mikeg@dali.gatech.edu (Mike Gourlay) (12/13/90)
Hi HP people, I hear that HPUX 8.0 will be shipped out in June or July. I also hear that X11R4 will be send with it. That is, if you have `software support.' What if you don't have software support? [Gripes about HP policy left for later.] Will someone please make a tar file of the distribution tape, and put it somewhere? Or send the tar file to those of us who have no support from HP? My local HP rep was confused when I told her I knew of people with X11R4 on an HP. She seems to think that it hasn't been released yet. People have it, though! Will the X11R4 binaries running on HPUX 6.5 also run on a similar machine running 7.0? Specifically, will the X binaries running on a HPUX 6.5 9000s319C+ run on a HPUX 7.0 9000s370? Libraries? Server? Client? Why does the X11R3 that comes with HPUX 7.0 not include some of the X routines that many public programs expect? Will X11R4 lack this hideous problem? Why does HP consider fixing their software bugs (called an update of the OS) something for which you have to pay big money for a support contract? It is unfair. If you pay for UNIX, you should get UNIX, and not have to pay incrementally to have HP give you UNIX over a period of time. It isn't as if the machine was inexpensive up front. It isn't as if HPUX is not the most incompatible port of UNIX and X on the planet. It most certainly is NOT that new versions of the OS are improvements to UNIX. The new versions are bug fixes. That is not right. X is in the public domain. HP makes so called `inhancements' (read "incompatible frilly fluff") to it and sells it for money. Motif is nice, but the other bugs and left-outs of HP's version of X on HPUX are not worth it. X should be compilable from the distribution at MIT, and so capable people should be able to compile X11R4 without waiting for the newest distribution from HP. Has anyone done that? I'll admit that I'm not a capable person when it comes to getting UNIX programs to run on HPUX. My local rep blaimed it on inherent incompatibilities in UNIX. WHAT?!?!?!!? That's ridiculous. Inherent incompatibilities in UNIX? No such thing. Between UNIX and HPUX, okay. Between SYSV and BSD, okay. Between _either_ BSD or SYSV and HPUX, I see incompatibility (or is it incompetence?) But not within UNIX. That, according to her, is why WE HAVE TO PAY HP TO GIVE US PUBLIC DOMAIN SOFTWARE AND BUG FIXES TO THEIR WEIRD, BUGGY, INCOMPLETE UNIX PORT AND SPECIAL VERSION OF X. For now, I run many of the X apps I need on a Sparc, and just have the Sparc use my display. EtherNet is wonderful. xhost + is fantastic. My HPUX is being used as a dumb X terminal!!! I'd like it to evolve into something more useful. Thank you in advance for your reply, Mike Gourlay mike@penguin.gatech.edu -------------------------------------------- I don't officially speak for Georgia Tech or GTRI; Why the hell should they care if I'm unhappy with HPUX, just because I work with it every single day? The upper echelons have Sparcs and SGI's.
milburn@me10.lbl.gov (John Milburn) (12/15/90)
In his article mikeg@dali.gatech.edu (Mike Gourlay) angrily writes: >Hi HP people, Well, I'm not an hp person, but here goes. > I hear that HPUX 8.0 will be shipped out in June or July. >I also hear that X11R4 will be send with it. That is, if you have >`software support.' HPUX rev 8.0 will indeed be a major upgrade to the operating system, and as such, HP deserves to recoup the investment they put into it. The mechanism for this is the software support contract, which I believe > What if you don't have software support? [Gripes about >HP policy left for later.] Will someone please make a tar file of the >distribution tape, and put it somewhere? Or send the tar file to those >of us who have no support from HP? I think HP would (rightfully) scream bloody murder if someone were to freely distribute such an archive. (Even _I_ wouldn't consider posting a full OS release. :-) The only reason you don't have support is because you choose not to. > My local HP rep was confused when I told her I knew of >people with X11R4 on an HP. She seems to think that it hasn't been >released yet. People have it, though! X11R4, as released by MIT, compiles and runs rather well under hpux 7.0. I have been using it since last May. Your rep is certainly uninformed, particularly with respect to PD software under hp-ux. This is a common problem, as most hp sales reps have no knowledge of the vast community of networked machines in the world. Try to get one to exchange e-mail with you! > Will the X11R4 binaries running on HPUX 6.5 also run >on a similar machine running 7.0? Specifically, will the X binaries >running on a HPUX 6.5 9000s319C+ run on a HPUX 7.0 9000s370? >Libraries? Server? Client? Most binaries which run under 6.5 will run under 7.0. I don't recall needing to recompile anything when we switched over ( about a year ago.) Of course, I never ran R4 under 6.5, so I can't answer the specific question. I was running MIT's R3 at the time of the switch, and didn't have any problem with it. > Why does the X11R3 that comes with HPUX 7.0 not include >some of the X routines that many public programs expect? Will X11R4 >lack this hideous problem? A number of things are indeed missing. That is what motivated me to get the release from MIT. I don't know HP's rational for releasing only a subset of the binaries and libraries. A larger set of the R3 stuff was made available on an unsupported SE tape, which you can ftp from me10.lbl.gov. If you really want a full X implementation, get the MIT distribution and compile it yourself. It can be had via anonymous ftp from expo.lcs.mit.edu. > Why does HP consider fixing their software bugs (called an update >of the OS) something for which you have to pay big money for a >support contract? It is unfair. If you pay for UNIX, you should get UNIX, >and not have to pay incrementally to have HP give you UNIX over a period >of time. It isn't as if the machine was inexpensive up front. It isn't as if >HPUX is not the most incompatible port of UNIX and X on the planet. It >most certainly is NOT that new versions of the OS are improvements >to UNIX. The new versions are bug fixes. That is not right. I agree with you to a certain extent. If you buy an hp box, with rev 6.x, you should be provided with bug fixes to that revision of the OS free of charge. However, if you choose not to purchase a support contract, you have no inherent right to future major releases. I find the cost of HP support contracts to be very reasonable, particularly the material only level of support. > X is in the public domain. HP makes so called `inhancements' >(read "incompatible frilly fluff") to it and sells it for money. HP bundles a subset of X with their systems, along with some hp enhancements, like mwm and hpterm. They do not sell any MIT software. They have every right to sell their enhancements. If you don't want what hp give you for free, or want more, get it from MIT. >Motif is nice, but the other bugs and left-outs of HP's version of X >on HPUX are not worth it. X should be compilable from the distribution >at MIT, and so capable people should be able to compile X11R4 without >waiting for the newest distribution from HP. Has anyone done that? I'll Many people have done this. >admit that I'm not a capable person when it comes to getting UNIX programs >to run on HPUX. My local rep blaimed it on inherent incompatibilities in >UNIX. WHAT?!?!?!!? That's ridiculous. Inherent incompatibilities in UNIX? Again, your local rep is uninformed. This is, I believe, a major problem with hp. The sales reps are (severe generalization here) remarkably unfamiliar with the systems they sell. [...] > For now, I run many of the X apps I need on a Sparc, and just >have the Sparc use my display. EtherNet is wonderful. xhost + is >fantastic. My HPUX is being used as a dumb X terminal!!! I'd like it >to evolve into something more useful. If there is enough general interest, I could make the R4 binairies and libraries (for s300 and 400 machines) available. > Thank you in advance for your reply, > Mike Gourlay > mike@penguin.gatech.edu -- JEMilburn@lbl.gov ...!ucbvax!lbl.gov!JEMilburn Disclaimer: I speak for myself. -- JEMilburn@lbl.gov ...!ucbvax!lbl.gov!JEMilburn
harry@hpcvlx.cv.hp.com (Harry Phinney) (12/15/90)
Mike Gourlay writes: > I hear that HPUX 8.0 will be shipped out in June or July. > I also hear that X11R4 will be send with it. That is, if you have > `software support.' No, X11R4 is part of HP-UX 8.0 regardless of whether or not one has software support :-) I can't comment on the release date of 8.0. I would suggest getting that information from your local HP representative. > What if you don't have software support? [Gripes about > HP policy left for later.] Will someone please make a tar file of the > distribution tape, and put it somewhere? Or send the tar file to those > of us who have no support from HP? HP-UX 8.0 has not yet been released and in any case I suspect this would be less than legal, as HP-UX is a licensed product from HP. > My local HP rep was confused when I told her I knew of > people with X11R4 on an HP. She seems to think that it hasn't been > released yet. People have it, though! The X11 release 4 distribution from MIT contains code to support both the series 300 and the series 800 machines. It is quite possible for people to compile this code and use it (that is why we contribute this code for the MIT releases). In addition, we have previously announced in this group that there is an unsupported, prerelease version of our product R4 server for the 800s available via anonymous ftp. Not all local HP representatives have the time to keep up on news/notes (apparently some posters don't either). This server is available from: hpcvaaz.cv.hp.com (15.255.72.15) in the file: hpcvaaz:~ftp/pub/unsupported/XServers/X.R4.800.Z There is also a "readme" file which you may wish to look at. The server binary is _compressed_ so be sure to _uncompress_ it before you bother trying to run it. In addition there is a released and supported R4 server available for the series 300 machines. This can be retrieved from: hpcvaaz:~ftp/pub/XServers/X.7.03.300.Z and: hpcvaaz:~ftp/pub/XServers/X.7.03.300.readme If you need some of the libraries not distributed with HP-UX, you can get hpcvaaz:~ftp/pub/MitX11R4/libs.s300.Z hpcvaaz:~ftp/pub/MitX11R4/libs.s800.Z These contain both libXaw and libXmu. Get the README file in that same directory for information on unpacking these archives. > Will the X11R4 binaries running on HPUX 6.5 also run > on a similar machine running 7.0? Specifically, will the X binaries > running on a HPUX 6.5 9000s319C+ run on a HPUX 7.0 9000s370? > Libraries? Server? Client? Yes, with a couple of caveats. HP-UX 7.0 added support for some additional graphics display systems. The HP-UX 6.5 X server will not work on those display systems released after HP-UX 6.5. In addition, most HP-UX 7.0 X installations will cause the user's environment to have the DISPLAY variable set to "local:0.0" or something similar. The HP-UX 6.5 (or earlier) client libraries/binaries will not recognize this as meaning a Unix-domain socket. These binaries will work fine if DISPLAY is set to `hostname`:0. > Why does the X11R3 that comes with HPUX 7.0 not include > some of the X routines that many public programs expect? We do not ship some of the libraries distributed in the MIT X Consortium distribution because of possible support problems. It may appear to be bureaucratic overhead, but for us to release a product we have to show that we've adequately tested it. This involves having a test suite which covers a certain percentage of the paths through the code, and enough test hours (with few enough bugs found) to ensure that the customer will be able to make good use of the product. We simply do not have the "engineering resources" (i.e. enough people) to adequately test all of the various pieces of the public X11 release. > Why does HP consider fixing their software bugs (called an update > of the OS) something for which you have to pay big money for a > support contract? ... > The new versions are bug fixes. That is not right. While new versions do indeed contain bug fixes, they also generally contain significant new functionality. Do you believe it is free for us to continue to improve and enhance our products? You may not care whether I get paid, but I do. My salary comes from customers who purchase our products. If we simply give away all future work, where will my salary come from? > X is in the public domain. No, X is not in the public domain. The code from the MIT X Consortium is copyrighted by a host of people and corporations, including HP. > HP makes so called `inhancements' I don't think we've ever claimed to make any "inhancements":-) We have certainly made enhancements to the X server to allow our Starbase, GKS, and PHIGS libraries to better operate within the X environment, and to take advantage of some of the features of our particular graphics systems. We have also added input extensions to allow the server to deal with multiple and varied input devices, and have donated this input code back to the MIT X Consortium. Please understand that we participate in the MIT X releases precisely because we know that some customers require the latest-greatest version of X more quickly than we can "productize" it. These customers (quite possibly including yourself) are more than welcome to use the code distributed by MIT, but must realize that it will lack some of the above-mentioned enhancements and has not undergone the same level of testing as our product releases. > X should be compilable from the distribution > at MIT, and so capable people should be able to compile X11R4 without > waiting for the newest distribution from HP. Has anyone done that? Many people have done exactly that. If you are having problems, I would suspect that it is with the build process for the MIT X distribution and not with anything specific to HP. The main reason we made the above-mentioned, ftp-able server available is to aid those customers who also need to be able to run Starbase applications within X windows. <Much flaming deleted> > Thank you in advance for your reply, While ranting and flaming may make one feel a bit better during moments of frustration, it doesn't help motivate anyone else to provide assistance. Also, saying thank you at the end does little to remove the bitter taste of the preceding flames. Harry Phinney harry@cv.hp.com
jbb@hpcvlx.cv.hp.com (Jim B. Byers) (12/15/90)
>I hear that HPUX 8.0 will be shipped out in June or July. >I also hear that X11R4 will be send with it. That is, if you have >`software support.' Yes the 7.03 server is being moved and improved into 8.0. >What if you don't have software support? [Gripes about >HP policy left for later.] Will someone please make a tar file of the >distribution tape, and put it somewhere? Or send the tar file to those >of us who have no support from HP? The MIT tape is widely available. People on the net have responded with ftp sites when software has been requested on the net. > My local HP rep was confused when I told her I knew of >people with X11R4 on an HP. She seems to think that it hasn't been >released yet. People have it, though! The 7.03 server is R4. There is also a post 7.03 server that is ftp-able and runs fast on a 98550 card that is available via anonymous ftp. See my posting in the responses to "Re: Dream system" (in comp.sys.hp) describing where to get it. This server was announced in comp.sys.hp quite a while back. > Will the X11R4 binaries running on HPUX 6.5 also run >on a similar machine running 7.0? Specifically, will the X binaries >running on a HPUX 6.5 9000s319C+ run on a HPUX 7.0 9000s370? >Libraries? Server? Client? That has been our experience. As a general practice it is always better to have libraries that were compliled for the OS you are linking on. I would very much using the post 7.03 server on 7.0 or 7.03 and not use the 6.5 server. I don't know of anyone using the 6.5 server on a latter release or know why they would. > Why does the X11R3 that comes with HPUX 7.0 not include >some of the X routines that many public programs expect? Will X11R4 >lack this hideous problem? The MIT tape contains a huge mix of MIT blessed/authored "core" programs, and a large bunch of contributed programs that are in various states of quality and reliability. Which ones are of interest to you? If they all are, have you considered installing the MIT tape itself? Which one would you expect us to answer questions on if they were provided? Which ones would you expect us to bugfix if you were on support services and found a problem? > Why does HP consider fixing their software bugs (called an update >of the OS) something for which you have to pay big money for a >support contract? It is unfair. If you pay for UNIX, you should get UNIX, >and not have to pay incrementally to have HP give you UNIX over a period >of time. This is a general argument about support contracts and whether all future versions should be free. I will not add to much here since this gets covered elsewhere for Unix, Dos and application software. My 2 cents is that I see considerable effort going into new functionality. I expect that I will pay to upgrade to new versions of any software I use be it Unix, Dos or whatever. Nothing is free. If all future releases were free then it would cost more upfront. This would be unfair to those who buy a product and rarely upgrade. > X is in the public domain. HP makes so called `inhancements' >(read "incompatible frilly fluff") to it and sells it for money. >Motif is nice, but the other bugs and left-outs of HP's version of X >on HPUX are not worth it. X should be compilable from the distribution >at MIT, and so capable people should be able to compile X11R4 without >waiting for the newest distribution from HP.... Well, yes there is a public version of X. It is a reference port. In order to do this properly, most peices are the least common denominator. For example, the server assumes a dumb frame buffer. If you want to use that great, it is available. We worked quite hard to assure that the MIT stuff will compile up correctly on our machines. We also went to great effort to provide screen drivers so that one could compile the MIT server and still get good performance. If you think that the MIT stuff is bug free when it is released then you have not followed the long list of patches :-). In working with MIT in this way we try to aid the users out there who need the latest and greatest leading edge (bleading edge?) stuff while at the same time realizing that what many customers ask for is stable solid code (in supported releases.) Where have you found or X11 to be incompatible? Compatibility with the standard release is religion here. We certainly add things that people have said they needed (devices other than a mouse and keyboard), the sox11 driver, Access to the 3D accelerators - but were have we lost compatibility? Jim Byers Interface Technology Operation Marketing/Lab team "The X11/Motif/Vue/Architect folks in Corvallis"
jsadler@misty.boeing.com (Jim Sadler) (12/15/90)
/ misty:comp.sys.hp / mikeg@dali.gatech.edu (Mike Gourlay) / 7:01 am Dec 13, 1990 / >Hi HP people, > > I hear that HPUX 8.0 will be shipped out in June or July. >I also hear that X11R4 will be send with it. That is, if you have >`software support.' > > What if you don't have software support? [Gripes about >HP policy left for later.] Will someone please make a tar file of the >distribution tape, and put it somewhere? Or send the tar file to those >of us who have no support from HP? Why should you get the additions that HP put into X for free, while I have to pay for it ? As you say below X is in the "public domain", I beleive it's not. It is freely distributed. If you want it, why not ftp it from uunet or mit. > > My local HP rep was confused when I told her I knew of >people with X11R4 on an HP. She seems to think that it hasn't been >released yet. People have it, though! As they say not all reps are created equal and its not always the reps fault. The 842 and 852 intro is a good example. The sales force didn't know about them until a few days before they were introduced. They didn't get briefed until after the intro and some of the magazines knew about them way before the field offices. I think marketing at Corp HQ. messed but good. > > Why does the X11R3 that comes with HPUX 7.0 not include >some of the X routines that many public programs expect? Will X11R4 >lack this hideous problem? Good question ! I'd sure like to know! It's not restricted to just X either other libraries are missing !! > > Why does HP consider fixing their software bugs (called an update >of the OS) something for which you have to pay big money for a >support contract? It is unfair. If you pay for UNIX, you should get UNIX, >and not have to pay incrementally to have HP give you UNIX over a period >of time. It isn't as if the machine was inexpensive up front. It isn't as if >HPUX is not the most incompatible port of UNIX and X on the planet. It >most certainly is NOT that new versions of the OS are improvements >to UNIX. The new versions are bug fixes. That is not right. I kinda agree somewhat, except for the "most incompatible" try Xenix or apollo "UN*X". Software only update service is $600 per year list. I would think that it would be cheaper than that for you with the edu discount. The last time I check Sun and DEC charged for their update service. Is it competative ?? > > X is in the public domain. HP makes so called `inhancements' >(read "incompatible frilly fluff") to it and sells it for money. Stuff deleted. I really do wish HP would provide libraries and utilities that are more upto date. > > For now, I run many of the X apps I need on a Sparc, and just >have the Sparc use my display. EtherNet is wonderful. xhost + is >fantastic. My HPUX is being used as a dumb X terminal!!! I'd like it >to evolve into something more useful. While ours aren't "dumb X terminal"'s. I understand the feeling. > > Thank you in advance for your reply, > Mike Gourlay > mike@penguin.gatech.edu > >-------------------------------------------- >I don't officially speak for Georgia Tech or GTRI; Why the hell >should they care if I'm unhappy with HPUX, just because I work >with it every single day? The upper echelons have Sparcs and SGI's. >---------- jim sadler 206-234-9009 email uunet!bcstec!jsadler | jsadler@misty.boeing.com This service is brought to you by the computing mafia of Boeing (BCS). Oh ya None of the above is an opinion of The Boeing Co.
bb@beach.cis.ufl.edu (Brian Bartholomew) (12/15/90)
ATTRIBUTIONS: >> = mikeg@dali.gatech.edu (Mike Gourlay) > = milburn@me10.lbl.gov (John Milburn) ----- SMOKE SCREENS OBSCURING "REAL COST OF OWNERSHIP": >> I hear that HPUX 8.0 will be shipped out in June or July. I also hear >> that X11R4 will be send with it. That is, if you have `software >> support.' > HPUX rev 8.0 will indeed be a major upgrade to the operating system, > and as such, HP deserves to recoup the investment they put into it. > The mechanism for this is the software support contract I believe the mechanism for this is the "Operating System Purchase Price". If the ownership and use of a HP workstation over a five year lifespan will cost you xx dollars in OS upgrades ("Software Support"), than that number should play a part in the purchase price computations. >> Why does HP consider fixing their software bugs (called an update of >> the OS) something for which you have to pay big money for a support >> contract? It is unfair. If you pay for UNIX, you should get UNIX, >> and not have to pay incrementally to have HP give you UNIX over a >> period of time. It isn't as if the machine was inexpensive up front. >> The new versions are bug fixes. That is not right. > I agree with you to a certain extent. If you buy an hp box, with rev > 6.x, you should be provided with bug fixes to that revision of the OS > free of charge. However, if you choose not to purchase a support > contract, you have no inherent right to future major releases. I find > the cost of HP support contracts to be very reasonable, particularly > the material only level of support. Again, if the ownership of a workable Operating System requires you to purchase "Software Support" as one of its components, then consider it part of the purchase price. If true, then this implies that "HP-UX" without "Software Support" is not a workable option. > HP bundles a subset of X with their systems, along with some hp > enhancements, like mwm and hpterm. How much extra cost and effort would it cost HP to port the rest of the MIT X goodies, and ship them out as part of theier X distribution? Another $30 tape? How hard can it be to finish the job? > If you don't want what hp give you for free, or want more, get it > from MIT. Free (cough, cough)? See above paragraphs. Incidently, Sun is joining the ranks of the foolish vendors by threatening to unbundle the C compiler from SunOS. One of UNIX's strengths has always been its opportunities for symbiosys between tools - something that is hard to demonstrate as existing to a non-programmer (like a manager contemplating a workstation purchase). ----- COST OF "UNIQUE" SOFTWARE: >> It isn't as if HPUX is not the most incompatible port of UNIX and X on >> the planet. It most certainly is NOT that new versions of the OS are >> improvements to UNIX. >> Motif is nice, but the other bugs and left-outs of HP's version of X >> on HPUX are not worth it. X should be compilable from the >> distribution at MIT, and so capable people should be able to compile >> X11R4 without waiting for the newest distribution from HP. > Many people have done this. Put a dollar value on the effort (time = money) to track down, upgrade, port, and otherwise bring HP-UX up to snuff with what you could purchase from other vendors. Add it to the purchase price. ----- SALES REPRESENTATIVE PROBLEMS: >> My local HP rep was confused when I told her I knew of people with >> X11R4 on an HP. She seems to think that it hasn't been released yet. > Your rep is certainly uninformed, particularly with respect to PD > software under hp-ux. This is a common problem, as most hp sales reps > have no knowledge of the vast community of networked machines in the > world. Try to get one to exchange e-mail with you! >> My local rep blaimed it on inherent incompatibilities in UNIX. > Again, your local rep is uninformed. This is, I believe, a major > problem with hp. The sales reps are (severe generalization here) > remarkably unfamiliar with the systems they sell. Here we have two more peoples' experiences with non-knowledgable sales reps. Didn't the "Open Letter to HP" do any good? Come on, HP, get with the program. You seem to be able to do very well with Lab equipment support; what is so terribly different about workstations? "Any sufficiently advanced technology is indistinguishable from a rigged demo." ------------------------------------------------------------------------------- Brian Bartholomew UUCP: ...gatech!uflorida!mathlab.math.ufl.edu!bb University of Florida Internet: bb@math.ufl.edu -- "Any sufficiently advanced technology is indistinguishable from a rigged demo." ------------------------------------------------------------------------------- Brian Bartholomew UUCP: ...gatech!uflorida!mathlab.math.ufl.edu!bb University of Florida Internet: bb@math.ufl.edu
mikeg@dali.gatech.edu (Mike Gourlay) (12/18/90)
John Milburn replies to my original article: }The only reason you don't have support is because you }choose not to. WRONG. The "only" reason I do not have support is that I do not run GTRI, that support is considered unaffordable, and in the past, we have not had good experience with HP computers. (We have great expereience with other hardware HP makes, like their 8510 network analyzer, I must say.) }If you really want a full X implementation, get the MIT }distribution and compile it yourself. As I posted, I do not have the disk space for doing that, since there can be NO down time for X on this, I need a running copy of R3 to run while I have the R4 source, object files, libraries and binaries. That's quite a bit more than I can fit. --------------------------------------------- Harry Phinney writes: } previously announced in this group }Not all local HP representatives have the time to keep up on news/notes }(apparently some posters don't either). Some of the posters have not read this group and all of its messages since the beginning of time. I have only had this box for one quarter. As for "Not all local HP reps..." having time to keep up with news, my reps don't have InterNet access, much less news from this group. } This server is available from: }hpcvaaz.cv.hp.com (15.255.72.15) in the file: }If you need some of the libraries not distributed with HP-UX, you can get }These contain both libXaw and libXmu. Get the README file in that }same directory for information on unpacking these archives. I have asked some times before, and got no reply, cincerning the Athena Widgets. My local rep told me that they were simply not available for HP. Naturally she's wrong, since at the very most inconvenient, one can make X from the MIT distribution, which (as an HP person points out) is not an optimal solution by any means. In fact, the main two reasons (in order) for me wanting to use R4 is an strong need for the GKS libraries, and generally R4 runs things that R3 does not that I want to run. }While new versions do indeed contain bug fixes, they also generally }contain significant new functionality. Do you believe it is free for us }to continue to improve and enhance our products? You may not care }whether I get paid, but I do. My salary comes from customers who }purchase our products. If we simply give away all future work, where }will my salary come from? As I posted, there is no such thing as new functionality as far as UNIX is concerned. Translated from HPese, this means "bug fixes," and "putting in standard UNIX libraries that HPUX lacked." Don't get sappy about your pay. Obviously GTRI does purchase HP products. If we don't buy the "software support," you'll still get paid. Whether HP pays you to do a complete, running version of UNIX, or an incomplete, buggy version of some weird hybrid of two flavors of UNIX is my concern. I'd rather you get payed for a good job. If HP tells you that incomplete, and bugs, are okay, then you can't help it. Fine. That's between HP and their customers, not you and their customers. Do not tell me that I or any set of customers is responsible for your pay. Whether that is true or not is irrelevant to this discussion. However, if you want to take up the matter of your pay with me, then I will take up the matter of a lousy job of programming with you. I'd rather keep this impersonal, and give the programmers the benefit of the doubt, and hope that no self respecting systems programmer would give their blessing to what is running on my machine right now. I'd rather like to think that HP said that "this is good enough for government work," and not give the programmers a part in the decision to market this OS. If you want to tell me that you condone HPUX 7.0, and that you think it is acceptable, then I see no problem with it being free, and you not getting payed. I suggest you drop this line of reasoning, and leave further comments about your pay out of this. ] X is in the public domain. }No, X is not in the public domain. The code from the MIT X Consortium }is copyrighted by a host of people and corporations, including HP. What's your point? GNU software is also copyrighted. I call it PD. If there is some legal technicality of words here, then maybe PD is not the term. I mean free, legal to distribute, use etc. I mean that I don't have to pay for it. If that is what PD means, then X is in the public domain. ] HP makes so called `inhancements' }I don't think we've ever claimed to make any "inhancements":-) That's that the local representative called them. }We have }certainly made enhancements to the X server to allow our Starbase, GKS, }and PHIGS libraries to better operate within the X environment, and to GKS and PHIGS work? Those are what I'm looking for. That's the reason I want R4 (other reasons less important.) I want to avoid implementing GKS, and R4 has it. If HP has it the I'm happy for Christmas. Now, where is my present? Where do I get a working copy of GKS for my HP? }take advantage of some of the features of our particular graphics }systems. We have also added input extensions to allow the server to }deal with multiple and varied input devices, and have donated this input }code back to the MIT X Consortium. Please understand that we }participate in the MIT X releases precisely because we know that some }customers require the latest-greatest version of X more quickly than we }can "productize" it. These customers (quite possibly including }yourself) are more than welcome to use the code distributed by MIT, but }must realize that it will lack some of the above-mentioned enhancements }and has not undergone the same level of testing as our product releases. I don't care if it's supported. }While ranting and flaming may make one feel a bit better during moments }of frustration, it doesn't help motivate anyone else to provide }assistance. I know. Money helps motivate anyone to provide assistance. ----------------------------------- Jim B. Byers replies: }The MIT tape contains a huge mix of MIT blessed/authored "core" }programs, and a large bunch of contributed programs that are in }various states of quality and reliability. Which ones are of interest }to you? Athena Widgets, GKS, and PHIGS are of interest to me. Are those considered by HP to be unreliable, and low quality } If they all are, have you considered installing the MIT tape }itself? Which one would you expect us to answer questions on if }they were provided? Which ones would you expect us to bugfix if }you were on support services and found a problem? If I were on support, I would expect support for Athena Widgets, GKS and PHIGS. }My 2 cents is that I see considerable effort going into new }functionality. I expect that I will pay to upgrade to new versions }of any software I use be it Unix, Dos or whatever. Nothing is free. Wrong. GNU, and X are free to be used by us. }If all future releases were free then it would cost more upfront [sic]. This }would be unfair to those who buy a product and rarely upgrade. It is unfair to pay any money for a UNIX port and have it not work. If some people do not need UNIX to work, then they get a bargain out of "software support," because they do not need it. I use much of what would be UNIX if HPUX implemented it. }We worked quite hard to assure that }the MIT stuff will compile up correctly on our machines. } We also went to }great effort to provide screen drivers so that one could compile the }MIT server and still get good performance. the MIT stuff will compile up }correctly on our machines. As it should be, since you advertise that these boxes run X. }Where have you found or X11 to be incompatible? Compatibility with }the standard release is religion here. We certainly add things that }people have said they needed (devices other than a mouse and keyboard), }the sox11 driver, Access to the 3D accelerators - but were have we }lost compatibility? Oh, that's an easy one. There are plenty of public domain programs that expect Athena Widgets that do not compile. I even had one program that used Motif widgets (one of the few I've found) that tried to use some Motif Widgets that were not on my machine, which is supposed to have it. As far as UNIX programs, I won't even bother telling you much more than that is a very, very uninformed, naive question, and a ridiculous claim. I agree thoroughly with Brian Bartholome's comments in his reply in this thread.
frank@grep.co.uk (Frank Wales) (12/18/90)
[Disclaimer: I am not, and never have been, an employee of HP, just a long-time satified customer, blah-blah-blah.] In article <1990Dec13.100152@dali.gatech.edu> mikeg@dali.gatech.edu (Mike Gourlay) writes: > I hear that HPUX 8.0 will be shipped out in June or July. >I also hear that X11R4 will be send with it. That is, if you have >`software support.' > > What if you don't have software support? [Gripes about >HP policy left for later.] Will someone please make a tar file of the >distribution tape, and put it somewhere? Or send the tar file to those >of us who have no support from HP? What a good idea. But why stop with HP? Why not make copies of all the common OSs from all vendors and put them somewhere public? Why bother actually paying for them at all? After all, they cost nothing to develop, duplicate and distribute, right? > Why does HP consider fixing their software bugs (called an update >of the OS) something for which you have to pay big money for a >support contract? Most software updates include new commands, new facilities and performance or compatibility improvements which are worth actually paying for. And remember that much of the code in HP-UX wasn't actually written by HP at all. When was the last time you had to debug a two million line package you didn't write? >It is unfair. No, it isn't. Name any commercial vendor who will provide on-going support and updates completely free of charge. >If you pay for UNIX, you should get UNIX, and not have to pay >incrementally to have HP give you UNIX over a period of time. Huh? I wasn't aware that I was getting HP-UX on the installment plan. UNIX is a moving target from any vendor, which is one reason why software updates exist and cost money. >It isn't as if the machine was inexpensive up front. I find HP machines quite competitive for price/performance, especially when cost of ownership and reliability are considered. >It isn't as if HPUX is not the most incompatible port of UNIX and >X on the planet. It doesn't sound like you've ported much to HP-UX that wasn't Sun or VAX dependent. It's very accommodating of foreign software. >It most certainly is NOT that new versions of the OS are improvements >to UNIX. The new versions are bug fixes. That is not right. You are correct; the statements preceding "That is not right" are not right. How many new versions have you seen? How many other vendors' versions of UNIX have you seen? How do you determine whose is better? > X is in the public domain. No, it isn't. It's freely distributable within certain constraints. >HP makes so called `inhancements' >(read "incompatible frilly fluff") to it and sells it for money. Sun, AT&T, DEC, and most other workstation vendors also make enhancements. Is everyone evil? >X should be compilable from the distribution >at MIT, and so capable people should be able to compile X11R4 without >waiting for the newest distribution from HP. Has anyone done that? You said earlier that they had. Besides, how long do you think it takes to test something the size of X, especially when you didn't write it? How long does it take to distribute zillions of copies of it? Do you think HP are just dawdling because they feel like it? >My local rep blaimed it on inherent incompatibilities in UNIX. WHAT?!?!?!!? >That's ridiculous. Inherent incompatibilities in UNIX? No such thing. See Configure by Larry Wall for some idea of the nonexistent incompatibilities between different UNIX versions. Read comp.unix.* and comp.lang.c for some more. >Between SYSV and BSD, okay. Which one of these isn't UNIX, since obviously they can't both be (UNIX isn't incompatible with itself, after all)? >Between _either_ BSD or SYSV and HPUX, I see incompatibility >(or is it incompetence?) Ahem. HP-UX is one of the most reliable UNIX versions I know of. It got that way because HP actually debug the thing before they ship it, unlike certain other workstation vendors I could mention. Their implementation of some things is also cleaner (e.g., context-dependent files versus mutant symlink file systems from Hell). Maybe you're not prepared to pay for reliability, and want Untested Creeping Features (tm) instead. Maybe that puts you in the minority of HP's customers. >That, according to her, is why WE HAVE TO PAY HP TO GIVE US PUBLIC >DOMAIN SOFTWARE AND BUG FIXES TO THEIR WEIRD, BUGGY, INCOMPLETE UNIX >PORT AND SPECIAL VERSION OF X. For some reason, the phrase "eat my shorts" comes to mind. It always seems to happen when someone's BIFF key gets stuck down. Can't imagine why. -- Frank Wales, Grep Limited, [frank@grep.co.uk<->uunet!grep!frank] Kirkfields Business Centre, Kirk Lane, LEEDS, UK, LS19 7LX. (+44) 532 500303
jsadler@misty.boeing.com (Jim Sadler) (12/18/90)
/ misty:comp.sys.hp / harry@hpcvlx.cv.hp.com (Harry Phinney) / 1:29 pm Dec 14, 1990 / >Mike Gourlay writes: MUCH DELETED. > >> Why does the X11R3 that comes with HPUX 7.0 not include >> some of the X routines that many public programs expect? > >We do not ship some of the libraries distributed in the MIT X Consortium >distribution because of possible support problems. It may appear to be >bureaucratic overhead, but for us to release a product we have to show >that we've adequately tested it. This involves having a test suite >which covers a certain percentage of the paths through the code, and >enough test hours (with few enough bugs found) to ensure that the >customer will be able to make good use of the product. We simply do not >have the "engineering resources" (i.e. enough people) to adequately >test all of the various pieces of the public X11 release. > I can accept the above. But why doesn't HP provide the untested code as "user supported" ? DEC does this with their install tapes. I'm not sure, but I believe they do it for no extra charge (at least it doesn't show up as a seperate charge). This is the first time that I have heard why HP doesn't ship certain libraries, up till now I thought someone was makeing capricious and arbitrary decisions on what to not included with HP-UX. Is the above reason why certain berkley utilities and librarys are missing ? MORE DELETED. > >While ranting and flaming may make one feel a bit better during moments ^^^^^^^ Two points I'd like to make: 1. If it was just moments that happened occasionally, I would agree with you. When it happens week after week in the normal course of your work it gets extremely fustrating. Please don't think I agree with everything that Mr. Gourlay say's, I don't, but I do empathize with him. These feeling are compounded if you work in a multi-vendor shop and you can walk to a brand X machine and it does what the HP didn't. 2. Please, Please don't take my or anyone else postings personally. Most of the time the reason I make critical type postings is because I can't get the real information anyother way and I hope it might influence someone to correct my perceived problem. >of frustration, it doesn't help motivate anyone else to provide >assistance. Also, saying thank you at the end does little to remove the >bitter taste of the preceding flames. > >Harry Phinney harry@cv.hp.com >---------- jim sadler 206-234-9009 email uunet!bcstec!jsadler | jsadler@misty.boeing.com This service is brought to you by the computing mafia of Boeing (BCS). Oh ya None of the above is an opinion of The Boeing Co.
jbb@hpcvlx.cv.hp.com (Jim B. Byers) (12/18/90)
OK, now we are getting somewhere. I understand you to be saying
the following things.
1) You are using an HP-UX machine.
2) Do not have a support contract because your organization decided
not to.
3) You have only had access to the net for the past quarter.
4) You have need for some X11 based programs that you have seen.
Someone on your site compiled them for another machine, but not for
HP-UX.
5) You would be willing to compile them if you have to. (It is assumed
that you would prefer a binary.
6) You need the Athena widgets to compile these it. You want these to
be shipped as part of HP-UX. If you were on support you would expect
us to bug fix these and work with you on work arounds etc.
7) You also need a copy of GKS. I take this to mean the xgks stuff
found in the contrib section of the MIT tape under
contrib/libraries/xgks
8) You cannot install the MIT tape and build what you want due to space
constraints.
9) You are angry and are letting us know.
Lets cover these.
Re: 4) You have some X11 programs that you want to use. These are
applications that you have seen, that may be available as source. I
understand these to be PD rather than purchased applications. I still
am not sure what these are. In the past requests for such binaries have
been fulfilled by the net. Someone took the time to compile them for
use on another system on your site. Other sites across the country
probably have compiled them for HP-UX. Give the net a try. No vendor
will ever provide every PD application that is passed around on the net. Why would they, they
Why would they? They are passed around quite efficiently on the net :-).
Re: 5) The obvious problem here is obtaining the right patches etc. to
the MIT tape. I know you have a great faith in the MIT tape software
(even in the contrib) section but it is software whose major goal is
features not stable code. This is not to say that the contributors,
HP included, do not strive for good code. My point is that this is the
first time that much of this code has seen the light of day.
Re: 6) I would recommend that you consider ftp-ing the libraries from
hpcvaaz as Harry describes in his response. You would prefer that
these were included in HP-UX. This is a tough one. We wrestled with
this one when the Athena widgets first came out (R2).As you mentioned
people would expect us to support/bug fix them if they were in HP-UX.
It frankly has not come up as much of an issue as the libraries are
available via hpcvaaz and most commercial developers use Motif if they
are widgeteers. In short we (here on the X11/Motif/VUE team ) haven't
heard many complaints. Really.
We still would have had a problem satisfying you with the 7.0 release.
The 7.0 release was completed *before* R4 came out. So we would have
had to rely on hpcvaaz to satisfy your needs.
In a similar vain you want the latest thing on the net but want it
to be there in the last standard release. You expect these to be
bug free and part of an extremely stable release. This is hard to
deliver on! Are these goals mutually exclusive or am I missing
something?
Re: 7) GKS has been on the HP-UX boxes for a long time. It has always
been available as a separate product. Since you need a PD type version
you probably are talking about the xgks from the contrib/toolkit section
of the MIT tape. As a point of reference, there are 9 toolkits in this
section. I have no direct info on xgks, since I have never been asked
about it. How about posting a request on the net (sansflames) and see
what kind of response you get. Figraro from template as well as an
HP implementation of Phigs can be purchased. PEX from the X consortium
is still rather young (I believe) and is I am unsure of its availability
outside the consortium. Someone else will have to help out on this.
Re: 8) The size of the MIT tape is huge, agreed. There are many parts
to it. Different people are interested in different parts. Thats the
beauty of PD software, but it is unrealistic to expect all vendors
to ship all parts to all people for "free".
Re: 9) Personally I thinks flames can be a good sign. The person
obviously feels strongly about the subject. On the other hand, keep
in mind that if you really want answers, it is not unreasonable to
try a more straight forward query first. The Athena widgets have
been on hpcvaaz for a long time. Various people on the net have
R4 binaries/source. All these could have been found without extrainious
heat.
I was very serious about the fact that compatibility is religion here.
While I am involved with only X11/Motif here which makes me naive
(as each of us is - knowing only a small fraction of the world's
knowledge) I know the immense amount of importance it take on here.
I am glad to have read your comments. And I hope that this helps!
Jim Byers
ITO Marketing/Lab Team
"Never attribute malice where ignorance will suffice."
A rule for livinggordon@maxwell.waterloo.edu (Gordon R. Strachan) (12/18/90)
In article <101950173@hpcvlx.cv.hp.com> jbb@hpcvlx.cv.hp.com (Jim B. Byers) writes: > >OK, now we are getting somewhere. I understand you to be saying >the following things. > >1) You are using an HP-UX machine. >2) Do not have a support contract because your organization decided > not to. >3) You have only had access to the net for the past quarter. >4) You have need for some X11 based programs that you have seen. > Someone on your site compiled them for another machine, but not for > HP-UX. stuff deleted >Lets cover these. > > >Re: 6) I would recommend that you consider ftp-ing the libraries from >hpcvaaz as Harry describes in his response. You would prefer that >these were included in HP-UX. This is a tough one. We wrestled with >this one when the Athena widgets first came out (R2).As you mentioned >people would expect us to support/bug fix them if they were in HP-UX. >It frankly has not come up as much of an issue as the libraries are >available via hpcvaaz and most commercial developers use Motif if they >are widgeteers. In short we (here on the X11/Motif/VUE team ) haven't >heard many complaints. Really. > > Okay, I interpret this to say that unless we complain, we won't get the Athena widget set shipped in v8.0. So okay, I'm complaining! I want the Athena widgets put into the standard HP distribution. What you say is in essentially correct. When I write X programs I use the motif widgets because they are better than the Athena widgets. But, the problem comes when you want to compile public domain code, the vast majority of which still use the Athena widgets. I don't think it is such a big deal for HP to stick the library in the distribution and mark it as unsupported. After all, wasn't that what the /usr/contrib directory was for, unsupported contributed software? At least make it an option for those of us that want it. The lack of Athena widgets and totally brain dead man page names for the Xt library functions are the two biggest problems with HP's X software. I would like to add though, that on the hole I am quite happy with HP's X software. I don't consider HP's additions to the X server to be minor. The reason I never installed the R4 stuff from MIT is I didn't want to lose these additions, most notably the starbase extensions. Also, I have been satisfied with HP software support. It seems to me, although this is just subjective, that HP has a lot more updates than other major vendors (ie Sun) and their O/S is more stable. The software support is expensive but, so far at least, I think it has been worth the money. > >I am glad to have read your comments. And I hope that this helps! > >Jim Byers >ITO Marketing/Lab Team Gordon
harry@hpcvlx.cv.hp.com (Harry Phinney) (12/19/90)
Jim Sadler writes: > But why doesn't HP provide the untested code > as "user supported" ? We've tried something like this in the past (the SE support tape), and we may do something similar in the future. Our support organization(s) have expressed reservations about such things, stating that customers expect support for anything and everything we ship regardless of labeling. While I personally feel we should provide such a package, I can't guarantee it will happen. We are trying to accomplish much the same goal with the ftp-able X stuff on hpcvaaz.cv.hp.com. > This is the first time that I > have heard why HP doesn't ship certain libraries, up till now I > thought someone was makeing capricious and arbitrary decisions on what > to not included with HP-UX. Is the above reason why certain berkley > utilities and librarys are missing ? I suspect that this is the reasoning, although I don't have first hand knowledge in the case of Berkeley stuff. While I personally have no explanation for the current lack of things like bcopy(), index/rindex, etc., I believe the situation will improve some with HP-UX 8.0. > 2. Please, Please don't take my or anyone else postings > personally. Most of the time the reason I make critical type > postings is because I can't get the real information anyother > way and I hope it might influence someone to correct my perceived > problem. I can't remember ever being offended by one of your postings. I am more than willing to listen to criticism and certainly try to learn from it. I recognize that the previous poster has some legitimate complaints and problems, but for me they tended to get lost within the unusually large amount of flames contained in the posting. Harry Phinney harry@cv.hp.com
erc@pai.UUCP (Eric F. Johnson) (12/21/90)
In a general controversy about what HP bundles with its version of the X Window System and Motif, there are a few points that should be made. 1) HP is apparently NOT shipping the Athena widget library with its version of the X Window System. This is libXaw.a in /usr/lib. Since Motif is not a free library, many people writing free X software use the Athena widgets. If you want to use free X software and that software needs the Athena widget library, you're stuck on an HP. Jim B. Byers of HP (jbb@hpcvlx.cv.hp.com) writes that fixing bugs and support are the main reasons why HP decided not to include the Xaw library, but I think the real issue is that both the Motif library (Xm) and the Athena widget library (Xaw) depend on the X Toolkit Intrinsics (Xt)--and the OSF had to change Xt for Motif 1.0. This means that HP would have to ship two versions of Xt (one with the OSF/Motif modifications, one without). This would be a very confusing situation. I can understand why HP didn't ship the Athena widget library, as neither did Data General or SCO, but I still wish I could compile Athena widget programs on an HP workstation (I'm not letting the others off the hook either). Happily, this situation is much better with the latest versions of Motif and X. With X11 Release 4 (up to patch 18) and Motif 1.1, both Xaw and Xm can share the same version of the Xt library. 2) I'm glad that HP is spending the time to fix bugs and make sure everything works right, but does it have to take so long? X11 Release 3 came out in 1988. I received HP-UX 7.0 (with R3) in spring 1990, well AFTER I had installed MIT X11 Release 4 on our company Sun workstations. (X11 Release 4 came out in January, 1990 to the public. HP, as an X Consortium member, had it much earlier.) Yes, I want HP to fix bugs, but I also work for a company that produces X software. The differences between X11 R3 and R4 are dramatic. Release 4 is the most professional release from MIT to date. It uses a lot less RAM and the performance is much better. In addition, a number of bugs were fixed in R4, that we still have to deal with in R3. The new routines added in R4, such as XGetGCValues() and XIconfiyWindow(), are very useful, but not available on my HP. (XIconfiyWindow() can be emulated under R3--and I've done it--but you'll have fun trying to make a portable R3 version of XGetGCValues() that doesn't break the rules.) In addition, if you ask on the net, the resounding response will be to not use R3 or Motif 1.0 for commercial applications (especially due to memory leaks). Everyone advises using R4 and Motif 1.1. This is an incentive not to sell HP-based systems, but instead sell DG or Sun or whatever (Boulware Technologies, my employer, sells industrial automation systems, reselling computers from a number of vendors). I like HP equipment and have found HP boxes to be incredibly durable and resistant to untold abuse on the factory floor. But, I wish the system was more up to date. Now, HP is saying we'll receive HP-UX 8.0 perhaps sometime in March. In real terms, I expect this to be May-June (based on the experience with HP-UX 7.0). X11 Release 5 is expected sometime about July, 1991, or about the time I'll be installing R4 on the HP's. To put the time issue in another manner: 3) Our firm IS on HP software support. Could HP please (pretty please with sugar on it) release interim versions of the OS with just one feature enhanced? In this case, I don't need everything HP-UX 8.0 offers, but I would have appreciated (and still would appreciate) a 7.0 with X11 R4 and Motif 1.1 before 8.0 comes out. The issue is probably moot with 8.0 now, but how about 8.0 with X11 R5 planned for, say, December 1991 (or about 6 months after the MIT release of R5)? This would keep my HP systems more up to date and would encourage selling HP, rather than other vendors' CPUs. It would also give you HP folks time to fix bugs and test out the system. Remember, HP will get R5 from MIT a lot sooner than I will (BTI is not a member of the X Consortium at this time). If cost is an issue, I will gladly trade every Software Release/Status Bulletin from now to enternity for interim bug fix and enhancement tapes from HP. (The SR/SB is utterly useless in its present form anyway.) To be more precise, I would rather get an HP tape with X11 R5 than any and all SR/SBs, espcially if I could get the tape BEFORE X11 R6 :-) comes out. It would also be nice if HP would tell its customers when (even a ballpark figure would help) HP plans to ship X11 R5. (I mention this because I had to endure a host of HP reps talking about the wonders of X11 R3 when HP-UX 7.0 came out. Not a single HP person in the 7.0 seminar even knew about R4--which I had already installed on Suns. None of these people could even tell me if HP had ANY planes to ever ship an R4 product. So, I'm asking now, will and if so, when will HP ship X11 R5? This may seem premature, but it would be nice if HP could clue us in to its future plans. I don't need a firm date, but something like "HP-UX 9.0 [or OSF/1?] will have that, and we expect it in 2nd Q 1992" would help.) 4) Why not compile the MIT R4 on my HP, you may ask. Why not? Because HP decided to ship only object modules (.o files) with the MIT X11 R4. This means that if I have a problem, and there undoubtedly will be problems, I am SOL. When problems arose on the Suns, I could modify the source. With HP, I don't have that luxury since HP decided somehow that they would only pretend to go along with the MIT release. I wouldn't mind this if I would get more up-to-date releases of X from HP. But I don't. I was still running R2 on HP-UX 6.5 AFTER I had installed R4 on Suns (HP-UX 7.0 appeared FULLY in about the May, 1990 time frame; R4 from MIT came out in Jan., 1990). Now, it looks like I will get HP R4 about the time I install R5 on the Suns. Argh. Please ship full sources with the MIT X11 R5 and not these object modules. 5) I, at least, am not asking HP to support, enhance and fix bugs on all the X Window contrib software. (I wouldn't wish that on anybody.) But, it would be nice if all items that are considered X Consortium standards are included with the HP version of X. Generally, this is the core part of the X release tapes. Since HP *IS* a member of the X Consortium, it shouldn't be too hard to ask MIT what exactly is considered a standard part and what isn't. I would appreciate the Athena widget library included as part of HP's X. (I generally try to use Motif instead of Xaw, but for a lot of Athena-based software it's easier to compile using the Xaw widgets than it is to port to Motif.) jbb@hpcvlx.cv.hp.com (Jim B. Byers of HP) writes: [...much deleted...] > 5) You would be willing to compile them if you have to. (It is assumed > that you would prefer a binary. > Re: 5) The obvious problem here is obtaining the right patches etc. to > the MIT tape. I know you have a great faith in the MIT tape software > (even in the contrib) section but it is software whose major goal is > features not stable code. This is not to say that the contributors, > HP included, do not strive for good code. My point is that this is the > first time that much of this code has seen the light of day. The first time this code saw the light of day was in January, 1990, or about a year ago. Yes, a lot of the contrib code had problems. In my tests, I could only get about 40 percent of the contrib code to compile under X11 R4 on a Sun. But, I couldn't even attempt to compile most of it on an HP, even now, due to the lack of the Xaw library (and R4 features). Forty percent is better than none, I'd say. > Re: 6) I would recommend that you consider ftp-ing the libraries from > hpcvaaz as Harry describes in his response. You would prefer that > these [Athena widgets] were included in HP-UX. So would I. > This is a tough one. We wrestled with > this one when the Athena widgets first came out (R2). I believe HP was pushing the HP widget set at that time; perhaps this was part of the wrestling? Nothing against the HP widget set, but why didn't you include the Athena set as well? You obviously had to have it, since just about every standard X program in /usr/bin/X11 on the HP uses the Athena widgets, including xterm, xcalc, xclock, xload, and so on. Did HP write their own replacement for the Athena widgets? Or, did HP port the Athena widgets and then decide not to include the widget library with the HP release? > As you mentioned > people would expect us to support/bug fix them if they were in HP-UX. > It frankly has not come up as much of an issue as the libraries are > available via hpcvaaz and most commercial developers use Motif if they > are widgeteers. In short we (here on the X11/Motif/VUE team ) haven't > heard many complaints. Really. Consider this a friendly complaint. > Re: 8) The size of the MIT tape is huge, agreed. There are many parts > to it. Different people are interested in different parts. Thats the > beauty of PD software, but it is unrealistic to expect all vendors > to ship all parts to all people for "free". I think the original poster was really asking for the Athena widget library, so he could compile Athena (Xaw) programs. > I was very serious about the fact that compatibility is religion here. > While I am involved with only X11/Motif here which makes me naive > (as each of us is - knowing only a small fraction of the world's > knowledge) I know the immense amount of importance it take on here. > I am glad to have read your comments. And I hope that this helps! > Jim Byers > ITO Marketing/Lab Team Don't get me wrong, I like HP systems and I appreciate the magnitude of the work HP folks do. But, a lot of the code I use every day includes the Athena widget set. It would be nice to have this on an HP. Even more of the code I use every day would be enhanced if it could run under X11 R4. Thanks for responding anyway, notwithstanding the tone of the original query that started this thread. (Just to clarify: That query was NOT written by me.) Have fun, -Eric -- Eric F. Johnson phone: +1 612 894 0313 BTI: Industrial Boulware Technologies, Inc. fax: +1 612 894 0316 automation systems 415 W. Travelers Trail email: erc@pai.mn.org and services Burnsville, MN 55337 USA
ian@dms.cdc.com (Ian Hogg) (12/27/90)
This in regards to discussions about HP's lag behind MIT X releases: I build the whole core distribution from HP. I generally just play with the server because the one built from the MIT tape is a whole lot slower than what we get on HP-UX tapes. However, I do install MIT libraries and useful clients. So I end up with a hybrid system that has HP's server and the MIT libraries. Since we license Motif I've got 1.1 compiled. I also keep HP's libraries and include files around in case we have problems. I also generally recompile things like libXaw with HP supplied Xlib and Xt. This has worked reasonably well. I really, really wish thar HP would get X releases out much sooner but I've found this solution to minimize the aches of compiling apps thar depend on newer releases of Xlib/Xt.