the number of mentions for Motif vs OPEN LOOK, I really doubt that this is a
good measure, given the independent aliases and support mechanisms.
+>
+> 2) The XView OpenLook toolkit source is freely-redistributable; and
+> included with the X11R4 source; you have to pay for Motif. Not
+> just to develop using it, but a per-customer right-to use. I'm
+> not about to spend the bucks for a Motif developer's licence for
+> my Amiga at home when I will be able to get the OpenLook toolkits
+> with X11.
+I pay nothing for Motif (yes it is imbedded in the workstation price)
+unless I want
+source. The libraries come with most of the workstations that I order and it is
+supported by the vendor. I do not HAVE to get R4 or build it or mess
+with anything. Now,
+I know that Motif isn't very mature yet. In fact, it leaks like a sieve
+in many cases,
+but, it will improve, and we have applications running under Motif
+running on 7
+platforms now and we only had to build it ourselves for the Sun systems.
What are you smoking ? The fact is, one way or another, you do pay for it. OSF/Motif
requires royalties. So maybe you didn't pay it directly - but you can be sure your
vendor did, and he is going to pass the cost on to you somehow.
+>
+> 3) Two OpenLook toolkits are provided with the SysV R.4 distribution and
+> in the Sun OpenWindows package. (OpenWindows is available to Sun
+> users for a nominal media charge. The right-to-use is included in
+> your SunOS licence.)
+Again, Motif toolkits are provided with nearly everything we are
+interested in. If it
+isn't provided now, it soon will be, or it is available for the asking.
+DG, Digital,
+HP/Apollo, IBM, Solbourne, SCO, etc.... The only significant player
+missing is Sun.
With all due respect to all the competent and committed vendors out there, I honestly
don't believe that you can stack alot of names together and say 'there's the marketplace'.
How much of the workstation market is DG anyway ? Or IBM for that matter ?
+>
+> 4) OpenLook defines a 2D look which works much better on monochrome and
+> low-res displays. (Have you ever seen Motif on monochrome? Fuzzy
+> looking isn't it...)
+We have defined monochrome resources which yield a perfectly acceptable
+appearance
+for Motif....
+>
+> 5) OpenLook has a smoother, more-consistant user interface. (Ok, this one
+> borders on opinion instead of pure fact. :-)
+It doesn't border on opinion, it clearly is an opinion :-)
+BTW, many of the statements that I make here are also opinions. But,
+they come from a
+lot of experience in this market place. I will not argue technical
+superiority of
+either toolkit. We have used both and found significant problems with
+both. But from the
+marketplace and recent announcements, the race nearly is over and the
+leader is clear.
+I hear that UI is close to endorsing Motif as a toolkit. As I understand
+it they only
+have to see that it works under SysV R.4. It will...
+I would not be surprised to see many major customers requiring vendor supplied
+Motif libraries in the near future.
I think that the the race has hardly begun. I believe that your impression that UI
is preparing to commit to Motif is wrong.
I believe that DEC and HP have done a fantastic job in marketing Motif. And there is
no substitute for creating fear, uncertainty and doubt, when it comes to marketing. But
have no doubt, what you are witnessing is marketing, and very little else...
Motif defines not just a look & feel, but an API. Frankly, given the fact that we are
only just on the threshhold of embracing multi-media technology, isn't it a bit early
to trot down a path that confines applications to a given API ? Oh, I guess we could
have decided that the IBM PC UI of circa 1981 was the absolute in computing, and then
never bothered to do the Macintosh. Or the NeXT for that matter...
Things are going to continue to change, as they always do. More and more non-technical
users are going to use workstations (I hope), whether they are running OPEN LOOK *or*
Motif. And the common look & feel will be what matters to those users, not the API.
The OPEN LOOK UI does not define an API, but it does attempt to provide a paradym for
the interaction of applications - important when you consider that 'smart' applications
need to integrate together if we ever want to see the kind of productivity on workstations
under Xwindows that we see on the Mac and NeXT.
Both Motif and OPEN LOOK will have their successes, and there is nothing like a little
healthy competition...
+----- End Included Message -----
[My opinions are most definately my own....]ries@venice.SEDD.TRW.COM (Marc Ries) (06/30/90)
In article <138082@sun.Eng.Sun.COM> patl@bodacia.Eng.Sun.COM (Pat Lashley) writes: ->In article <1990Jun28.111050.17353@ncsuvx.ncsu.edu> willis@ecoult.ncsu.edu.UUCP (Bill Willis) writes: ->> ... ->>Stick with Motif... It looks like it will be around longer and touch ->>more platforms... ->A common mis-perception. Let me present the following counter-arguments: ->1) Count the number of applications actually shipping. OpenLook wins -> by almost 2:1. Count the number of SunView applications total. -> Easily the second most widely used look and feel in the world. -> Think about how easy most of them are to port using xview.... Maybe if the counter-argument is "Count the number of applications actually shipping ON SUNs" and I would be more inclined to agree. If one is not running Suns/ATTs/SolBs, then as an end-user/developer I see it just the opposite: It's more like 3:1 MOTIF vs. OpenLook. Personally, I care less about WHICH GUI is being used versus what APPLICATIONS/DEVELOPMENT TOOLS ACTUALLY EXIST for my particular hardware platform.
guy@auspex.auspex.com (Guy Harris) (07/01/90)
Please keep your lines within 80 columns, thanks....
>Motif defines not just a look & feel, but an API.
Well, actually, OSF apparently now has a certification process for
toolkits, so they've finally realized that "The Toolkit is *NOT* the
Look&Feel". Given that there are non-Xm toolkits under development (and
some may already exist) that offer a Motif L&F, this would seem to be a
sensible acknowledgment of reality....rmf@media.uucp (Roger Fujii) (07/01/90)
willis@ecoult.ncsu.edu (Bill Willis) writes: >> >> 2) The XView OpenLook toolkit source is freely-redistributable; and >> included with the X11R4 source; you have to pay for Motif. Not >> just to develop using it, but a per-customer right-to use. I'm >> not about to spend the bucks for a Motif developer's licence for >> my Amiga at home when I will be able to get the OpenLook toolkits >> with X11. >I pay nothing for Motif (yes it is imbedded in the workstation price) >unless I want >source. The libraries come with most of the workstations that I order and it is >supported by the vendor. I do not HAVE to get R4 or build it or mess >with anything. Now, >I know that Motif isn't very mature yet. In fact, it leaks like a sieve >in many cases, >but, it will improve, and we have applications running under Motif >running on 7 >platforms now and we only had to build it ourselves for the Sun systems. There are several vendors who DO NOT bundle Motif with their systems. This poses several non-trivial problems. Let's say a person writes an application with uil. He must choose one of these headaches: 1) the application will be HUGE (no shared libraries) 2) distribute the shared libraries with the application and pay OSF their distribution fee. 3) Compatibility risks because the shared library that the user provides (either vendor supplied or self generated) doesn't work like they should. Using publically distributable components provide another alternative: I can provide EVERYTHING that is needed to assure that the product runs. Granted that this may seem extreme, but there are *lots* of vendor supplied X servers that don't work quite right (backing store is broken, event handling is messed up....), and this may be the *only* way of making your program run correctly. >> >> 3) Two OpenLook toolkits are provided with the SysV R.4 distribution and >> in the Sun OpenWindows package. (OpenWindows is available to Sun >> users for a nominal media charge. The right-to-use is included in >> your SunOS licence.) >Again, Motif toolkits are provided with nearly everything we are >interested in. If it >isn't provided now, it soon will be, or it is available for the asking. >DG, Digital, >HP/Apollo, IBM, Solbourne, SCO, etc.... The only significant player >missing is Sun. Apollos don't come with uil from what I heard. Real Soon Now doesn't help if what you need is needed today. If everyone can get it for free, then great. But if one has to PAY for them...... >> >> 5) OpenLook has a smoother, more-consistant user interface. (Ok, this one >> borders on opinion instead of pure fact. :-) >It doesn't border on opinion, it clearly is an opinion :-) >BTW, many of the statements that I make here are also opinions. But, >they come from a >lot of experience in this market place. I will not argue technical [ stuff deleted ] Actually, OpenLook should be more consistant because their style guides are *infinitely* better than the one OSF provides. Motif is great if you have a SINGLE monolithic program, but it sucks big time if you have a multi-window application. The guides are very vague about when you should do things and when you don't. I am not impress with the restrictions that their widget set provides.... >I would not be surprised to see many major customers requiring vendor supplied >Motif libraries in the near future. sad, but true. It is a shame that a windowing system whose idea is that the interface should fit the user (and not the other way around) would be condemned to such a fate. Personally, I think it would be great if vendors provided multiple interfaces and the the user pick and choose, but OSF licensing would make this too cumbersome. -- Roger Fujii - Media Cybernetics Phone: (301)495-3305 Internet: rmf%media@uunet.uu.net UUCP: {uunet,hqda-ai}!media!rmf
toml@ninja.Solbourne.COM (Tom LaStrange) (07/02/90)
|> >Motif defines not just a look & feel, but an API. |> |> Well, actually, OSF apparently now has a certification process for |> toolkits, so they've finally realized that "The Toolkit is *NOT* the |> Look&Feel". Given that there are non-Xm toolkits under development (and |> some may already exist) that offer a Motif L&F, this would seem to be a |> sensible acknowledgment of reality.... I just got the certification package and from what I can tell it is to certify a port of Xm to a given vendor's hardware. It is an API certification process, not a "Look&Feel" certification process. So it seems that the only certified OSF/Motif toolkits will be Xm ports. If I have interpreted this wrong I would certainly like someone from OSF to let me know. -- Tom LaStrange Solbourne Computer Inc. ARPA: toml@Solbourne.COM 1900 Pike Rd. UUCP: ...!{boulder,sun}!stan!toml Longmont, CO 80501
jimf@SABER.COM (07/02/90)
| +DG, Digital, | +HP/Apollo, IBM, Solbourne, SCO, etc.... The only significant player | +missing is Sun. [...] |How much of the workstation market is DG anyway ? Or IBM for that matter ? HP/Apollo and DEC form quite a large chunk, I daresay larger than Sun. Nor would I count IBM out (regardless of my personal opinions towards the company). From my standpoint, there is one major player -- Sun -- pushing Open Look. Every other major player is pushing Motif. We market a product on Sun and DEC platforms now, with IBM and HP ports in progress and the potential for several others. Of these vendors and potential vendors, only Sun wants Open Look. We don't want to abandon Sun, our major platform today, but can we realistically abandon *every* other major vendor out there? I think not. Neither Open Look nor Motif is a good environment for developing complex applications, and they differ enough that it would be difficult to support a product which makes use of both of them. The lack of quality in both XView and Open Look makes development tough enough for either (doing both is nigh impossible). Personally I think that Sun/AT&T should abandon their UI crusade so that we developers can get on with our lives. If they instead built a competing Motif library (or perhaps a merged OL/Motif) and put it out there with the same terms as XView, we'd get around the painful licensing of Motif and the moronic competing-UI problems that we developers have to deal with now. But that's wishful thinking, I know. Happy hacking, Jim frost saber software jimf@saber.com
alan@cogswell.Jpl.Nasa.Gov (Alan S. Mazer) (07/02/90)
In article <100920232@hpcvlx.cv.hp.com>, ben@hpcvlx.cv.hp.com (Benjamin Ellsworth) writes: | | > 1) Count the number of applications actually shipping. OpenLook wins | > by almost 2:1. Count the number of SunView applications total. | > Easily the second most widely used look and feel in the world. | | By some estimates, there are more lines of program written in COBOL | currently running in the the world than all other languages (execpt | maybe Fortran) combined. Do we therefore conclude that COBOL is the | language of choice for new program development? No, you conclude that you better get a computer that runs COBOL if you want access to that software. If you want to rewrite that software in C, go ahead, but don't expect _me_ to pay you to do it. I'd rather buy a COBOL-running machine. Just as I'd rather use a window system for which there is already software available. -- Alan # <this space available> ..!ames!elroy!alan alan@elroy.jpl.nasa.gov
chan@hpfcmgw.HP.COM (Chan Benson) (07/03/90)
>>Motif defines not just a look & feel, but an API. > >Well, actually, OSF apparently now has a certification process for >toolkits, so they've finally realized that "The Toolkit is *NOT* the >Look&Feel". Given that there are non-Xm toolkits under development (and >some may already exist) that offer a Motif L&F, this would seem to be a >sensible acknowledgment of reality.... Indeed, I believe the people who wrote Looking Glass (a desktop manager for X11) used their own Motifish library. (If I'm wrong please correct me; that was the impression I got from a trade rag article and we all know how accurate they can be). Also, OpenDialog from the Apollo folks has a Motif mode that is not really Motif. -- Chan
drl@VUSE.VANDERBILT.EDU (David R. Linn) (07/03/90)
>Personally I think that Sun/AT&T should abandon their UI crusade so >that we developers can get on with our lives. If they instead built a >competing Motif library (or perhaps a merged OL/Motif) and put it out >there with the same terms as XView, we'd get around the painful >licensing of Motif and the moronic competing-UI problems that we >developers have to deal with now. Does anyone know of "merged OL/Motif" packages beside: - N3 (AT&T/USO) - OI (Solbourne + others) - Interviews V3 (proposed) David
price@wsl.dec.com (Chuck Price) (07/03/90)
Some Sun afficionado writes: > With all due respect to all the competent and committed vendors out there, > I honestly > don't believe that you can stack alot of names together and say > 'there's the marketplace'. > How much of the workstation market is DG anyway ? Or IBM for that matter ? String them all together and you get a marketplace much larger than Sun's. Sun *might* have 25% of the Unix W/S marketplace today. They are the only one not doing Motif. Digital alone, and Digital and H/P combined in particular, make for a significant market. And anyone who discounts IBM as a player is just being silly. <Biased opinion> The claim that OPEN LOOK is better because you can get XView free clinches the argument for me. You get what you pay for. I'd suggest that Motif is the hands down winner. My prediction is that it will be bundled on every major hardware manufacturer's platform by this time next year. <Bias off> <Opinion> Software vendors *want* a common API. That is the problem with OPEN LOOK. It doesn't have one. I personally don't have an issue with OPEN LOOK. As long as a S/W vendor is running on the DECstation, you'll make me happy. My personal recommendation is Motif, because I believe there is more thought and technology behind it. But I saw OPEN LOOK running on a DECstation in Sun's booth at Xhibition, and it was faster than OPEN LOOK running on their own Sun workstations. If a software vendor wants to use it on our equipment, more power to them. But I submit that there is less risk to a software vendor in using a system which has the support of the vast majority of system suppliers in the industry, versus a plethora of different implementations of a look and feel supported by the minority. <End Opinion> -chuck
sami@strawdog.Eng.Sun.COM (Sami Shaio) (07/03/90)
In article <1990Jul3.004745.7400@wrl.dec.com> price@decisv.enet.dec.com (Chuck Price) writes: > ><Biased opinion> > >The claim that OPEN LOOK is better because you can get XView free clinches >the argument for me. You get what you pay for. I'd suggest that Motif is > (stuff deleted) ><Bias off> >-chuck Isn't it somewhat ironic to be using that kind of argument in this newsgroup? After all, the MIT implementation of X is also free and that doesn't seem to be stopping many of you. Perhaps arguments based on technical data might seem more convincing than "My toolkit costs more than yours". --sami
barnett@grymoire.crd.ge.com (Bruce Barnett) (07/03/90)
In article <9007021559.AA00522@lance> jimf@SABER.COM writes:
HP/Apollo and DEC form quite a large chunk, I daresay larger than Sun.
Nor would I count IBM out (regardless of my personal opinions towards
the company).
From my standpoint, there is one major player -- Sun -- pushing Open
Look. Every other major player is pushing Motif.
McNealy made an interesting point at a recent talk.
First of all, the HP/Apollo and DEC "chunk" includes non-workstation
sales, which you must factor out when comparing market share.
Second, each has several binary formats that must be supported.
HP/Apollo have at least four different machine architectures/formats.
DEC has two. What percentage of the marketplace is any ONE of those?
(A side question - Will DecWindows and NewWave be phased out when
they ship Motif?)
Scott McNealy counters that you can reach 33% of the market share with
one binary for a Sparc. He also claims there will be 1 million
Sparc or Sparc-compatibles sold in one year's time.
Software vendors "go where the money is".
I don't think anyone can say that the war is won until we see third
party vendors making millions of $$ shipping X windows-based software.
--
Bruce G. Barnett barnett@crd.ge.com uunet!crdgw1!barnettittai@shemesh.GBA.NYU.EDU (Ittai Hershman) (07/03/90)
> Scott McNealy counters that you can reach 33% of the market share with > one binary for a Sparc. He also claims there will be 1 million > Sparc or Sparc-compatibles sold in one year's time. > > Software vendors "go where the money is". Scott McNealy makes a good point, but not when it comes to windowing systems. I happen to be typing this on a Sparcstation running Motif; many others are doing the same thing. There are more important things to do than quibble about which GUI is better/nicer/cleaner. Motif is perfectly adequate and its PM look and feel is a big advantage in a world dominated by DOS. If Sun wants to continue to expend resources on pursuing this GUI war idiocy, that is their problem. I, for one, will continue to buy Suns based on all the other factors which I think make them the best workstation for my money, and obtain Motif from a third party (just like add-on memory simms :-). Software vendors will "go where the money is" -- and it is becoming evident that Motif not OpenLook is that place. Even on Sun hardware. -Ittai
de5@STC06.CTD.ORNL.GOV (SILL D E) (07/03/90)
In article <1990Jul3.004745.7400@wrl.dec.com> somebody writes: > >You get what you pay for. Bull. GNU Emacs is far superior to editors costing much more. -- -- Dave Sill (de5@ornl.gov) These are my opinions. Martin Marietta Energy Systems Workstation Support
de5@STC06.CTD.ORNL.GOV (SILL D E) (07/03/90)
In article <BARNETT.90Jul3063026@grymoire.crd.ge.com> barnett@crdgw1.ge.com writes: > >Scott McNealy counters that you can reach 33% of the market share with >one binary for a Sparc. 33% of all workstations are SPARC's running OpenLook? No way. Even if 33% of all workstations *shipping* were SPARC's--which I doubt--I suspect more than a handful are running the MIT distribution or some other non-OpenLook X. -- Dave Sill (de5@ornl.gov) These are my opinions. Martin Marietta Energy Systems Workstation Support
ben@hpcvlx.cv.hp.com (Benjamin Ellsworth) (07/03/90)
> No, you conclude that you better get a computer that runs COBOL if you > want access to that software. That's fair. There are a number of other valid conclusions. My point was that the primary thesis --OL is the future-- is not served by an argument regarding the size of installed base. --- Ben #continue disclaimer
rhoward@msd.gatech.edu (Robert L. Howard) (07/03/90)
In article <3569@shemesh.GBA.NYU.EDU> ittai@shemesh.GBA.NYU.EDU (Ittai Hershman) writes: >There are more important things to do than quibble about which GUI is >better/nicer/cleaner. Motif is perfectly adequate and its PM look and ^^^^^^^ >feel is a big advantage in a world dominated by DOS. ^^^^ Yes, the computing world is somewhat dominated by DOS but PM is *OS/2* which *nobody* really uses. (The only people I know that have even tried were merely curious types.) I know that MS Windows 3.0 is similar and that it runs under DOS, but I don't really believe many people are going to end up using that as a solution. I think the garbage about using Motif *just* because it looks like PM is a bit simplistic. (I do believe there may be *other* good reasons to use (pick) Motif, but this ain't one of them.) Robert -- | Robert L. Howard | Georgia Tech Research Institute | | rhoward@msd.gatech.edu | STL / MSD | | (404) 528-7165 | Atlanta, Georgia 30332 | | UUCP: ...!{allegra,amd,hplabs,ut-ngp}!gatech!msd!rhoward |
barnett@grymoire.crd.ge.com (Bruce Barnett) (07/04/90)
In article <9007031518.AA03015@stc06.CTD.ORNL.GOV> de5@STC06.CTD.ORNL.GOV (SILL D E) writes: > >Scott McNealy counters that you can reach 33% of the market share with >one binary for a Sparc. 33% of all workstations are SPARC's running OpenLook? No way. Even if 33% of all workstations *shipping* were SPARC's--which I doubt--I suspect more than a handful are running the MIT distribution or some other non-OpenLook X. Those are his numbers. My blue sky guess is 80-90% of the sparcs are running SunView. I suppose his logic is that those people will go to OpenWindows when the products they need are available. There is some sense to that, because the OpenWindows environment is very similar to SunView, and most people can convert over to it with little effort. Old Menus's and defaults are converted, the SunView programs are replaced by OpenWindows tools, etc. When someone raised the point that just because Sun has 33% of the market, not all Sun's being shipped were Sparcs. McNealy stood his ground and repeated his claim. I guess this is an extrapolation based on current and expected shipments of Sparcs and Sparc clones. -- Bruce G. Barnett barnett@crd.ge.com uunet!crdgw1!barnett
price@wsl.dec.com (Chuck Price) (07/04/90)
In article <138275@sun.Eng.Sun.COM>, sami@strawdog.Eng.Sun.COM (Sami Shaio) writes: |> In article <1990Jul3.004745.7400@wrl.dec.com> price@decisv.enet.dec.com (Chuck Price) writes: |> > |> ><Biased opinion> |> > |> >The claim that OPEN LOOK is better because you can get XView free clinches |> >the argument for me. You get what you pay for. I'd suggest that Motif is |> > (stuff deleted) |> ><Bias off> |> >-chuck |> |> Isn't it somewhat ironic to be using that kind of argument in this |> newsgroup? |> |> After all, the MIT implementation of X is also free and that doesn't seem |> to be stopping many of you. Perhaps arguments based on technical data |> might seem more convincing than "My toolkit costs more than yours". |> |> --sami It doesn't stop people from using it - but the refusal of hardware vendors (particularly Sun) to spend the extra resources to make it product quality has hurt its ability to penetrate the windowing market. Look at all the press stating that X is slow - Digital's implementation of X, and our performance contributions to X11R4, refute this statement. But those vendors who merely ship the MIT sample server as their X server, or otherwise spin their wheels in pointless counterproductive "competition", only reinforce that opinion. In the case of XView, you will most likely get what you pay for: poor performance relative to production quality toolkits, and an API and implementation with minimal remedial support. I knew the statement that "you get what you pay for" would generate controversy. I agree that there are good and many exceptions to the rule. But in many cases, it is a fact that you can't get something for nothing. A software manufacturer would be making a mistake to bet a business on free software, without any expectation of "paying". In the case of most free software, that payment is in the form of in-house support of the code, including the possibility of having to perform up front porting and bug fixing (and future porting). When you get supported software (like Motif) from a vendor like Digital or H-P, they are providing that service for you. You should expect and demand high quality, high performance implementations, and an adherance to standards. It doesn't necessarily mean that you get it, but because you paid money (by buying the workstation) you at least have some legal recourse. With free software, you are frequently on your own. This above should not be interpreted as an argument for making a selection purely on cost. I merely caution the software manufacturer to realistically weigh the risks and benefits associated with all the packages before making a decision. Those who heard me speak at Unix Expo know that while I support my company's decision to offer Motif, I personally don't take as much stock in the consistency-across-applications argument as do some. As I presented in my talk, I believe it is much more important for the software vendor to provide application consistency across platforms. I do not buy into the position that AcmeCAD should look like an OPEN LOOK application on Sun workstations and a Motif application on DECstations. I submit that AcmeCAD should look like AcmeCAD, so that when the user moves from seat to seat (or when management decides to throw our a hardware vendor), retraining on the same product is not required. (Because of my corporate bias, I would be tickled pink if the vendor decided that the application consistency would be provided by Motif. But if they pick XView, I'll do everything in my power to make sure that they ship a quality product on the DECstation and VAXstation, regardless). I am for toolkits which allow this kind of consistency. Since XView sources are available on the MIT tape, I have no problem with software vendors using this, as long as they understand the inherent limitations imposed (like push-pins not working properly with anything but olwm, etc.). I think it is up to us as X Consortium members to insure that underpinnings are in place to insure the smooth integration of a variety of different toolkits onto one desktop. Personally, I like the variety inherent in a world offering multiple, interoperable toolkits. <Another Bias Alert> I find it ironic that the vendor who just a year ago was bashing the entire industry for not being "open", is now adopting a position that says "we can do whatever we want because we're big". Could it be because they have lost the high ground in hardware and software technology, and are struggling to define product differentiation in commodity market? <End alert> My humble apologies to those programmers who have spent countless hours developing free software, and who might have taken offense at my "get what you pay for" comment. Also my thanks - here I am using Xrn, and I didn't pay a cent for it [of course it breaks about once every 45 minutes and I have to restart it, but it's still much better than rn! :-) ]. -chuck (Please round up the usual disclaimers and insert them here...)
nazgul@alphalpha.com (Kee Hinckley) (07/04/90)
In article <1990Jul1.061632.20280@media.uucp> rmf@media.uucp (Roger Fujii) writes: >willis@ecoult.ncsu.edu (Bill Willis) writes: >There are several vendors who DO NOT bundle Motif with their systems. >This poses several non-trivial problems. Which exist regardless of the toolkit you use. >Let's say a person writes an application with uil. He must choose one of these >headaches: > 1) the application will be HUGE (no shared libraries) One reason why no one in their right mind would use UIL :-). But it has nothing to do with the advantages or disadvantages of the Motif toolkit. > 2) distribute the shared libraries with the application and pay > OSF their distribution fee. > 3) Compatibility risks because the shared library that the user > provides (either vendor supplied or self generated) doesn't > work like they should. Note that if you wanted to use shared Xview libraries you would have exactly the same problem. Moreso - less vendors bundle xview. What has this got to do with Motif? >would be condemned to such a fate. Personally, I think it would be >great if vendors provided multiple interfaces and the the user pick and >choose, but OSF licensing would make this too cumbersome. Funny that Solbourne didn't think so. Sure I agree, it would be great. Too bad there isn't a toolkit in the world which is capable of *fully* supporting multiple GUIs. Why don't we start writing code now with what we've got instead of arguing about what we ought to have. I have no desire to be writing applications for OS/2 five years from now. -- Alphalpha Software, Inc. | motif-request@alphalpha.com nazgul@alphalpha.com |----------------------------------- 617/646-7703 (voice/fax) | Proline BBS: 617/641-3722 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.
smith@scotty (Steven Smith) (07/04/90)
One other note on this subject. Has anyone considered the volume of software that is already written in SunView that will convert easily to XView? I am sure that Sun has thought about this. The process is quite trivial for most of the code, and any vendor attempting such a conversion instantly widens his own market. Personally, I see nothing wrong with multiple GUI's, as end users can adjust more easily than we give them credit for. I have a new package for DNA analysis written in XView today. Thats because eight months ago XView was available as source. The toolkit supports many great features transparenly (infinite split screen, tear off/pinned menus, etc.), and it looks good. People who argue for one toolkit or another probably haven't tried using both. Myself included. But if the toolkit dows what you need, and you can get the support that you need as well, then Look and Feel really aren't that big of an issue. Personal opinions..... Steve Smith Research Programmer U of Illinios
stpeters@dawn.crd.ge.COM (Dick St.Peters) (07/04/90)
Chuck Price (price@decwrl.dec.com) writes: > Sun *might* have 25% of the Unix W/S marketplace today. The last numbers I saw said Sun had 36+% of the market for the first quarter of this calendar year, up substantially from their ~28% for all of last year. > The claim that OPEN LOOK is better because you can get XView free clinches > the argument for me. You get what you pay for. Then NeWS is better than X, isn't it? > Software vendors *want* a common API. I suspect they want an API that's common across all the platforms they write for but don't care whether it's common with other software vendors. -- Dick St.Peters, GE Corporate R&D, Schenectady, NY stpeters@dawn.crd.ge.com uunet!dawn.crd.ge.com!stpeters
reha@cunixf.cc.columbia.edu (Reha Elci) (07/04/90)
In article <1990Jul3.193445.17201@wrl.dec.com> price@decisv.enet.dec.com (Chuck Price) writes: >only reinforce that opinion. In the case of XView, you will most likely get >what you pay for: poor performance relative to production quality toolkits, >and an API and implementation with minimal remedial support. > There was a post indicating that sofware vendors go where the money is. This is absolutely true! You have to look at this issue from a company's point of view. The task is to select a platform+UI that will serve its users' needs. Moreover, the combination must be supportable. This implies that the vendor of the software and the hardware must have excellent service and support. Free software contains, almost by definition, no support. A company who decides to use such software is on its own with respect to fixing of bugs, presence of viruses and a support base to solve problems and even just answer a few questions. People who advocate free software have NO idea how critical quality is! Real money and reputation is on the line every time you commit to a piece of software that is a critical part in an operation. A $50 paid to a vendor or even a $1000 paid to a vendor who provides support is usually dwarfed by the amount you stand to lose if something goes wrong. You think NYSE should use free software for booking trades? Most software vendors are not interested in hackers who save money for years to buy a workstation and because that is NOT where the money is. So one should discount the opinions of that group (which is also the group who is for free software). >I knew the statement that "you get what you pay for" would generate >controversy. I agree that there are good and many exceptions to the rule. >With free software, you are frequently on your own. > No! There are NO exceptions to the rule! GNU emacs, from a company's perspective is NOT a viable answer; as good as it is for hackers. When I get a tape from FSF, it comes with NO guarantees. Well, that is the whole point; a company wants packages which come with guarantees. When free software comes with 7x24 software support and guarantees, then I will alter my position on this issue. XView, low quality and very little service is not a good alternative, neither is MIT distribution of anything! X is available, with full support from OSF members. DEC has their own server, display postscript (from original vendor who supports it -- not like SUN's clone), and own improved implementation of Motif and DWT which is fully supported. IBM announcements are in the same direction with the addition of NeXTStep with full support and service. Even Motif from OSF with their inadequate support might not cut it. To recap, the money is with companies who are willing to pay for quality, service and support. Big companies spend literally millions of dollars for software and hardware; what makes people think they care about +- $100 for good software? Such companies choosing Motif as the platform will force the software vendors in that direction; it they do not comply they will be out of business for neglecting that market that actually does not mind paying extra for added service. >company's decision to offer Motif, I personally don't take as much stock >in the consistency-across-applications argument as do some. As I presented >in my talk, I believe it is much more important for the software vendor >to provide application consistency across platforms. I do not buy into > >-chuck That is definitely the wrong attitude. Look and feel across applications is EXTREMELY important. To understand this take a look at the learning times for software packages on the Mac. If all had their own implementation of widgets and modes of operation, do you think Macs would have been a success. Just about every literature on user interface design stresses the importance of this point. A text widget MUST work exactly the same way across all programs under the same window manager! I definitely should be able to cut and paste text across applications. All user interface objects like push buttons, file selection boxes, help widgets MUST look and feel the same. If I have to learn the l&f of a new widget set every time I buy software and I have to go and explain to the users this new methodology, productivity would suffer to a very large extent. When you do not have consistency accross applications, you have just violated countless golden rules of user interface design. That is simply BAD user interface design. Application consistency across platforms is moot; most large companies which will buy this software from vendors also happen to have single vendor policies. Surely, even they had different hardware all over the place, they would at least pick the same UI. If marketing data proves otherwise, I would be very interested to see it. > >(Please round up the usual disclaimers and insert them here...) same here.
gsh7w@astsun9.astro.Virginia.EDU (Greg S. Hennessy) (07/05/90)
Reha Elci writes:
#No! There are NO exceptions to the rule! GNU emacs, from a company's
#perspective is NOT a viable answer; as good as it is for hackers. When I
#get a tape from FSF, it comes with NO guarantees.
Lotus-123 comes with no guarentees. You should then conclude that
Lotus is NOT a viable answer for a company's perspective.
--
-Greg Hennessy, University of Virginia
USPS Mail: Astronomy Department, Charlottesville, VA 22903-2475 USA
Internet: gsh7w@virginia.edu
UUCP: ...!uunet!virginia!gsh7wsami@ruffles.Eng.Sun.COM (Sami Shaio) (07/06/90)
In article <1990Jul4.145214.6357@cunixf.cc.columbia.edu> reha@cunixf.cc.columbia.edu (Reha Elci) writes: >In article <1990Jul3.193445.17201@wrl.dec.com> price@decisv.enet.dec.com (Chuck Price) writes: > >>I knew the statement that "you get what you pay for" would generate >>controversy. I agree that there are good and many exceptions to the rule. >>With free software, you are frequently on your own. >> > >No! There are NO exceptions to the rule! GNU emacs, from a company's >perspective is NOT a viable answer; as good as it is for hackers. When I >get a tape from FSF, it comes with NO guarantees. Well, that is the whole >point; a company wants packages which come with guarantees. When free >software comes with 7x24 software support and guarantees, then I will >alter my position on this issue. > >XView, low quality and very little service is not a good alternative, neither >is MIT distribution of anything! X is available, with full support from OSF >members. DEC has their own server, display postscript (from original vendor You shouldn't confuse how a product is distributed with how it will be supported. Sun has chosen to distribute XView as a free product with source. However this doesn't mean it's unsupported! In fact, there were press reports about Sun getting third-party companies to support XView on other platforms. And Sun supports XView on the sun platform. It just happens to be free. --sami ps: Please note that I speak for myself only. (standard disclaimer)
chrise@bcsaic.UUCP (Chris Esposito) (07/06/90)
In article <1210034@hpfcmgw.HP.COM> chan@hpfcmgw.HP.COM (Chan Benson) writes: >Indeed, I believe the people who wrote Looking Glass (a desktop manager >for X11) used their own Motifish library. (If I'm wrong please correct >me; that was the impression I got from a trade rag article and we all "Motifish" - what a delightful word. I've always suspected there was something fishy about OSF & Motif. I'll ask my grocer if I can get a pound of this new sort of fish the next time I'm in there. "Open Look fish" doesn't have nearly as nice a ring to it. But seriously, folks... The folks from Visix (Looking Glass developers) were just here for a demo. What they told me was that they have an Xlib-based (yes, all the way down at that level) toolkit that they used to get the Motif look & feel. They claimed large performance improvements over implementations using the standard Motif libraries. The demo was on a DEC 3100 and performance was quite impressive. Looking Glass does not distinguish between mouse buttons (Looking Glass user's Mnaual, page 11); from what I remember of the Motif Style Guide, each mouse button has a different name/function assigned to it. While such trivial "violations" of the Style Guide mean very little to me, they may be much more important to others. -- "A waist is a terrible thing to mind" - Dan Quayle at a Weight Watchers group ------------------------------------------------------------------------------ Chris Esposito | Internet: chrise@atc.boeing.com Boeing Advanced Technology Center | uucp: ...!uw-june!bcsaic!chrise
iarrobin@wookie.Central.Sun.COM (Mike Iarrobino - Systems Engineer) (07/06/90)
In article <9007021559.AA00522@lance> jimf@SABER.COM writes: >|How much of the workstation market is DG anyway ? Or IBM for that matter ? > >HP/Apollo and DEC form quite a large chunk, I daresay larger than Sun. >Nor would I count IBM out (regardless of my personal opinions towards >the company). Hmmm, let's look at some real numbers. ------------------------------------------------------------------------ The following is from Computer Reseller News, June 25th, 1990, p. 14: "WORLDWIDE RISC WORKSTATION/SERVER MARKET BY MICROPROCESSOR, VENDOR" 1989-1990 Unit Shares 1989 1990 WORKSTATIONS SERVERS WORKSTATIONS SERVERS TOTAL SPARC 61% 45% 74% 50% Sun 61 41 61 38 Solbourne 0 4 1 5 Toshiba 0 0 9 0 other 0 0 4 7 TOTAL MIPS 22 21 16 24 Silicon Graphics 11 3 4 5 Digital 10 10 10 10 MIPS 1 8 1 5 other 0 0 1 4 HP-PA 6 13 3 10 IBM RT/RS-6000 5 9 2 7 other 6 10 4 10 ------------------------------------------------------------------------ These figures clearly show Sun Microsystems as the leading vendor of RISC workstations. -- /\ \\ \ Michael V. Iarrobino \ \\ / Systems Engineer / \/ / / Sun Microsystems / / \//\ One Cranberry Hill \//\ / / Lexington, MA 02173 / / /\ / / \\ \ Phone: (617) 861-2241 \ \\ EMail: iarrobin@East.Sun.COM \/
willis@ecoult.ncsu.edu (Bill Willis) (07/06/90)
In article <2142@east.East.Sun.COM>, iarrobin@wookie.Central.Sun.COM (Mike Iarrobino - Systems Engineer) writes: > > These figures clearly show Sun Microsystems as the leading vendor of > RISC workstations. > I thought we had been talking about ALL workstations, not just RISC workstations. There are a significant number of non-RISC workstations in this market place and they change the numbers quoted in your post quite a bit. Bill Willis, Director willis@ecovsa.ncsu.edu Engineering Computer Operations willis%ecovsa@ncsuvx.ncsu.edu Box 7901, School of Engineering North Carolina State University (919) 737-2458 We have met the enemy, and they are us --- Pogo
mayer@hplabsz.HPL.HP.COM (Niels Mayer) (07/06/90)
In article <2142@east.East.Sun.COM> iarrobin@wookie.Central.Sun.COM (Mike Iarrobino - Systems Engineer) writes: >In article <9007021559.AA00522@lance> jimf@SABER.COM writes: >>|How much of the workstation market is DG anyway ? Or IBM for that matter ? >> >>HP/Apollo and DEC form quite a large chunk, I daresay larger than Sun. >>Nor would I count IBM out (regardless of my personal opinions towards >>the company). > >Hmmm, let's look at some real numbers. > >The following is from Computer Reseller News, June 25th, 1990, p. 14: >"WORLDWIDE RISC WORKSTATION/SERVER MARKET BY MICROPROCESSOR, VENDOR" ^^^^ > These figures clearly show Sun Microsystems as the leading vendor of > RISC workstations. Given the background of the discussion (workstation market share), you've just posted a completely useless and misleading comparison. We're not talking about market share of RISC workstations, we're talking about market share of workstations, and in particular, the total number of workstations running Motif versus Open Kook. It doesn't matter diddly squat whether the workstation is using a SPARC, HP's PA-RISC, MIPS, 68xxx, 88xxx, i386, i486, etc. processor. Nice try... ---------- PS: I had a real good time at the Xhibition tradeshow watching Sun marketing droids burst their cranial bloodvessels everytime they saw an ISV's sparcstation running a Motif application. "You gots to chill" -- EPMD. ------------------------------------------------------------------------------ Disclaimer: I'm speaking for myself only. ------------------------------------------------------------------------------- Niels Mayer -- hplabs!mayer -- mayer@hplabs.hp.com Human-Computer Interaction Department Hewlett-Packard Laboratories Palo Alto, CA. *
nazgul@alphalpha.com (Kee Hinckley) (07/06/90)
In article <BARNETT.90Jul3063026@grymoire.crd.ge.com> barnett@crdgw1.ge.com writes: >McNealy made an interesting point at a recent talk. ... >Second, each has several binary formats that must be supported. I'm sorry. I think this is a pretty funny argument coming from a company that not only has three different machine architectures, but also has three different *data* representations. Even DEC and Apollo managed to get that one right. But in any case, the marketshare numbers I've seen were numbers of workstations period. That should be fairly obvious, if they weren't, IBM would be listed as number one. -kee -- Alphalpha Software, Inc. | motif-request@alphalpha.com nazgul@alphalpha.com |----------------------------------- 617/646-7703 (voice/fax) | Proline BBS: 617/641-3722 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) (07/06/90)
In article <138275@sun.Eng.Sun.COM> sami@strawdog.Eng.Sun.COM (Sami Shaio) writes: >In article <1990Jul3.004745.7400@wrl.dec.com> price@decisv.enet.dec.com (Chuck Price) writes: ... >>The claim that OPEN LOOK is better because you can get XView free clinches >>the argument for me. You get what you pay for. I'd suggest that Motif is ... >Isn't it somewhat ironic to be using that kind of argument in this >newsgroup? > >After all, the MIT implementation of X is also free and that doesn't seem >to be stopping many of you. Perhaps arguments based on technical data >might seem more convincing than "My toolkit costs more than yours". Support is the difference. If I were stuck with the reference implementation of X on my workstation I wouldn't be using X. But I'm using an optimized server from my vendor. That's what Motif is going to have the XView won't. As for why you see a whole bunch of engineers suddenly making marketing arguments (and looking pretty silly doing it), there are a couple answers. For one thing, the toolkits out there are pretty primitive, buggy and weak. There just isn't a lot to be said technically for any of them. However, given time and support they'll get better, and then we can start arguing about technical merits. Right now it's just a matter of getting enough acceptance to make it worthwhile to enhance the toolkit. The other reason is more basic, it has become very clear (or should I say, Sun has (through their success) made it very clear) that technical excellence, while nice, is not essential to the survival of a particular product. Technically, Apollo had a much better system - builtin distributed file system, demand paging across the network, transparent support for diskless nodes (plug it in and boot it), graphical windowing system, system-wide editor available in every window, objected-oriented typed file system - and most of that was available close to 10 years ago. Unfortunately, having all that neat stuff is not enough, you've got to be able to sell it. Motif has suceeded in selling itself, and (here's the good news) it's not any worse than any of the other toolkits out there. If you want my opinion (can't imagine why you would), the biggest mistake that the OL crowd make was going with multiple APIs. Sure, you can argue that it's too early to standardize on an API ("what about voice and video," they say). But in fact if you don't standardize on an API then you have just fragmented the Unix market even more, and the potential market share of my product is now smaller. What Sun forgets is that it is possible to extend an API to future systems without becoming backwards incompatible. Sure, it's not always the best thing, and if you were to rewrite it you might do it differently, but it can be done (and as things move along, you can faze in new APIs). But arguing that it's too early to standardize on one, so instead we should standardize on many is just silly. That's just another way of saying "Sun wants a SunView API, and AT&T wants their own, so we'll do two". This reminds me somewhat of the original X announcement, which Sun declined to come to despite multiple invitations (well, actually they tried to crash it at the last minute). The reason Bill Joy gave, in an article he wrote for one of the Unix mags, was that he refused to stand up there and state that X was *the* standard windowing system. I believe he felt it was too early to standardize on one windowing system. (Of course what was missing from that article was the fact that the announcement in fact did *not* state that X was _the_ standard windowing system, only that it was _a_ standard windowing system; but then, one shouldn't let the facts get in the way.) -kee -- Alphalpha Software, Inc. | motif-request@alphalpha.com nazgul@alphalpha.com |----------------------------------- 617/646-7703 (voice/fax) | Proline BBS: 617/641-3722 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.
jimf@SABER.COM (07/06/90)
|People who advocate free software have NO idea how critical |quality is! This is a sweeping generalization which is not true since a good many people who supply free software also supply commercial software. |GNU emacs, from a company's |perspective is NOT a viable answer; as good as it is for hackers. When I |get a tape from FSF, it comes with NO guarantees. Well, that is the whole |point; a company wants packages which come with guarantees. Ok, we have several commercial software packages here. I found the license agreements to two of them: Sun Microsystems TOPS: -- cut here -- TOPS warrants to the original licensee only that the Software shall operate substantially in accordance with the description Provided in the accompanying Documentation. [...] IN NO EVENT SHALL TOPS NOR ANYONE ELSE WHO HAS BEEN INVOLVED IN THE CREATION, PRODUCTION OR DELIVERY OF THE SOFTWARE, HARDWARE OR DOCUMENTATION BE LIABLE FOR ANY SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES WHETHER BASED UPON BREACH OF WARRANTY, BREACH OF CONTRACT, NEGLIGENCE, STRICT LIABILITY IN TORT OR ANY OTHER LEGAL THEORY. Such damages include but are not limited to, loss of profits, loss of savings or revenue, loss of use of the Software or Hardware and any associated equipment or software, cost of capital, cost of any substitute equipment, facilities or services, downtime, the claims of third parties including customers and injury to property. -- cut here -- In English, that means "we promise that it operates pretty much as our documentation says it does, but if it breaks and you loose anything, we're not going to pay for it." Adobe Type Manager: -- cut here -- [...] The Software will be free from errors under normal use and service for a period of ninety (90) days after delivery to you. For purposes of this Agreement an "error" shall be a flaw in the Software which causes it not to reproduce the Adobe version of the typefaces when used on a computer utilizing the Adobe Type Manager software, the Display PostScript system from Adobe or an Adobe PostScript interpreter. You must report all defects and return the Software to the location where you obtained it with a copy of your receipt within such period to be elibible for warranty service. If the Software fails to comply with this warranty, Adobe will, at its option and cost, replace the diskettes not meeting the warranty or refund the license fee for the Software. [...] IN NO EVENT WILL ADOBE BE LIABLE TO YOU FOR ANY CONSEQUENTIAL OR INCIDENTAL DAMAGES, INCLUDING ANY LOST PROFITS OR LOST SAVINGS, OR FOR ANY CLAIM BY ANY PARTY, EVEN IF AN ADOBE REPRESENTATIVE HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. -- cut here -- In English this says "We promise that this generates fonts. If it doesn't, we'll either give you your money back or replace the disks, at our option. If it breaks and you loose anything, we're not going to pay for it." These statements are very standard amongst commercial software. GNU has no warranty. They don't promise that their stuff does anything in particular and if you loose something, they're not going to pay for it. I don't see a whole lot of difference in their warranties except that with GNU they won't repay your purchase price (which you didn't pay anyway). Your only guarantee you have is that a for-profit company probably won't stay in business long if their product doesn't work (although there are obvious exceptions). Given the success to failure rate in our business, this isn't saying much. Welcome to the Real World. jim frost saber software jimf@saber.com
geoff@nluug.nl (G. Coupe EPD/74 O75/1435) (07/07/90)
In article <11025@hydra.gatech.EDU> rhoward@msd.gatech.edu (Robert L. Howard) writes: >I think the garbage about using Motif *just* because it looks like >PM is a bit simplistic. (I do believe there may be *other* good reasons >to use (pick) Motif, but this ain't one of them.) Beg to disagree. We have a mixed community of PC and technical workstation users. There is a big benefit to putting Motif on the workstations and Windows 3.0 on the PCs *because* of the same look and feel - the users can feel at home on either platform, and they don't curse in frustration (and lose productivity through mistakes). Lower retraining costs, lower end-user support costs. Many users switch tools during the course of their day's work: running technical applications, and then writing up results on a PC. We expect some users to access Motif clients via an X server running under Windows, and the similarity of the look and feel will, we hope, create the illusion of seamless integration (at least as far as the user is concerned). - Geoff Coupe These are my views, not necessarily my employers
peter_colby@viewlogic.COM (Peter Colby) (07/07/90)
Hey, did this get to you?? If it did, please send me a better
route!! If it didn't I'll keep trying!
Peter Colby
peter@viewlogic.comnick@bischeops.UUCP (Nick Bender) (07/07/90)
In article <1990Jul4.145214.6357@cunixf.cc.columbia.edu>, reha@cunixf.cc.columbia.edu (Reha Elci) writes: > > No! There are NO exceptions to the rule! GNU emacs, from a company's > perspective is NOT a viable answer; as good as it is for hackers. When I > get a tape from FSF, it comes with NO guarantees. Well, that is the whole > point; a company wants packages which come with guarantees. When free > software comes with 7x24 software support and guarantees, then I will > alter my position on this issue. > I hate to burst your bubble, but support is now available for most of the GNU stuff. Here is a blurb from info I received from them: Cygnus support has been supporting free software for major companies and laboratories since January, 1990. Our initial product, _Leveraged Support_, provides large sites with significant cost savings while giving them access to our world-class free software developers. We have recently expanded our support offerings with Core Support, which provides similiar benefits on a smaller scale. The address: Cygnus Support 814 University Ave. Palo Alto, CA 94301 +1 415 322 3811 FAX 415 326 1669 As for the need for quality, GNU emacs is far and away the best editor I have ever used. And gcc is regularly sited for producing smaller and faster optimized object code. And flex and bison are recognized to be more powerfull than lex and yacc. And look at the portability of this stuff - emacs runs on damn near anything resembling U**x. Freeware and other hacks are one thing, but in general I have found FSF code to be of high quality. I would have no problems recomending it to any commercial company, especially with the support available from Cygnus. Nick Bender nick@bis.com uunet!bischeops!nick
mouse@SHAMASH.MCRCIM.MCGILL.EDU (der Mouse) (07/07/90)
> People who advocate free software have NO idea how critical quality > is! ...to whom? (Yes, I know you implicitly said, but you're missing the point.) Quality is *not* critical - to me. Or to lots of other people. And on another point, *what* quality? We are running an operating system here which is allegedly vendor-supported and presumably has the "quality" you're talking about. I've found at least as many bugs in it as I have in X, and with X, I can *fix* the bugs. On the spot. If you add in the cost of a service contract guaranteeing the sort of response time possible to someone with all the source right there, it's out of reach even for many of the "large" buyers. (I once had occasion to do a little consulting for a "large buyer", whose name I won't give here, and we had access to the vendor support people, so I have some idea what I'm talking about when I say that.) > No! There are NO exceptions to the rule! GNU emacs, from a > company's perspective is NOT a viable answer; as good as it is for > hackers. When I get a tape from FSF, it comes with NO guarantees. > Well, that is the whole point; a company wants packages which come > with guarantees. When free software comes with 7x24 software support > and guarantees, then I will alter my position on this issue. Let me quote a bit from the license agreement for Turbo C. L I M I T E D W A R R A N T Y With respect to the physical diskette and physical documentation enclosed herein, Borland International Inc. ("Borland") warrants the same to be free of defects in materials and workmanship for a period of 60 days from the date of purchase. [stuff about how to claim against the previous sentence's promises.] Borland International, Inc. specifically disclaims all other warranties, expressed or implied, including but not limited to implied warranties of merchantability and fitness for a particular purpose [...]. If someone can produce evidence that this is atypical, you may have a point. But otherwise, I think your argument founders on the facts: nobody guarantees software, not even the commercial stuff. As for support, I've had a couple of occasions to contact vendor support at various companies. None of them compares to what I see for X simply by posting to xpert/comp.windows.x. >> I personally don't take as much stock in the >> consistency-across-applications argument as do some. > That is definitely the wrong attitude. Look and feel across > applications is EXTREMELY important. Again - to whom? To you? To your company? Certainly not to me. der Mouse old: mcgill-vision!mouse new: mouse@larry.mcrcim.mcgill.edu
oz@yunexus.yorku.ca (Ozan Yigit) (07/07/90)
In article <2142@east.East.Sun.COM> iarrobin@wookie.Central.Sun.COM (Mike Iarrobino - Systems Engineer) writes: >... figures omitted ... > >These figures clearly show Sun Microsystems as the leading vendor of >RISC workstations. So, Dennis Ritchie is right again. oz
vladimir@prosper (Vladimir G. Ivanovic) (07/07/90)
In article <9007062359.AA01925@shamash.McRCIM.McGill.EDU>, mouse@SHAMASH (der Mouse) writes: [...] >nobody guarantees software, not even the commercial stuff. As for [...] Wrong. CE Software, Central Point Software, ALSoft, Preferred Publishers and Symantec, just to mention a few Macintosh vendors, as well as DEC (last time I checked - 2 years ago) guarantee that their software does what the docs say it does, and if it can't fixed it in a reasonable amount of time, your money will be refunded. There are time limits (60 days - 1 year). Dantz Development will refund your money if you're dissatisfied within 60 days. Some vendors DO stand behind thier products. -- Vladimir
barnett@grymoire.crd.ge.com (Bruce Barnett) (07/07/90)
In article <1990Jul6.135852.22415@alphalpha.com> nazgul@alphalpha.com (Kee Hinckley) writes: >Second, each has several binary formats that must be supported. I'm sorry. I think this is a pretty funny argument coming from a company that not only has three different machine architectures, but also has three different *data* representations. Even DEC and Apollo managed to get that one right. It's true that the data representations are not compatible between the Sun lines. My personal view is that this forces programmers to write programs the correct way (i.e. portable), instead of a convenient, non-portable way. Done the right way, a data file can be shared across Dec, Apollo, HP, MIPS, IBM, and Sun. -- Bruce G. Barnett barnett@crd.ge.com uunet!crdgw1!barnett
rlk@THINK.COM (Robert L Krawitz) (07/07/90)
Date: Fri, 6 Jul 90 19:59:46 -0400 From: der Mouse <mouse@shamash.mcrcim.mcgill.edu> > People who advocate free software have NO idea how critical quality > is! And on another point, *what* quality? We are running an operating system here which is allegedly vendor-supported and presumably has the "quality" you're talking about. I've found at least as many bugs in it as I have in X, and with X, I can *fix* the bugs. On the spot. If you add in the cost of a service contract guaranteeing the sort of response time possible to someone with all the source right there, it's out of reach even for many of the "large" buyers. Note that even a service contract is no guarantee, if the vendor does not acknowledge the bug. We currently have a very serious problem with the kernel from a major vendor (accessing a very large directory over NFS apparently causes occasional system crashes). Said vendor has not acknowledged the problem. Yes, I have had disagreements with providers of free software about whether something's a bug or not (not even as clear cut as the example above). The last time we disagreed on the issue, I simply made the change myself. There is at least one company that I'm aware of (Cygnus Support; my only connection is that I know one of the principals) that's in the business of providing paid support for users of free software, to address the support issue very specifically. So to say that "free == unsupported" and "paid == supported" is wrong. ames >>>>>>>>> | Robert Krawitz <rlk@think.com> 245 First St. bloom-beacon > |think!rlk (assistant postmaster) Cambridge, MA 02142 harvard >>>>>> . Thinking Machines Corp. (617)876-1111
mouse@LARRY.MCRCIM.MCGILL.EDU (der Mouse) (07/08/90)
>> nobody guarantees software, not even the commercial stuff. > Wrong. > CE Software, Central Point Software, ALSoft, Preferred Publishers and > Symantec, just to mention a few Macintosh vendors, as well as DEC > (last time I checked - 2 years ago) guarantee that their software > does what the docs say it does, and if it can't fixed it in a > reasonable amount of time, your money will be refunded. If that's a guarantee, well, then all free software is guaranteed. Most of the big free software places (MIT X group, FSF, Larry Wall, etc) actually fix things faster than most commercial vendors (once you can get them to agree that such-and-such needs fixing - a problem which (as someone pointed out here recently) exists with the commercial outfits as well). der Mouse old: mcgill-vision!mouse new: mouse@larry.mcrcim.mcgill.edu
stripes@eng.umd.edu (Joshua Osborne) (07/09/90)
In article <1990Jul3.201647.3436@alphalpha.com> nazgul@alphalpha.com (Kee Hinckley) writes: >In article <1990Jul1.061632.20280@media.uucp> rmf@media.uucp (Roger Fujii) writes: [...] >> 2) distribute the shared libraries with the application and pay >> OSF their distribution fee. >> 3) Compatibility risks because the shared library that the user >> provides (either vendor supplied or self generated) doesn't >> work like they should. >Note that if you wanted to use shared Xview libraries you would have >exactly the same problem. Moreso - less vendors bundle xview. What >has this got to do with Motif? Except with XView (2) would be "distribute the shared libraries with the application", note the missing part "and pay OSF their distribution fee", XView is free, you can include the shared lib with your code if you think you need to. (Without paying any money, or filling out any forms) [...] -- stripes@eng.umd.edu "Security for Unix is like Josh_Osborne@Real_World,The Mutitasking for MS-DOS" "The dyslexic porgramer" - Kevin Lockwood "Don't try to change C into some nice, safe, portable programming language with all sharp edges removed, pick another language." - John Limpert
nazgul@alphalpha.com (Kee Hinckley) (07/11/90)
In article <1990Jul9.032137.23688@eng.umd.edu> stripes@eng.umd.edu (Joshua Osborne) writes: >Except with XView (2) would be "distribute the shared libraries with >the application", note the missing part "and pay OSF their distribution fee", It is not clear to me that this is the case. I believe, although I can not tell for sure from the contract, that I do NOT have to pay a distribution fee (which by the way, is at _most_ $40/machine) if the machine in question already has Motif on it from another source. Or in other words, I'd just have to charge $40 extra for the Sun version of my product :-). -kee -- Alphalpha Software, Inc. | motif-request@alphalpha.com nazgul@alphalpha.com |----------------------------------- 617/646-7703 (voice/fax) | Proline BBS: 617/641-3722 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.
sean@dsl.pitt.edu (Sean McLinden) (07/15/90)
> People who advocate free software have NO idea how critical quality > is! This type of thinking has, quite simply, not been born out by experience. Without a doubt the three most significant advances in software in the past century have been Unix, X, and TCP/IP (the latter two attributable, in part, to Unix). I can make a strong argument for these (and PLEASE, don't mention ANYTHING produced by Microsoft, which is a marketing company not a software company, as being a competitor). Each of those systems was available in source form at little or no cost. Smart vendors[1] realized that the "network" is a tremendous programming resource; one which could not be afforded by almost any commercial concern. Wanna find the bugs in your code? Post it to the net. You'll get not only bug fixes but enhancements and extensions more than you can believe. So how do you make money? You sell a supported version and you charge for the support. Those people who MUST have a reliable system will pay for it; the rest will contibute to it's enhancement (like the old adage in science "There are two kinds of people in science: those that read the literature and those that contribute to it.") Whether or not people like Stallman are correct on philosophical grounds he has proven to be correct on practical grounds. After all, a lot of people are making a living selling TeX support but you can still get it for free. Sean McLinden Decision Systems Laboratory University of Pittsburgh [1] Actually, most of the money to be made is by startups and entrepre- neurs. To date, few if any large vendors have been smart enough to realize this which is why we are getting a VMS which is POSIX compliant rather than a Unix which runs VMS applications.