INFO-VAX Tue, 25 Sep 2007 Volume 2007 : Issue 524 Contents: Article of Interest. Re: Article of Interest. Re: despair Re: despair Re: despair Re: despair Re: despair Re: despair Re: despair Re: despair Re: despair Re: despair Re: despair Re: despair Re: despair Re: despair Re: despair Expected votes for a boot node Re: Expected votes for a boot node Re: Expected votes for a boot node Re: Expected votes for a boot node Re: Expected votes for a boot node Re: Expected votes for a boot node Re: Expected votes for a boot node Re: Expected votes for a boot node Re: Expected votes for a boot node Free software hacking,anti-hacking,registry tweaks,compter tricks Re: Impact VMS version on CPU speed Re: UK Based Alpha Station For Sale. Re: UK Based Alpha Station For Sale. ---------------------------------------------------------------------- Date: Tue, 25 Sep 2007 07:51:36 -0400 From: Chuck Aaron Subject: Article of Interest. Message-ID: I think they need to be running VMS! http://www.cnn.com/2007/US/09/24/homelandsecurity.computers/index.html ------------------------------ Date: Tue, 25 Sep 2007 13:22:34 -0000 From: IanMiller Subject: Re: Article of Interest. Message-ID: <1190726554.772902.254490@w3g2000hsg.googlegroups.com> Reading that article it sounds more like operational incompetence rather than anything to do with the platform they where using. ------------------------------ Date: 25 Sep 2007 11:53:18 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: despair Message-ID: <5lsb5dF9qudpU1@mid.individual.net> In article <1190596861.683981.103920@k79g2000hse.googlegroups.com>, AEF writes: > On Sep 23, 8:59 pm, b...@cs.uofs.edu (Bill Gunshannon) wrote: >> In article <1190592151.788617.205...@g4g2000hsf.googlegroups.com>, >> AEF writes: >> >> > On Sep 23, 5:10 pm, b...@cs.uofs.edu (Bill Gunshannon) wrote: >> >> In article , >> >> "Tom Linden" writes: >> >> >> > On Sat, 22 Sep 2007 10:35:17 -0700, Bill Gunshannon >> >> > wrote: >> >> >> >> In article , >> >> >> "Tom Linden" writes: >> >> >> >>> Penny-wise, pound-foolish? The development and maintenance of C code is >> >> >>> more costly. So pay now or pay later. >> >> >> >> Sorry, no basis in fact for that. Might actually be cheaper as the pool >> >> >> of available C Programmers is considerably higher than say the pool of >> >> >> available Fortran, COBOL or (gasp) PL/I programmers. And if the shop has >> >> >> reasonable coding standards, maintenance does not have to be a nightmare. >> >> >> (I know, having worked in large shops where turnover every 1-3 years is >> >> >> prett close to 100%.) >> >> >> >> bill >> >> >> > No, I do have a basis for that (1) 41 years of experience >> >> >> More than me, but not by much if you count the times I wandered into >> >> the IT field for short periods of time before I took it on full-time. >> >> >> > and (2) if you >> >> > compare >> >> > PL/I and C programs for equivalent functionality, you will find that the C >> >> > version >> >> > will have 2 to 3 times as many source language statements. >> >> >> And that means? Somehow, I don't think number of statements is a strong >> >> basis for deciding the economy of a an Information System. If it were, >> >> there would be a lot more stuff written in BASIC. :-) And loaded with >> >> GOTO's. >> >> >> And let's not forget that PL/I compilers are rarer than PL/I programmers >> >> which kind of limits it's usefulness beyond some very small niches. >> >> > Weren't you the one who said bad code comes solely from bad >> > programmers and that the language makes zero difference? Did I miss >> > something? >> >> What does that have to do with the fact that PL/I compilers are rarer >> than PL/I programmers? Or the number of statements it takes to write >> a program in any given language? > > Sorry, I was referring mostly to your excessive GO TO's complaint > w.r.t. BASIC. (I should have snipped the succeeding paragraph.) Well, I thought pretty much everyone agreed with Dykstra regarding the use of GOTO's, but maybe I'm wrong. (there are other things where Dykstra and I part company so maybe some don't even agree with this.) > I > thought that you implied that BASIC tends to give you bad code because > of all the GO TO's, in contradiction to your claim about bad code > coming solely from programmers. Good programmers don't use GOTO's, even in BASIC. :-) > And can you write good code using MS- > DOS? Well, MSDOS isn't a language so I don't understand the question. But if you meant can I write good code to run under MSDOS, certainly. > > It seems to me that always blaming the programmer regardless of what > language he is stuck with is like blaming a rescue worker for failing > a rescue regardless of what equipment he has to work with (say a > fireman armed with a single fire extinguisher vs. a fire truck with > working hydrant). A smart fireman knows when he doesn't have the equipment to perform the task and he doesn't waste time, effort or his life trying. SE is the same way. One of the first parts of SE is selecgting the right tools for the job. That means choosing the right language for the task at hand. Anything less is not SE it is hacking. Because your boss won't buy another compiler is not an excuse, it just means your shop does not practice SE. You can change that or live with the consequences. But you can't blame that one language for not being the right tool for the job. I have never worked in a one language shop. > > As to appropriateness of different languages for different purposes > (as mentioned by another poster), in physics from 1983 through 1991 I > used FORTRAN to write programs to analyze data (as did most physicists > at the time -- I don't know what they use now!). Would another > language been as good or better? Of course there's code for analyzing > experimental data, and other code for performing theoretical > calculations. Att that time, Fortran was the right tool for that job. Probably still is today, although real Frotran compilers seem to be getting rare as well. I have found myself maintaining business applications written in Fortran by engineers (in their free time one summer just to keep them busy). They wrote in Fortran because it was the only language they knew. The program were abominations and I got the task of maintaining them because I came into a shop knowing Fortran and they had these legacy programs..... My first suggestion was to re-write them all in COBOL like the rest of the business applications. bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: 25 Sep 2007 12:05:26 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: despair Message-ID: <5lsbs5F9qudpU2@mid.individual.net> In article , Michael Kraemer writes: > > This is still FORTRAN turf, simply because of legacy code. > As recently as 6 months ago I saw a major Beltway-Bandit advertising for people to do Fortran at a very high-tech Navy site. They were even willing to train. I probably should have thrown my resume into the ring!! But it was a very high cost of living area and much too close to Florida. :-) bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: 25 Sep 2007 12:12:31 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: despair Message-ID: <5lsc9fF9qudpU3@mid.individual.net> In article , VAXman- @SendSpamHere.ORG writes: > In article <5lo8a3F96a8nU1@mid.individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: > {...snip...} >>The Amish would disagree with your notion regarding powertools just >>as I disagree with your notion regarding the computer languages!! >>Both of them are merely your opinion. > > Many of the younger generation Amish are moving away from those > traditions and accepting some of the ways of the outside. Many people accept MS as the epitome of computing today. What's your point? Anyone who knows the Amish and is familiar with the quality of their products would know that the young aren't leaving cause they can do a better job with modern tools. They are leaving because the outside world has forced its way into the Amish lifestyle and the young people have had their first taste of hedonism. It's a lot like the garden of eden or Pandora'a Box. What they don't have looks desirable but they don't realize that they are making a one way trip. As for the quality of their products, go visit an Amish farm sometime. Look at the quality of the house, the barn, the outbuildings. The food. The clothing. Anything they make. And then tell me you can buy better quality (not style, quality) in Walwart. And, guess what, it's all made in America, too. :-) bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: Tue, 25 Sep 2007 05:53:32 -0700 From: AEF Subject: Re: despair Message-ID: <1190724812.867435.319180@y42g2000hsy.googlegroups.com> On Sep 25, 7:53 am, b...@cs.uofs.edu (Bill Gunshannon) wrote: > In article <1190596861.683981.103...@k79g2000hse.googlegroups.com>, > AEF writes: > > > > > On Sep 23, 8:59 pm, b...@cs.uofs.edu (Bill Gunshannon) wrote: > >> In article <1190592151.788617.205...@g4g2000hsf.googlegroups.com>, > >> AEF writes: > > >> > On Sep 23, 5:10 pm, b...@cs.uofs.edu (Bill Gunshannon) wrote: > >> >> In article , > >> >> "Tom Linden" writes: > > >> >> > On Sat, 22 Sep 2007 10:35:17 -0700, Bill Gunshannon > >> >> > wrote: > > >> >> >> In article , > >> >> >> "Tom Linden" writes: > > >> >> >>> Penny-wise, pound-foolish? The development and maintenance of C code is > >> >> >>> more costly. So pay now or pay later. > > >> >> >> Sorry, no basis in fact for that. Might actually be cheaper as the pool > >> >> >> of available C Programmers is considerably higher than say the pool of > >> >> >> available Fortran, COBOL or (gasp) PL/I programmers. And if the shop has > >> >> >> reasonable coding standards, maintenance does not have to be a nightmare. > >> >> >> (I know, having worked in large shops where turnover every 1-3 years is > >> >> >> prett close to 100%.) > > >> >> >> bill > > >> >> > No, I do have a basis for that (1) 41 years of experience > > >> >> More than me, but not by much if you count the times I wandered into > >> >> the IT field for short periods of time before I took it on full-time. > > >> >> > and (2) if you > >> >> > compare > >> >> > PL/I and C programs for equivalent functionality, you will find that the C > >> >> > version > >> >> > will have 2 to 3 times as many source language statements. > > >> >> And that means? Somehow, I don't think number of statements is a strong > >> >> basis for deciding the economy of a an Information System. If it were, > >> >> there would be a lot more stuff written in BASIC. :-) And loaded with > >> >> GOTO's. > > >> >> And let's not forget that PL/I compilers are rarer than PL/I programmers > >> >> which kind of limits it's usefulness beyond some very small niches. > > >> > Weren't you the one who said bad code comes solely from bad > >> > programmers and that the language makes zero difference? Did I miss > >> > something? > > >> What does that have to do with the fact that PL/I compilers are rarer > >> than PL/I programmers? Or the number of statements it takes to write > >> a program in any given language? > > > Sorry, I was referring mostly to your excessive GO TO's complaint > > w.r.t. BASIC. (I should have snipped the succeeding paragraph.) > > Well, I thought pretty much everyone agreed with Dykstra regarding > the use of GOTO's, but maybe I'm wrong. (there are other things where > Dykstra and I part company so maybe some don't even agree with this.) You lost me here. Who's saying GOTO's are a good thing? > > thought that you implied that BASIC tends to give you bad code because > > of all the GO TO's, in contradiction to your claim about bad code > > coming solely from programmers. > > Good programmers don't use GOTO's, even in BASIC. :-) > > > And can you write good code using MS- > > DOS? > > Well, MSDOS isn't a language so I don't understand the question. But > if you meant can I write good code to run under MSDOS, certainly. Why not? You can write "batch programs" with it. That's programming. You can write "programs" in DCL, too. How are they not languages? > > It seems to me that always blaming the programmer regardless of what > > language he is stuck with is like blaming a rescue worker for failing > > a rescue regardless of what equipment he has to work with (say a > > fireman armed with a single fire extinguisher vs. a fire truck with > > working hydrant). > > A smart fireman knows when he doesn't have the equipment to perform > the task and he doesn't waste time, effort or his life trying. SE > is the same way. One of the first parts of SE is selecgting the > right tools for the job. That means choosing the right language for > the task at hand. Anything less is not SE it is hacking. Because > your boss won't buy another compiler is not an excuse, it just means > your shop does not practice SE. You can change that or live with the > consequences. But you can't blame that one language for not being the > right tool for the job. I have never worked in a one language shop. Is your hyphen-key broken or is it your fault hyphens are missing above? :-) I still think it's plausible that a bad language can make it harder to write good code and therefore a larger percentage of code written in that language (all other things being equal) will be bad. You can still blame the programmer, but there will still be more bad code. But I fully agree that you should use a language appropriate to the task! I think it's always appropriate to do what's appropriate. How could it not be?! :-) > > As to appropriateness of different languages for different purposes > > (as mentioned by another poster), in physics from 1983 through 1991 I > > used FORTRAN to write programs to analyze data (as did most physicists > > at the time -- I don't know what they use now!). Would another > > language been as good or better? Of course there's code for analyzing > > experimental data, and other code for performing theoretical > > calculations. > > Att that time, Fortran was the right tool for that job. Probably > still is today, although real Frotran compilers seem to be getting > rare as well. I have found myself maintaining business applications > written in Fortran by engineers (in their free time one summer just > to keep them busy). They wrote in Fortran because it was the only > language they knew. The program were abominations and I got the > task of maintaining them because I came into a shop knowing Fortran > and they had these legacy programs..... My first suggestion was to > re-write them all in COBOL like the rest of the business applications. OK, so it wasn't a physics "shop". Right? I can't imagine doing physics in COBOL! > > bill > > -- > Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves > b...@cs.scranton.edu | and a sheep voting on what's for dinner. > University of Scranton | > Scranton, Pennsylvania | #include AEF ------------------------------ Date: 25 Sep 2007 13:09:23 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: despair Message-ID: <5lsfk3F9qnrbU1@mid.individual.net> In article <1190724812.867435.319180@y42g2000hsy.googlegroups.com>, AEF writes: > On Sep 25, 7:53 am, b...@cs.uofs.edu (Bill Gunshannon) wrote: >> >> Well, MSDOS isn't a language so I don't understand the question. But >> if you meant can I write good code to run under MSDOS, certainly. > > Why not? You can write "batch programs" with it. No, not with MSDOS. With COMMAND.COM maybe, but not with MSDOS. MSDOS is a kernel and nothing more. It's not a language, it's an attempt at an operating system (although it is no more an OS than CP/M was.) > That's programming. > You can write "programs" in DCL, too. How are they not languages? Yes, you can write programs in DCL. But you can't write programs in VMS. See the difference? > >> > It seems to me that always blaming the programmer regardless of what >> > language he is stuck with is like blaming a rescue worker for failing >> > a rescue regardless of what equipment he has to work with (say a >> > fireman armed with a single fire extinguisher vs. a fire truck with >> > working hydrant). >> >> A smart fireman knows when he doesn't have the equipment to perform >> the task and he doesn't waste time, effort or his life trying. SE >> is the same way. One of the first parts of SE is selecgting the >> right tools for the job. That means choosing the right language for >> the task at hand. Anything less is not SE it is hacking. Because >> your boss won't buy another compiler is not an excuse, it just means >> your shop does not practice SE. You can change that or live with the >> consequences. But you can't blame that one language for not being the >> right tool for the job. I have never worked in a one language shop. > > Is your hyphen-key broken or is it your fault hyphens are missing > above? :-) What did you want hyphenated? I'll hyphenate anything you ask for. :-) But I don't see anything in the paragraph preceding that should have been. > > I still think it's plausible that a bad language can make it harder to > write good code and therefore a larger percentage of code written in > that language (all other things being equal) will be bad. You can > still blame the programmer, but there will still be more bad code. But it is only some small group of individuals who are deciding what paticular features make "bad programming languages". Do you really think that Dennis Ritchie and Brian Kernighan set out to write a bad language when they designed C? Or is it just possible they knew what they were doing and every feature of C was put there for a particular reason? > > But I fully agree that you should use a language appropriate to the > task! I think it's always appropriate to do what's appropriate. How > could it not be?! :-) > >> > As to appropriateness of different languages for different purposes >> > (as mentioned by another poster), in physics from 1983 through 1991 I >> > used FORTRAN to write programs to analyze data (as did most physicists >> > at the time -- I don't know what they use now!). Would another >> > language been as good or better? Of course there's code for analyzing >> > experimental data, and other code for performing theoretical >> > calculations. >> >> Att that time, Fortran was the right tool for that job. Probably >> still is today, although real Frotran compilers seem to be getting >> rare as well. I have found myself maintaining business applications >> written in Fortran by engineers (in their free time one summer just >> to keep them busy). They wrote in Fortran because it was the only >> language they knew. The program were abominations and I got the >> task of maintaining them because I came into a shop knowing Fortran >> and they had these legacy programs..... My first suggestion was to >> re-write them all in COBOL like the rest of the business applications. > > OK, so it wasn't a physics "shop". Right? I can't imagine doing > physics in COBOL! It was a school, actually. And they had all these engineering professors with nothing to do int he summer. So someone came up with the bright idea of giving them a program spec and having them spend their summer writting an Information System to meet the spec. What they got was typical engineering programming. A really bad piece of spaghetti Firtran with several thousand lines of code, no documentation and the writers were gone within the next two years so even if you assumed they could figure it out, they weren't there any more. I have other interesting stories about non-programmers writting Fortran, but we'll save them for another day. :-) bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: 25 Sep 2007 13:15:35 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: despair Message-ID: <5lsfvmF9qnrbU2@mid.individual.net> In article , VAXman- @SendSpamHere.ORG writes: > In article <5lsc9fF9qudpU3@mid.individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: >> >> >>In article , >> VAXman- @SendSpamHere.ORG writes: >>> In article <5lo8a3F96a8nU1@mid.individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: >>> {...snip...} >>>>The Amish would disagree with your notion regarding powertools just >>>>as I disagree with your notion regarding the computer languages!! >>>>Both of them are merely your opinion. >>> >>> Many of the younger generation Amish are moving away from those >>> traditions and accepting some of the ways of the outside. >> >>Many people accept MS as the epitome of computing today. What's your >>point? >> >>Anyone who knows the Amish and is familiar with the quality of their >>products would know that the young aren't leaving cause they can do >>a better job with modern tools. They are leaving because the outside >>world has forced its way into the Amish lifestyle and the young people >>have had their first taste of hedonism. It's a lot like the garden of >>eden or Pandora'a Box. What they don't have looks desirable but they >>don't realize that they are making a one way trip. >> >>As for the quality of their products, go visit an Amish farm sometime. >>Look at the quality of the house, the barn, the outbuildings. The food. >>The clothing. Anything they make. And then tell me you can buy better >>quality (not style, quality) in Walwart. And, guess what, it's all >>made in America, too. :-) > > Bill, don't forget where I was born and raised. My family almost moved > to the heartland Lancaster Country too. I have never had the pleasure of meeting you and have no idea where you are from. But I know many people who look at the Amish and just laugh. Sad really. I could never change (voluntarily) to that life- style, as I have spent too much time "outside" and as I said, it's like Pandora's box, once opened you can't put the demons back in, but I can easily see that there are some desirable things to it. I have, taken one step in that direction, though, by all but abandoning TV. Abd, the less I watch, the less I miss it. :-) bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: 25 Sep 2007 08:20:35 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: despair Message-ID: In article <1190724812.867435.319180@y42g2000hsy.googlegroups.com>, AEF writes: > > You lost me here. Who's saying GOTO's are a good thing? > All hail BLISS, which has no GOTO and yet provides a syntax for each of the legitimate uses of GOTO. > > OK, so it wasn't a physics "shop". Right? I can't imagine doing > physics in COBOL! No, but I've seen orbital mechanics done using a FORTRAN IV style that we refered to as COMEFROM statements. ------------------------------ Date: Tue, 25 Sep 2007 06:48:19 -0700 From: AEF Subject: Re: despair Message-ID: <1190728099.052447.35980@d55g2000hsg.googlegroups.com> On Sep 25, 9:09 am, b...@cs.uofs.edu (Bill Gunshannon) wrote: > In article <1190724812.867435.319...@y42g2000hsy.googlegroups.com>, > AEF writes: > > > On Sep 25, 7:53 am, b...@cs.uofs.edu (Bill Gunshannon) wrote: > > >> Well, MSDOS isn't a language so I don't understand the question. But > >> if you meant can I write good code to run under MSDOS, certainly. > > > Why not? You can write "batch programs" with it. > > No, not with MSDOS. With COMMAND.COM maybe, but not with MSDOS. > MSDOS is a kernel and nothing more. It's not a language, it's an > attempt at an operating system (although it is no more an OS than > CP/M was.) I always called it MS-DOS. > > > That's programming. > > You can write "programs" in DCL, too. How are they not languages? > > Yes, you can write programs in DCL. But you can't write programs in VMS. > See the difference? In this case, yes. In the Windows/DOS/COMMAND.COM/whatever-the-hell- you-call-it I'm never really sure what's going on! > >> > It seems to me that always blaming the programmer regardless of what > >> > language he is stuck with is like blaming a rescue worker for failing > >> > a rescue regardless of what equipment he has to work with (say a > >> > fireman armed with a single fire extinguisher vs. a fire truck with > >> > working hydrant). > > >> A smart fireman knows when he doesn't have the equipment to perform > >> the task and he doesn't waste time, effort or his life trying. SE > >> is the same way. One of the first parts of SE is selecgting the > >> right tools for the job. That means choosing the right language for > >> the task at hand. Anything less is not SE it is hacking. Because > >> your boss won't buy another compiler is not an excuse, it just means > >> your shop does not practice SE. You can change that or live with the > >> consequences. But you can't blame that one language for not being the > >> right tool for the job. I have never worked in a one language shop. > > > Is your hyphen-key broken or is it your fault hyphens are missing > > above? :-) > What did you want hyphenated? I'll hyphenate anything you ask for. :-) > But I don't see anything in the paragraph preceding that should have been. > one-language shop In "Anything less is not SE it is hacking" you need something between SE and it. I'm not sure whether a colon, semicolon, or dash is best, but you need *something* there. Granted it's not a hyphen. > > > > I still think it's plausible that a bad language can make it harder to > > write good code and therefore a larger percentage of code written in > > that language (all other things being equal) will be bad. You can > > still blame the programmer, but there will still be more bad code. > > But it is only some small group of individuals who are deciding what > paticular features make "bad programming languages". Do you really > think that Dennis Ritchie and Brian Kernighan set out to write a bad > language when they designed C? Or is it just possible they knew what > they were doing and every feature of C was put there for a particular > reason? No, yes, but the road to hell is paved with good intentions, no? I haven't used C much but I was commenting more on good tools vs. bad tools in general. > > > > > > But I fully agree that you should use a language appropriate to the > > task! I think it's always appropriate to do what's appropriate. How > > could it not be?! :-) > > >> > As to appropriateness of different languages for different purposes > >> > (as mentioned by another poster), in physics from 1983 through 1991 I > >> > used FORTRAN to write programs to analyze data (as did most physicists > >> > at the time -- I don't know what they use now!). Would another > >> > language been as good or better? Of course there's code for analyzing > >> > experimental data, and other code for performing theoretical > >> > calculations. > > >> Att that time, Fortran was the right tool for that job. Probably > >> still is today, although real Frotran compilers seem to be getting > >> rare as well. I have found myself maintaining business applications > >> written in Fortran by engineers (in their free time one summer just > >> to keep them busy). They wrote in Fortran because it was the only > >> language they knew. The program were abominations and I got the > >> task of maintaining them because I came into a shop knowing Fortran > >> and they had these legacy programs..... My first suggestion was to > >> re-write them all in COBOL like the rest of the business applications. > > > OK, so it wasn't a physics "shop". Right? I can't imagine doing > > physics in COBOL! > > It was a school, actually. And they had all these engineering > professors with nothing to do int he summer. So someone came > up with the bright idea of giving them a program spec and having > them spend their summer writting an Information System to meet > the spec. What they got was typical engineering programming. > A really bad piece of spaghetti Firtran with several thousand > lines of code, no documentation and the writers were gone within > the next two years so even if you assumed they could figure it > out, they weren't there any more. > > I have other interesting stories about non-programmers writting > Fortran, but we'll save them for another day. :-) > > bill > > -- > Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves > b...@cs.scranton.edu | and a sheep voting on what's for dinner. > University of Scranton | > Scranton, Pennsylvania | #include AEF ------------------------------ Date: 25 Sep 2007 14:11:06 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: despair Message-ID: <5lsj7qF9qn9dU1@mid.individual.net> In article <1190728099.052447.35980@d55g2000hsg.googlegroups.com>, AEF writes: > On Sep 25, 9:09 am, b...@cs.uofs.edu (Bill Gunshannon) wrote: >> In article <1190724812.867435.319...@y42g2000hsy.googlegroups.com>, >> AEF writes: >> >> > On Sep 25, 7:53 am, b...@cs.uofs.edu (Bill Gunshannon) wrote: >> >> >> Well, MSDOS isn't a language so I don't understand the question. But >> >> if you meant can I write good code to run under MSDOS, certainly. >> >> > Why not? You can write "batch programs" with it. >> >> No, not with MSDOS. With COMMAND.COM maybe, but not with MSDOS. >> MSDOS is a kernel and nothing more. It's not a language, it's an >> attempt at an operating system (although it is no more an OS than >> CP/M was.) > > I always called it MS-DOS. Must make it hard getting your point accross to the helpdesk when you have a problrm. :-) > >> >> > That's programming. >> > You can write "programs" in DCL, too. How are they not languages? >> >> Yes, you can write programs in DCL. But you can't write programs in VMS. >> See the difference? > > In this case, yes. In the Windows/DOS/COMMAND.COM/whatever-the-hell- > you-call-it I'm never really sure what's going on! See comment above: :-) HP: "VMS Tech Support, can I help you?" Customer: "Yes, I wrote a program in VMS and it doesn't work. What's wrong with VMS?" > >> >> > It seems to me that always blaming the programmer regardless of what >> >> > language he is stuck with is like blaming a rescue worker for failing >> >> > a rescue regardless of what equipment he has to work with (say a >> >> > fireman armed with a single fire extinguisher vs. a fire truck with >> >> > working hydrant). >> >> >> A smart fireman knows when he doesn't have the equipment to perform >> >> the task and he doesn't waste time, effort or his life trying. SE >> >> is the same way. One of the first parts of SE is selecgting the >> >> right tools for the job. That means choosing the right language for >> >> the task at hand. Anything less is not SE it is hacking. Because >> >> your boss won't buy another compiler is not an excuse, it just means >> >> your shop does not practice SE. You can change that or live with the >> >> consequences. But you can't blame that one language for not being the >> >> right tool for the job. I have never worked in a one language shop. >> >> > Is your hyphen-key broken or is it your fault hyphens are missing >> > above? :-) > >> What did you want hyphenated? I'll hyphenate anything you ask for. :-) >> But I don't see anything in the paragraph preceding that should have been. >> > one-language shop > > In "Anything less is not SE it is hacking" you need something between > SE and it. I'm not sure whether a colon, semicolon, or dash is best, > but you need *something* there. Granted it's not a hyphen. I am sure my third-grade english teacher appreciates you re-inforcing all the stuff I slept through. :-) > >> >> >> > I still think it's plausible that a bad language can make it harder to >> > write good code and therefore a larger percentage of code written in >> > that language (all other things being equal) will be bad. You can >> > still blame the programmer, but there will still be more bad code. >> >> But it is only some small group of individuals who are deciding what >> paticular features make "bad programming languages". Do you really >> think that Dennis Ritchie and Brian Kernighan set out to write a bad >> language when they designed C? Or is it just possible they knew what >> they were doing and every feature of C was put there for a particular >> reason? > > No, yes, but the road to hell is paved with good intentions, no? If you don;t know how to program in C, don't program in C. How hard to understand is that? If you can't program, find a job you can do. > > I haven't used C much but I was commenting more on good tools vs. bad > tools in general. A hammer in the hands of most people is more an implement of destruction than construction. Does that mean the hammer was badly designed? Or that we should get rid of them in favor of something more idiot proof? Or, maybe we just need to accept that everyone isn't a carpenter. Now, re-target that logic!! bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: Tue, 25 Sep 2007 07:31:57 -0700 From: AEF Subject: Re: despair Message-ID: <1190730717.571788.154430@22g2000hsm.googlegroups.com> On Sep 25, 9:48 am, AEF wrote: > On Sep 25, 9:09 am, b...@cs.uofs.edu (Bill Gunshannon) wrote: > > > In article <1190724812.867435.319...@y42g2000hsy.googlegroups.com>, > > AEF writes: > > > > On Sep 25, 7:53 am, b...@cs.uofs.edu (Bill Gunshannon) wrote: > > > >> Well, MSDOS isn't a language so I don't understand the question. But > > >> if you meant can I write good code to run under MSDOS, certainly. > > > > Why not? You can write "batch programs" with it. > > > No, not with MSDOS. With COMMAND.COM maybe, but not with MSDOS. > > MSDOS is a kernel and nothing more. It's not a language, it's an > > attempt at an operating system (although it is no more an OS than > > CP/M was.) > > I always called it MS-DOS. > > > > > > That's programming. > > > You can write "programs" in DCL, too. How are they not languages? > > > Yes, you can write programs in DCL. But you can't write programs in VMS. > > See the difference? > > In this case, yes. In the Windows/DOS/COMMAND.COM/whatever-the-hell- > you-call-it I'm never really sure what's going on! To give a better answer, I thought that the entities corresponding to DCL and VMS were both called MS-DOS. [...] > > -- > > Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves > > b...@cs.scranton.edu | and a sheep voting on what's for dinner. > > University of Scranton | > > Scranton, Pennsylvania | #include > > AEF New(!) from IDG books: MS-DOS for Dummkopfs AEF ------------------------------ Date: Tue, 25 Sep 2007 09:22:48 -0700 From: Doug Phillips Subject: Re: despair Message-ID: <1190737368.637829.213480@19g2000hsx.googlegroups.com> On Sep 25, 8:02 am, Ron Johnson wrote: > On 09/25/07 06:53, Bill Gunshannon wrote: > > > In article <1190596861.683981.103...@k79g2000hse.googlegroups.com>, > > AEF writes: > [snip] > >> Sorry, I was referring mostly to your excessive GO TO's complaint > >> w.r.t. BASIC. (I should have snipped the succeeding paragraph.) > > > Well, I thought pretty much everyone agreed with Dykstra regarding > > the use of GOTO's, but maybe I'm wrong. (there are other things where > > Dykstra and I part company so maybe some don't even agree with this.) > > You're telling DCL programmers not to use GOTO? > I'm sure we've all seen some not-simple DCL code that attempts to do so. Not just DCL, either. Some languages don't have all of the structuring tools and the *careful* use of GOTO's makes code *more* readable than trying to impose goto-less structure on a non-structured language. > >> I > >> thought that you implied that BASIC tends to give you bad code because > >> of all the GO TO's, in contradiction to your claim about bad code > >> coming solely from programmers. > > > Good programmers don't use GOTO's, even in BASIC. :-) > > Good programmers don't *need* to use GOTO in languages that have > rich control structures. > I agree with you. But goto's should be used *judiciously* in those languages that do not have such rich structure control, rather than invent clever ways to *seemingly* avoid them. Many structuring tools are goto's in disguise. Statements like "exit loop" or such are simply goto's without a label, and if the loop is 6 nesting levels in, the code can be worse than spaghetti. Setting and testing and tracking a bunch of status switches is *not* better than one "goto end_of_job" in my opinion. It's the over-use, mis-use and abuse of goto's that makes bad code; not the goto statement itself. Using modular coding techniques can often compensate for the lack of good structuring tools, but even that can be abused and the code can become harder to read than the thoughtful use of a goto with a meaningful label. Good *experienced* programmers will make the best use of the tools they have and write readable and efficient code. I'll take a good lazy programmer over a clever programmer any day. > For example, everyone thinks that COBOL code will, by necessity, > will be spaghetti. Not true. Even COBOL-77 has enough (sadly > underused) control structures to make it into macaroni, and COBOL-85 > need only be risotto. > I think it might be a pasta-for-dinner night, for some reason;-) I learned BAL before I learned COBOL or any other HLL, and having that experience I recognized what instructions were probably being generated beneath the HLL code. As languages added structure, I saw the value and used it. The first exposure I had to PASCAL showed what a structured language *could* do, but it also showed me how structure could be abused and used to create code that is even more unmaintainable than *good* unstructured code. To be able to chose whatever language I thought would be best for each particular application is a luxury I've never had. If others have been able to do that, they are very fortunate. ------------------------------ Date: Tue, 25 Sep 2007 09:36:53 -0700 From: Doug Phillips Subject: Re: despair Message-ID: <1190738213.191833.293810@50g2000hsm.googlegroups.com> On Sep 25, 11:08 am, "Richard B. Gilbert" wrote: > VAXman- wrote: > > In article <5lsfvmF9qnr...@mid.individual.net>, b...@cs.uofs.edu (Bill Gunshannon) writes: > > {...snip...} > > >>I have never had the pleasure of meeting you and have no idea where > >>you are from. But I know many people who look at the Amish and just > >>laugh. Sad really. I could never change (voluntarily) to that life- > >>style, as I have spent too much time "outside" and as I said, it's > >>like Pandora's box, once opened you can't put the demons back in, but > >>I can easily see that there are some desirable things to it. I have, > >>taken one step in that direction, though, by all but abandoning TV. > >>Abd, the less I watch, the less I miss it. :-) > > > I'm a Pennsy farm boy from the Lehigh Valley area. > > > As for TV, it's just a passing fad. I find nothing on it that I > > could not do without. I *do* laugh at the people that plan their > > lives around the airing of a show. That is just pathetic. I do > > not go to the theater either for movies. I am proud to say that > > I am probably the only person who has not seen Star Wars. > > It's just barely possible that you can still be a human being if you > haven't seen Star Wars. But if you have missed Monty Python's "Holy > Grail" and "Life of Brian" there is just no hope for you. > > I grant you that broadcast TV is hardly worth the effort required to > turn the set on but there ARE very occasional exceptions. These > exceptions will, without exception, be canceled after a season or two > but they're great while they last. I miss "Soap". "Fresno" was good. > I have both on DVD. Watching David Attenborough's "Planet Earth" series was an incredible experience. It's on DVD for those who missed it. Condemning TV because of all of the bad shows is like condemning the internet because of all of the porn-sites. There is some extremely good TV programming, both educational and entertaining, it's not that rare, and no one has to watch the garbage. ------------------------------ Date: 25 Sep 2007 16:44:30 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: despair Message-ID: <5lss7eF9rssoU1@mid.individual.net> In article , VAXman- @SendSpamHere.ORG writes: > In article <5lsfvmF9qnrbU2@mid.individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: > {...snip...} >>I have never had the pleasure of meeting you and have no idea where >>you are from. But I know many people who look at the Amish and just >>laugh. Sad really. I could never change (voluntarily) to that life- >>style, as I have spent too much time "outside" and as I said, it's >>like Pandora's box, once opened you can't put the demons back in, but >>I can easily see that there are some desirable things to it. I have, >>taken one step in that direction, though, by all but abandoning TV. >>Abd, the less I watch, the less I miss it. :-) > > I'm a Pennsy farm boy from the Lehigh Valley area. So, what did you do that got you banished to a hell like NJ? > > As for TV, it's just a passing fad. I find nothing on it that I > could not do without. I *do* laugh at the people that plan their > lives around the airing of a show. That is just pathetic. I do > not go to the theater either for movies. I am proud to say that > I am probably the only person who has not seen Star Wars. I still go to the periodiv movie, but usually only when I am on a military base. Tickets cost $3 there and I can't remember the last movie I was that was worth more than that. bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: 25 Sep 2007 17:04:07 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: despair Message-ID: <5lstc6F9qvcuU1@mid.individual.net> In article , VAXman- @SendSpamHere.ORG writes: > In article <5lss7eF9rssoU1@mid.individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: >> >> >>In article , >> VAXman- @SendSpamHere.ORG writes: >>> In article <5lsfvmF9qnrbU2@mid.individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: >>> {...snip...} >>>>I have never had the pleasure of meeting you and have no idea where >>>>you are from. But I know many people who look at the Amish and just >>>>laugh. Sad really. I could never change (voluntarily) to that life- >>>>style, as I have spent too much time "outside" and as I said, it's >>>>like Pandora's box, once opened you can't put the demons back in, but >>>>I can easily see that there are some desirable things to it. I have, >>>>taken one step in that direction, though, by all but abandoning TV. >>>>Abd, the less I watch, the less I miss it. :-) >>> >>> I'm a Pennsy farm boy from the Lehigh Valley area. >> >>So, what did you do that got you banished to a hell like NJ? > > I came to NJ for college. Sadly, nobody told me that NJ was like the > Raid 'Roach Motel' where bugs check in but they never check out. The > greatest hope I hold is to get out of this shithole before my life's > clock has made its final tick. Well, don't look at coming back this way. The NJites and NYCites have all but taken over. Eastern PA (north of Philly, at least) is little more than a suburb of NYC and the original residents are paying dearly for that "priveledge". I look forward to eventually getting a decent job offer from way south of the Mason-Dixon line and getting out of here myself. bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: Tue, 25 Sep 2007 10:16:46 -0700 From: Doug Phillips Subject: Re: despair Message-ID: <1190740606.096887.286860@o80g2000hse.googlegroups.com> On Sep 25, 11:36 am, Doug Phillips wrote: > On Sep 25, 11:08 am, "Richard B. Gilbert" > wrote: > > > > > VAXman- wrote: > > > In article <5lsfvmF9qnr...@mid.individual.net>, b...@cs.uofs.edu (Bill Gunshannon) writes: > > > {...snip...} > > > >>I have never had the pleasure of meeting you and have no idea where > > >>you are from. But I know many people who look at the Amish and just > > >>laugh. Sad really. I could never change (voluntarily) to that life- > > >>style, as I have spent too much time "outside" and as I said, it's > > >>like Pandora's box, once opened you can't put the demons back in, but > > >>I can easily see that there are some desirable things to it. I have, > > >>taken one step in that direction, though, by all but abandoning TV. > > >>Abd, the less I watch, the less I miss it. :-) > > > > I'm a Pennsy farm boy from the Lehigh Valley area. > > > > As for TV, it's just a passing fad. I find nothing on it that I > > > could not do without. I *do* laugh at the people that plan their > > > lives around the airing of a show. That is just pathetic. I do > > > not go to the theater either for movies. I am proud to say that > > > I am probably the only person who has not seen Star Wars. > > > It's just barely possible that you can still be a human being if you > > haven't seen Star Wars. But if you have missed Monty Python's "Holy > > Grail" and "Life of Brian" there is just no hope for you. > > > I grant you that broadcast TV is hardly worth the effort required to > > turn the set on but there ARE very occasional exceptions. These > > exceptions will, without exception, be canceled after a season or two > > but they're great while they last. I miss "Soap". "Fresno" was good. > > I have both on DVD. > > Watching David Attenborough's "Planet Earth" series was an incredible > experience. It's on DVD for those who missed it. Condemning TV because > of all of the bad shows is like condemning the internet because of all > of the porn-sites. There is some extremely good TV programming, both > educational and entertaining, it's not that rare, and no one has to > watch the garbage. Correction: Alastair Fothergill produced "Planet Earth" -- Attenborough narrated the BBC release and Sigourney Weaver narrated the US release. ------------------------------ Date: Tue, 25 Sep 2007 05:09:30 -0400 From: JF Mezei Subject: Expected votes for a boot node Message-ID: <3e818$46f8d051$cef8887a$9314@TEKSAVVY.COM> I have a node which is "critical" in that it serves as a boot node to a workstation, as well as serving a very important disk to 2 nodes. (as well as the system disk to the workstation). VELO - VAX (own system disk) CHAIN - ALPHA (own system disk, serves important DATA disk) BIKE - ALPHA (workstation, boots via MOP from CHAIN. (each currently has 1 vote). After a power failure, I have seen a situation where the VAX was asking for date-time and thus not far enough to contribute its vote, and this prevented the boot node from booting. And without it, the workstation failed it MOP requests and never got anywhere. AKA: all nodes stayed down because the VAX was waiting for someone to enter date-time after a power failure. So I am considering setting the boot node's EXPECTED_VOTES down to 1 (equal to the votes it has). This would allow it to boot alone, The vax and workstation would have expected_votes of 2 ensuring they must see at least one other node before booting. however, this leads to the possibility of a partitioned cluster. CHAIN crashes, and its ethernet is disconnected. VELO and BIKE likely hang due to IO on the disks formerly served by CHAIN, but they remain "UP" from a cluster point of view. CHAIN reboots, but with its ethernet still disconnected. Thus forming a second instance of the cluster. Ethernet is reconnected. CHAIN VELO and BIKE now see each other. QUESTION: Is it correct to assume that VELO or BIKE will tell CHAIN to perform hara-kiri ? (since they have 2 votes, thy would have the power to tell the 1 vote CHAIN to go and get lost, right ?) --------- In a case where CHAIN lives in its own partitioned cluster while BIKE/VELO are hung, could I mitigate the problem of partitioned cluster by having CHAIN not start any of its applications ? (thus no modificatiosn to the data disk accessed by all nodes). ? In other words, in SYSTARTUP_VMS, I proceed far enough to ensure disks are available to other nodes and far enough to ensure MOP requests can be served. And I then wait in a loop for at least one of the 2 other nodes to become visible to F$CSID before proceeding with the rest of the startup process. So while CHAIN would exist in its own partitioned custer, it would remain stuck in SYSTARTUP_VMS, no applications started and no users logging in. Would it be correct to assume that no damage to the data disks would occur ? (CHAIN would not attempt to modify records that were locked by another node since CHAIN's apps would not be started). ------------------------------ Date: Tue, 25 Sep 2007 09:44:45 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: Expected votes for a boot node Message-ID: In article <3e818$46f8d051$cef8887a$9314@TEKSAVVY.COM>, JF Mezei writes: > I have a node which is "critical" in that it serves as a boot node to a > workstation, as well as serving a very important disk to 2 nodes. (as > well as the system disk to the workstation). > > VELO - VAX (own system disk) > CHAIN - ALPHA (own system disk, serves important DATA disk) > BIKE - ALPHA (workstation, boots via MOP from CHAIN. > (each currently has 1 vote). Make ALL logical disks shadow sets with at least two members, and for non-system disks, don't have all the members on the same node. This is the proper way to do what you want, disks are cheap, and it will solve all of your problems. (You might want to add some extra code to the startup so that you avoid this: MOUNT command doesn't see all the members, mounts the disk, and a few seconds later the other member comes in with a MOUNT command issued from another node, resulting in a full copy. I have some well tested code if you are interested.) > After a power failure, I have seen a situation where the VAX was asking > for date-time and thus not far enough to contribute its vote, and this > prevented the boot node from booting. And without it, the workstation > failed it MOP requests and never got anywhere. AKA: all nodes stayed > down because the VAX was waiting for someone to enter date-time after a > power failure. This is probably due to the battery in the VAX being dead. Replace it, and it won't ask for date and time. (Maybe you can even override the asking even if the battery is dead, I don't know.) > So I am considering setting the boot node's EXPECTED_VOTES down to 1 > (equal to the votes it has). This would allow it to boot alone, The vax > and workstation would have expected_votes of 2 ensuring they must see at > least one other node before booting. > > however, this leads to the possibility of a partitioned cluster. Yes, a VERY BAD THING. Don't do it. ------------------------------ Date: Tue, 25 Sep 2007 12:51:18 +0200 From: Joseph Huber Subject: Re: Expected votes for a boot node Message-ID: JF Mezei wrote: > I have a node which is "critical" in that it serves as a boot node to a > workstation, as well as serving a very important disk to 2 nodes. (as > well as the system disk to the workstation). > > VELO - VAX (own system disk) > CHAIN - ALPHA (own system disk, serves important DATA disk) > BIKE - ALPHA (workstation, boots via MOP from CHAIN. > (each currently has 1 vote). > Since there are no dual pathed disks, and all depends on CHAIN to work, why do You give the other two nodes votes at all? I would set votes=0 on VELO and BIKE, and set expected=1 on all nodes. Then the cluster is operational only if CHAIN is up. (I assume common files served by CHAIN). And in case You want to operate the Vax alone, boot it temporarily with vaxcluster=0. -- Joseph Huber - http://www.huber-joseph.de ------------------------------ Date: Tue, 25 Sep 2007 07:59:25 -0400 From: "Richard B. Gilbert" Subject: Re: Expected votes for a boot node Message-ID: <46F8F81D.20905@comcast.net> Phillip Helbig---remove CLOTHES to reply wrote: > In article <3e818$46f8d051$cef8887a$9314@TEKSAVVY.COM>, JF Mezei > writes: > > >>I have a node which is "critical" in that it serves as a boot node to a >>workstation, as well as serving a very important disk to 2 nodes. (as >>well as the system disk to the workstation). >> >>VELO - VAX (own system disk) >>CHAIN - ALPHA (own system disk, serves important DATA disk) >>BIKE - ALPHA (workstation, boots via MOP from CHAIN. >>(each currently has 1 vote). > > > Make ALL logical disks shadow sets with at least two members, and for > non-system disks, don't have all the members on the same node. This is > the proper way to do what you want, disks are cheap, and it will solve > all of your problems. If you think you can solve ALL of JF's problems, I salute you!!!! ;-) ------------------------------ Date: Tue, 25 Sep 2007 05:08:42 -0700 From: FrankS Subject: Re: Expected votes for a boot node Message-ID: <1190722122.778106.303010@n39g2000hsh.googlegroups.com> On Sep 25, 5:09 am, JF Mezei wrote: > I have a node which is "critical" in that it serves as a boot node to a > workstation, as well as serving a very important disk to 2 nodes. (as > well as the system disk to the workstation). > > VELO - VAX (own system disk) > CHAIN - ALPHA (own system disk, serves important DATA disk) > BIKE - ALPHA (workstation, boots via MOP from CHAIN. > (each currently has 1 vote). Two other options (in addition to Joseph Huber's solution): a) Set VOTES=2 on CHAIN and leave VOTES=1 on the other nodes. Set EXPECTED_VOTES=4 on all nodes. Quorom will be calculated at 3, which means VELO and BIKE could not form a cluster on their own. CHAIN would have to be running with at least one of the other nodes in order to make quorum. b) Set VOTES=2 on CHAIN, VOTES=1 on VELO, and VOTES=0 on BIKE. Set EXPECTED_VOTES=3 on all nodes. Quorum will be calculated at 2. That allows CHAIN to boot without either of the other two nodes being ready. I vaguely remember something about not letting LAVC satellite nodes have a vote, so this would meet that requirement with respect to BIKE. I'm not sure I understand your concern about CHAIN having access to the data drive after a cluster transition. If something happens to VELO or BIKE and it doesn't rejoin the cluster within the timeout window then why wouldn't you want VMS and/or RMS to clean things up and let CHAIN get back to work? Besides, doing something in SYSTARTUP_VMS wouldn't handle the case where all three nodes were running and BIKE or VELO disappears (since presumably startup would have long since been completed). ------------------------------ Date: 25 Sep 2007 08:05:48 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: Expected votes for a boot node Message-ID: In article <3e818$46f8d051$cef8887a$9314@TEKSAVVY.COM>, JF Mezei writes: > > So I am considering setting the boot node's EXPECTED_VOTES down to 1 > (equal to the votes it has). This would allow it to boot alone, The vax > and workstation would have expected_votes of 2 ensuring they must see at > least one other node before booting. I would set CHAIN's VOTES and expected votes to 1 and set the other two sysytems' VOTES to 0. That way the critical node is correctly identified and no fragmented cluster can form. On the rare occaision when I wanted to run the VAX without the Alphas I would do a conversational boot. I would also look at why the VAX was waiting for the time prompt. Mine does because the battery is dead again, but then I have to be there to boot it anyhow (I don't keep my hobbyist cluster up all the time, and it can't remember the language or boot device without the battery). I replaced the battery once, it lasted about two years. If you don't have time limits in your licenses (hobbyist licenses do) and you can survive some temporary mis-time-stamping, you can set VMS not to prompt for the time and then use a time server to fixup the time. ------------------------------ Date: Tue, 25 Sep 2007 06:17:21 -0700 From: FrankS Subject: Re: Expected votes for a boot node Message-ID: <1190726241.609010.88410@k79g2000hse.googlegroups.com> On Sep 25, 9:10 am, FrankS wrote: > After thinking about this for a little while (it's amazing what a nice > brisk walk can do for you) I came up with this simple solution: Oops. I forgot about the issue with the VAX not booting. Richard Gilbert may be on to something: We can't solve all your problems. (Although I mean only related to this one set of questions.) :-) ------------------------------ Date: Tue, 25 Sep 2007 19:22:49 +0200 From: "Gorazd Kikelj" Subject: Re: Expected votes for a boot node Message-ID: > > VELO - VAX (own system disk) > CHAIN - ALPHA (own system disk, serves important DATA disk) > BIKE - ALPHA (workstation, boots via MOP from CHAIN. > (each currently has 1 vote). > You already find out, that CHAIN IS THE MOST important system in the cluster. No other node can work without it. If you add a third system disk (BIKE get it's own system disk), then yes, give a 1 vote to each system and set EXPECTED_VOTES to 3. But in your situation, if CHAIN is down, cluster is down. So there is no point to put votes on VELO or BIKE. Set CHAIN 1 vote and others 0 votes. Put EXPECTED_VOTES to 1 and all will be well. If CHAIN goes down, cluster stall and wait for it. When it return, processing contiues. If one or both other nodes goes down, cluster is operational. No split clusters... So EXPECTED_VOTES = 1 VELO = 0 CHAIN = 1 BIKE = 0 If you want to have more resistant configuration, you'll need few more disk drives. One system disk for BIKE, One data disk for BIKE One data disk for VELO (optional) and then shadow DATA disk across the cluster. Then you can play with votes. In current situation you are without data if CHAIN is down. Best, Gorazd ------------------------------ Date: Tue, 25 Sep 2007 13:24:47 -0400 From: JF Mezei Subject: Re: Expected votes for a boot node Message-ID: FrankS wrote: > Besides, doing something in SYSTARTUP_VMS wouldn't handle the case > where all three nodes were running and BIKE or VELO disappears (since > presumably startup would have long since been completed). It is my understanding that EXPECTED_VOTES is used during early boot and that once the node has joined a cluster, normal quorum calculations happen and EXPECTED_VOTES isn't used anymore for quorum calculation. Once into the CLUSTER, CHAIN will have seen 3 votes total and will want a quorum of 2. Loss of VELO/BIKE will result in CHAIN being below quorum and stalling. Yes, BIKE at the moment is a satellite node, but I still want to give it a vote so it can help VELO stay up if CHAIN disapears temporarily and BIKE to continue to work its DECTERMS as long as they don't need to write/read to one of CHAIN's disks. AKA: the applications may stall due to disks being in mount verify, but I can still continue to use DECTERMS that have already started, and I can still do local stuff on VELO (such as access serial ports, telnet to router/switche etc). ------------------------------ Date: Tue, 25 Sep 2007 02:41:33 -0700 From: free Subject: Free software Message-ID: <1190713293.159812.51940@k79g2000hse.googlegroups.com> Click to download Free SMS Software in India http://www.cannyinfotech.com/mobile.html Your Friend ------------------------------ Date: Tue, 25 Sep 2007 01:13:11 -0700 From: "e.expelliarmus" Subject: hacking,anti-hacking,registry tweaks,compter tricks Message-ID: <1190707991.592552.157900@w3g2000hsg.googlegroups.com> well when i was in 10th me and my gang ruled the school.... i used to hack anybody who comes in my way and send proposel to anyone i get..... check this out buddies. kool website for: * hacking and anti hacking tricks * anti hackng tricks. * registry tweaks * orkut tricks * small virus * computer tricks and loads of different tricks... www.realm-of-tricks.blogspot.com www.registrydecoded.blogspot.com ------------------------------ Date: 25 Sep 2007 07:57:36 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: Impact VMS version on CPU speed Message-ID: In article <1190689803.195093.64270@g4g2000hsf.googlegroups.com>, Neil Rieck writes: > I don't have access to a Pascal complier so have nothing to add to > this thread other than this: I have done ad-hoc bench-marking in BASIC > and believe that every release of OpenVMS is faster than the previous > version. But here are two really important points: VMS 4.0 was, for us, obviously slower than 3.7. But the Fortran compiler's then new global optimization made up for it so our applications continued to run at acceptable speeds. The slowdown was in the I/O subsystem which had clustering layers added and impacted even non-clustered systems. Before I recompiled all our applications I was worried that our 11/780's might not be able to handle our applications under 4.0. Since we had a great number of UNIBUS and MASSBUS attached custom peripherals our possible upgrade paths were limitted. ------------------------------ Date: 25 Sep 2007 07:51:14 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: UK Based Alpha Station For Sale. Message-ID: In article <46f893bd$0$26349$4c368faf@roadrunner.com>, "Schroeder, AJ" writes: > > And I'm in the USA as well. That almost narrows it down to a few days drive each way! 8-) ------------------------------ Date: Tue, 25 Sep 2007 09:09:09 -0500 From: "Schroeder, AJ" Subject: Re: UK Based Alpha Station For Sale. Message-ID: Bob Koehler wrote: > In article <46f893bd$0$26349$4c368faf@roadrunner.com>, "Schroeder, > AJ" writes: >> >> And I'm in the USA as well. > > That almost narrows it down to a few days drive each way! 8-) lol... Sorry - it was late. I am in the Southeastern Wisconsin area... specifically Milwaukee. ------------------------------ End of INFO-VAX 2007.524 ************************