- COFF - www.tuhs.org

Re: [COFF] [TUHS] "7th Edition UNIX TM?? Summary" => "â??7th Edition UNIX â?? Summaryâ => â?????7th Edition UNIX â?????? Summaryâ???

by Greg 'groggy' Lehey

[redirecting to COFF] On Wednesday, 15 April 2020 at 18:19:57 +1000, Dave Horsfall wrote: > On Wed, 15 Apr 2020, Don Hopkins wrote: > >> I love how in a discussion of how difficult it was to publish a book on >> Unix with the correct punctuation characters 42 years ago, we still >> can???t even quote the title of the book in a discussion about Unix >> without the punctuation characters degrading and mutating each round >> trip. > > Well, I'm not the one here using Windoze... Arguably Microsoft does it better than Unix. Most of the issues are related to the character encoding. And as Don's headers say: X-Mailer: Apple Mail (2.3608.60.0.2.5) I agree with Don. I use mutt, which has many advantages, but sane character encoding isn't one of them. Greg -- Sent from my desktop computer. Finger grog(a)lemis.com for PGP public key. See complete headers for address and phone numbers. This message is digitally signed. If your Microsoft mail program reports problems, please read http://lemis.com/broken-MUA

5 years, 2 months

1
0
0 0

Re: [COFF] [TUHS] Plan 9 from outer space ?

by Adam Thornton

So it could be the lack of televised sports getting to me in these shelter-in-place days, but, I mean, sure, I guess I’ll throw in some bucks for a pay-per-view of a Pike/Thompson cage match. FIGHT! Followups set. > On Apr 18, 2020, at 6:28 PM, Rob Pike <robpike(a)gmail.com> wrote: > It wasn't my intention. > On Sun, Apr 19, 2020 at 11:12 AM Ken Thompson <ken(a)google.com> wrote: >> >> you shouldn't have shut down this discussion. >> On Sat, Apr 18, 2020 at 3:27 PM Rob Pike <robpike(a)gmail.com> wrote: >>> ``because''.

5 years, 2 months

2
1
0 0

COBOL.

by Dan Cross

So I imagine that most readers of this list have heard that a number of US states are actively looking for COBOL programmers. If you have not, the background is that, in the US, a number of unemployment insurance systems have mainframe backends running applications mostly written in COBOL. Due to the economic downturn as a result of COVID-19, these systems are being overwhelmed with unprecedented numbers of newly-unemployed people filing claims. The situation is so dire that the Governor of New Jersey mentioned it during a press conference. This has led to a number of articles in the popular press about this situation, some rather sensational: "60 year old programming language prevents people filing for unemployment!" E.g., https://www.cnn.com/2020/04/08/business/coronavirus-cobol-programmers-new-j… On the other hand, some are rather more measured: https://spectrum.ieee.org/tech-talk/computing/software/cobol-programmers-an… I suspect the real problem is less "COBOL and mainframes" and more organizational along the lines of lack of investment in training, maintenance and modernization. I can't imagine that bureaucrats are particularly motivated to invest in technology that mostly "just works." But the news coverage has led to a predictable set of rebuttals from the mainframe faithful on places like Twitter; they point out that COBOL has been updated by recent standards in 2002 and 2014 and is being unfairly blamed for the present failures, which arguably have more to do with organizational issues than technology. However, the pendulum seems to have swung too far with their arguments in that they're now asserting that COBOL codebases are uniformly masterworks. I don't buy that. I find all of this interesting. I don't know COBOL, nor all that much about it, save for some generalities about its origin and Grace Hopper's involvement in its creation. However, in the last few days I've read up on it a bit and see very little to recommend it: the type and scoping rules are a mess, things like the 'ALTER' statement and the ability to cascade procedure invocations via the 'THRU' keyword seem like a recipe for spaghetti code, and while they added an object system in 2002, it doesn't seem to integrate with the rest of the language coherently and I don't see it doing anything that can't be done in any other OO language. And of course the syntax is abjectly horrible. All in all, it may not be the cause of the current problems, but I don't know why anyone would be much of a fan of it and unless you're already sitting on a mountain of COBOL code (which, in fairness, many organizations in government, insurance and finance are...) I wouldn't invest in it. I read an estimate somewhere that there are something like 380 billion lines of COBOL out there, and another 5 billion are written annually (mostly by body shops in the BRIC countries?). That's a lot of code; surely not all of it is good. So....What do folks think? Is COBOL being unfairly maligned simply due to age, or is it really a problem? How about continued reliance on IBM mainframes: strategic assets or mistakes? - Dan C. (PS: I did look up the specs for the IBM z15. It's an impressive machine, but without an existing mainframe investment, I wouldn't get one.)

5 years, 2 months

10
13
0 0

OT BitKeeper: Creating a remote SSH repo

by Angel M Alganza

Hello: Im apologize for the OT but Im desperately looking for the answer to a question regarding BitKeeper and, since I heard about its vitues and also about it becoming open source in this list, I thought you guys wouldn't mind helping me find the answer to my problem. I've been using Git for some time now, but after hearing here about BK, Im trying it out to see if I can change to it. At the moment I'm quite happy with it and I very much like it's smaller than Git. In Git I use a remote SSH "original" repo instead of a public (HTTP) service, since I want to have distributed copies (several workstations and laptops) and a "non interactive" copy on a server of mine. I'm not sure I am making a lot of sense, since I'm not very expert on Git either (I just use a few basic functions) and English isn't my first language. What I actually need/want is to create a remote SSH repo from my local (working) repo. I haven't found how to do that neither in the documentation nor the Users Forum. Alas I havent found a mailing list to subscribe to and ask for help. That's why I'm asking you guys for help. Again sorry for the OT and thank you so much for reading until here. :) Cheers, Ángel

5 years, 2 months

4
11
0 0

On having a slash

by Dave Horsfall

Way back then, when dinosaurs strode the earth and System/360 reigned supreme, we were taught to slash our zeros and sevens (can't quite find the glyphs for them right now) in order to distinguish them from Oscars and Ones for the benefit of the keypunch girls (yes, really; I had the hots for one of them at one time, but someone else took her) on those green sheets. In the time-mean I also saw a slash through "Z" (Zulu, Zed, Zee) in order to distinguish it from a "2" (FIGURES TWO); WTF? And you *never* slashed your ones, lest thou ended up with the contents of an 029 chad-box over thine head... These days, I merely slash my sevens by habit and it doesn't even raise an eyebrow; even Oz Pigeon Post's mail mangler seems to accept it[*]. What're other bods' experiences? [*] Don't even mention Redfern, OK? Just don't... And as for getting off at Redfern, well, I did a few times when I had a contract job around there. -- Dave, bored at home

5 years, 2 months

15
20
0 0

Re: [COFF] On having a slash

by Rudi Blom

Remember I learned slashing my zeroes, small bar in sevens and zets, curly letter X. This was when a studied Mathematics fro a while at the University of Technology in Delft, the Netherlands.. I think it had to do with avoiding misunderstandings in scribbled formulars in a time we still did a lot of real writing semester papers. This was around mid-70tish for me.

5 years, 3 months

1
0
0 0

Discord chat in about 13 hours?

by Warren Toomey

Anybody feel like a text/voice chat on the ClassicCmp Discord server in about 13 hours, say 2200 UTC? #coff and the General voice channel. I'll pop on for an hour but start whenever you feel like. Cheers, Warren -- Sent from my Android phone with K-9 Mail. Please excuse my brevity.

5 years, 3 months

2
2
0 0

Happy birthday, ARPAnet!

by Dave Horsfall

RFC 1 "Host Software" was published on this day in 1969 by Steve Crocker; it described the IMP (the forerunner of today's routers). -- Dave

5 years, 3 months

2
2
0 0

Discord channels for COFF/TUHS

by Warren Toomey

All, Silent700 has graciously set up channels underneath the ClassicCMP Discord server for COFF and TUHS: COFF: https://discordapp.com/channels/642419539946110997/675860587355308049 TUHS: https://discordapp.com/channels/642419539946110997/675860622738718775 I thought it made sense to have the channels on a server where there was already old computer people (old for either or both :-). Cheers, Warren

5 years, 3 months

2
3
0 0

What sort of "hard drive" is this?

by Arthur Krewat

Doesn't ring any bells for me, so probably not DEC up to a certain point. https://www.surplusshed.com/pages/item/R1111.html art k.

5 years, 3 months

1
0
0 0

Re: [COFF] [TUHS] The most surprising Unix programs

by Dave Horsfall

On Thu, 19 Mar 2020, Mike Markowski wrote: >> I've been using my trusty HP-42S for so long that I can hardly remember >> how to use a "normal" calculator :-) > > When my classmate's calculator died during an engineering exam, he asked > if he could borrow my spare. I handed him my HP 32s and after a minute > he whispered, "Where's the equals key?" He gave my calculator back. > :-) I did that to a financial controller in a previous life; she was not amused... Hey, it was the only calculator that I had! I could see her helplessly looking for the "=" key, then I took pity on her. -- Dave

5 years, 3 months

7
7
0 0

Re: [COFF] [TUHS] The most surprising Unix programs

by Grant Taylor

+COFF On 3/20/20 8:03 AM, Noel Chiappa wrote: > Maybe I'm being clueless/over-asking, but to me it's appalling that > any college student (at least all who have _any_ math requirement at > all; not sure how many that is) doesn't know how an RPN calculator > works. I'm sure that there are some people, maybe not the corpus you mention, that have zero clue how an RPN calculator works. But I would expect anybody with a little gumption to be able to poke a few buttons and probably figure out the basic operation, or, ask if they are genuinely confused. > It's not exactly rocket science, and any reasonably intelligent > high-schooler should get it extremely quickly; just tell them it's > just a representational thing, number number operator instead of > number operator number. I agree that RPN is not rocket science. And for basic single operation equations, I think that it's largely interchangeable with infix notation. However, my experience is, as the number of operations goes up, RPN can become more difficult to use. This is likely a mental shortcoming on my part. But it is something that does take tractable mental effort for me to do. For example, let's start with Pythagorean Theorem a² + b² = c² This is relatively easy to enter in infix notation on a typical scientific calculator. However, I have to stop and think about how to enter this on an RPN calculator. I'll take a swing at this, but I might get it wrong, and I don't have anything handy to test at the moment. [a] [enter] [a] [enter] [multiply] [b] [enter] [b] [enter] [multiply] [add] [square root] # to solve for c (12 keys) Conversely infix notation for comparison. [a] [square] [plus] [b] [square] [square root] (6 keys) As I type this, I realize that I'm using higher order operations (square) in infix than I am in RPN. But that probably speaks to my ignorance of RPN. I also realize that this equation does a poor job at demonstrating what I'm trying to convey. — Or perhaps what I'm trying to convey is incorrect. — I had to arrange sub-different parts of the equation so that their results ended up together on the stack for them to be the targets of the operation. I believe this (re)arrangement of the equation is where most of my mental load / objection comes from with RPN. I feel like I have to process the equation before I can tell the calculator to compute the result for me. I don't feel like I have this burden with infix notation. Aside: I firmly believe that computers are supposed to do our bidding, not the other way around. s/computers/calculators/ > I know it's not a key intellectual skill, but it does seem to me to > be part of comon intellectual heritage that everyone should know, > like musical scales or poetry rhyming. Have you ever considered > taking two minutes (literally!) to cover it briefly, just 'someone > tried to borrow my RPN calculator, here's the basic idea of how they > work'? I'm confident that 80% of people, more of the corpus you describe, could use an RPN calculator to do simple equations. But I would not be surprised if many found that the re-arrangement of equations to being RPN friendly would simply forego the RPN calculator for simpler arithmetic operations. I think some of it is a mental question: Which has more mental load, doing the annoying arithmetic or re-arranging to use RPN. I believe that for the simpler of the arithmetic operations, RPN is going to be more difficult. All of this being said, I'd love to have someone lay out points and / or counterpoints to my understanding. -- Grant. . . . unix || die

5 years, 3 months

4
3
0 0

Re: [COFF] [TUHS] The most surprising Unix programs

by Clem Cole

Moving to COFF ... On Fri, Mar 20, 2020 at 1:24 PM Grant Taylor via TUHS <tuhs(a)minnie.tuhs.org> wrote: > Would you humor me with an example of what you mean by "thinking on the > fly"? Either I'm not understanding you or we think differently. > I'll take a stab at it in a minute. But first, I never cared either way. In college, I had an SR50 and my GF had an HP45. I would say, between my EE friends we were probably split 50/50 between TI and HP. Generally, it was the RPN centric crew were fiercely loyal as in the editor wars but would grab whichever was near me when we all were working a problem set; but I knew a couple of folks that hated RPN too. It's possible, because of my undiagnosed dyslexia at the time, but I would grab the closest calculator, pause to see which is was and then start entering things as needed. But like Jon -- if I had the TI in my hands, I found myself copying the equation. I was trying to pay attention to what button I was pressing to check for any keystroke entry errors. Both types had all of the same math functions so there was little difference in the number of strokes, other than not needing parentheses on HP and how you entered the calculation. With the HP, I was more aware of that equation I was calculating because I was having to make sure I entered it in the proper order so I could get the right answer. In my case, I was probably a tad more careful because I was being forced to thinking in terms of precedence - but I was thinking about the equation. Whereas with the TI I was just hitting the button per the equation on the paper. I typed a tad faster on the TI than the HP because I was not thinking as much but ... I probably made more typing errors there because I thought less about what I was doing. Clem

5 years, 3 months

5
8
0 0

Re: [COFF] [TUHS] The most surprising Unix programs

by Grant Taylor

Aside: I'm sending this reply to TUHS where the message that I'm replying to came from. But i suspect that it should migrate to COFF, which I'm CCing. On 3/20/20 5:48 AM, paul(a)guertin.net wrote: > I teach math in college, and I use an RPN calculator as well (it's > just easier). Would you please elaborate on "it's just easier"? I'm asking from a point of genuine curiosity. I've heard many say that RPN is easier, or that it takes fewer keys, or otherwise superior to infix notation. But many of the conversations end up somewhat devolving into religious like comments about preferences, despite starting with honest open-minded intentions. (I hope this one doesn't similarly devolve.) I've heard that there are fewer keys to press for RPN, but the example equations presented have been effectively he same. I've heard that RPN is mentally easier. But I apparently don't know enough RPN to be able to think in RPN natively to evaluate myself. I dabble with RPN, including keeping my main calculator app on my smart phone in RPN mode. So I am genuinely interested in understanding why you say that RPN is just easier. > Sometimes, during an exam, a student who forgot to bring their > calculator will ask if they can borrow mine. I always say "sure, but > you'll regret it" and hand them the calculator. After wasting one or > two minutes, they give it back. ~chuckle~ > (Note that I always make sure no calculator is needed for my exams, > but it's department policy to authorise non programmable calculators, > and it seems to reassure students to have the calculator on the desk, > so I don't mind.) > ACK -- Grant. . . . unix || die

5 years, 3 months

1
0
0 0

Old non-Unix software and manuals.

by Peter Jeremy

As many of you may be aware, Bruce D. Evans <bde(a)freebsd.org> died in mid-December. I am currently looking through his digital estate on behalf of his family and the FreeBSD Project. I have discovered that he kept an extensive collection of 5¼" floppy disks. I haven't looked through them but they appear to include things like OS-9 and Hitachi Peach files (and presumably Minix stuff, though I haven't found any of his Minix work). He also has a selection of newletters from an Australian Peach users group. Is there any interest in this material from a historicial perspective? -- Peter Jeremy

5 years, 3 months

10
12
0 0

Popular Programming languages over time

by David Barto

https://www.youtube.com/watch?v=Og847HVwRSI It’s kind of amazing how long Pascal hangs on to #1. And you can watch the ascent of web programming. David

5 years, 3 months

7
7
0 0

Speaking of floppy controllers,

by Grant Taylor

On 3/8/20 9:39 PM, Warner Losh wrote: > floppy controller supports the full range of crazy that once roamed > the earth Does anyone have any knee jerk reaction to the idea of putting a 5¼" floppy drive on a USB-to-Floppy (nominally 3½") adapter? Do I want to avoid tilting at this windmill? Am I better off installing the 5¼" floppy inside the computer and connecting directly to the motherboard? I'm only wanting to pull files off of 5¼" disks. At most I'll want to dd the disks to an image. That being said, I wonder if I should also be collecting any different types of images. (This may mean mobo instead of USB.) Thank you for any pro-tips that you can provide. -- Grant. . . . unix || die

5 years, 3 months

4
6
0 0

Re: [COFF] [TUHS] The most surprising Unix programs

by Clem Cole

Moving to COFF .... On Tue, Mar 17, 2020 at 10:58 AM Larry McVoy <lm(a)mcvoy.com> wrote: > As much as I don't care for Forth, man do I wish it had become the standard > for boot proms, it might not be my cup of tea but I could make it do what > I needed it to do. Amen bro... Sun did a nice job on that. Although the Alpha Boot ROMs were pretty good too. At least they were UNIX like and were extensible like the Sun boot ROMs. HP's were better than a PC BIOS, but they were pretty awful. > Can't say the same for UEFI, I disable that crap. > Well, it beats the crap out of IBM's BIOS, but that bar is very low. UEFI was sort of a 'camel' (a horse designed by committee) and too many people peed on it. Intel created EFI to try to fix BIOS and then people went nuts. Apple's version is the best of them, but as you say, they all suck if you have seen anything better. A big problem IMO is that EFI tried to be somewhat compatible. In the end, they were not, so you got the worst of both (new interfaces and legacy functionality). Server systems that support IPMT have Minux under the covers in coprocessor, which using a coprocessor is also how Apple runs UEFI. With IMPT, it is sort of sad more of it is not really exposed, but you need the added cost of the coprocessor. Plus it adds a new security domain, which many people complain about. I try to know as little about it as possible to get my work done, but exposing more of that interface might help.

5 years, 3 months

2
1
0 0

21st Century Equivalent to 'learn'?

by Warren Toomey

Hi all, I'm looking for an interactive tool to help students learn the Unix/Linux command line. I still remember the "learn" tool. Is there an equivalent for current systems? I have tried to forward-port the old learn sources to current Linux but my patience ran out :-) Thanks in advance for any tips/pointers. Cheers, Warren

5 years, 4 months

12
15
0 0

Fwd: [ih] NCP and TCP implementations

by Clem Cole

Given the recent discussion of pipes and networking ... I'm passing this along for those that might not have seen it. ---------- Forwarded message --------- From: Jack Haverty via Internet-history Date: Tue, Mar 10, 2020 at 1:30 PM Subject: Re: [ih] NCP and TCP implementations To: *Internet-History* The first TCP implementation for Unix was done in PDP-11 assembly language, running on a PDP-11/40 (with way too little memory or address space). It was built using code fragments excerpted from the LSI-11 TCP implementation provided by Jim Mathis, which ran under SRI's home-built OS. Jim's TCP was all written in PDP-11 assembler. The code was cross-compiled (assembled) on a PDP-10 Tenex system, and downloaded over a TTY line to the PDP-11/40. That was much easier and faster than doing all the implementation work on the PDP-11. The code architecture involved putting TCP itself at the user level, communicating with its "customers" using Unix InterProcess Communications facilities (Rand "Ports"). It would have been preferable to implement TCP within the Unix kernel, but there was simply not enough room due to the limited address space available on the 11/40 model. Later implementations of TCP, on larger machines with twice the address space, were done in the kernel. In addition to the Berkeley BSD work, I remember Gurwitz, Wingfield, Nemeth, and others working on TCP implementation for the PDP-11/70 and Vax. The initial Unix TCP implementation was for TCP version 2 (2.5 IIRC), as was Jim's LSI-11 code. This 2.5 implementation was one of the players in the first "TCP Bakeoff" organized by Jon Postel and carried out on a weekend at ISI before the quarterly Internet meeting. The PDP-11/40 TCP was modified extensively over the next year or so as TCP advanced through 2.5, 2.5+, 3, and eventually stabilized at TCP4 (which it seems we still have today, 40+ years later!) The Unix TCP implementation required a small addition to the Unix kernel code, to add the "await" and "capac" system calls. Those calls were necessary to enable the implementation of user-level code where the traditional Unix "pipeline" model of programming (input->process->process...->output) was inadequate for use in multi-computer programming (such as FTP, Telnet, etc., - anywhere where more than one computer was involved). The code to add those new system calls was written in C, as was almost all of the Unix OS itself. The new system calls added the functionality of "non-blocking I/O" which did not previously exist. It involved very few lines of code, since there wasn't room for very many more instructions, and even so it required finding more space by shortening many of the kernel error messages to save a few bytes here and there. Randy Rettberg and I did that work, struggling to understand how Unix kernel internals worked, since neither of us had ever worked with Unix before even as a user. We did not try to "get it right" by making significant changes to the basic Unix architecture. That came later with the Berkeley and Gurwitz efforts. The PDP-11/40 was simply too constrained to support such changes, and our mission was to get TCP support on the machine, rather than develop the OS. I think I speak authoritatively here, since I wrote and debugged that first Unix TCP code. I still have an old, yellowing listing of that first Unix TCP. FWIW, if there's interest in why certain languages were chosen, there's a very simple explanation of why the Unix implementation was done in assembler rather than C, the native language of Unix. First, Jim Mathis' code was in assembler, so it was easy to extract large chunks and paste them into the Unix assembler implementation. Second, and probably most important, was that I was very accustomed to writing assembler code and working at the processor instruction level. But I barely knew C existed, and was certainly not proficient in it, and we needed the TCP working fast for use in other projects. The choice was very pragmatic, not based at all on technical issues of languages or superiority of any architecture. /Jack Haverty On 3/9/20 11:14 PM, vinton cerf via Internet-history wrote: > Steve Kirsch asks in what languages NCP and TCP were written. > > The Stanford first TCP implementation was done in BCPL by Richard Karp. > Another version was written for PDP-11/23 by Jim Mathis but not clear in > what language. Tenex was probably done in C at BBN. Was 360 done in PL/1?? > Dave Clark did one for IBM PC (assembly language/??) > > Other recollections much appreciated. > > vint -- Internet-history mailing list Internet-history(a)elists.isoc.org https://elists.isoc.org/mailman/listinfo/internet-history

5 years, 4 months

2
4
0 0

Re: [COFF] 21st Century Equivalent to 'learn'?

by Rudi Blom

I started using 'cpio' in the 80tish and still use it, especially transferring files and complete directories between various UNIX versions like SCOUNIX 3.2V4.2, Tru64, HP-UX 11i.. The main option I use with cpio is (of course) "-c" and only occasionally "-u"

5 years, 4 months

1
0
0 0

some PDP11 stuff

by mylists＠posteo.de

Hi, while exploring the gopherspace (YES! Still existing, growing community) I found this gopher page: gopher://pdp11.tk/1 which can be reached with Lynx for example. Unfortunately I cannot evaluate the items there, but may be it is worth a look by someone knowledgeable. Cheers! mcc

5 years, 4 months

2
1
0 0

52-pin D-Sub?

by Adam Thornton

I work at an astronomy facility. I get to do some fun dumpster diving. I recently have pulled out of the trash a plugboard with a male and a female D-Sub 52 connector. 3 rows of pins, 17-18-17. I took the connectors off the board: there's nothing back there, so this thing only ever existed so you could plug the random cable you found into it and its friends to see what the cable fit. I can't find much evidence that a 52-pin D-Sub ever existed. Is this just Yet Another Physics Experiment thing where, hey, if your instrument already costs three million dollars, what's a couple of grand for machining custom connectors? Or was it once a thing? (also posting to cc-talk) Adam

5 years, 4 months

8
25
0 0

Re: [COFF] 52-pin D-Sub?

by Rudi Blom

Not UNIX, not 52-pin, but old, old and serial Your mission, should you choose to accept it, is to save data from a computer that should have died aeons ago ... Tap into the serial line – what could be simpler? Alas, the TI was smart enough to spot the absence of the rattly old beast ("the software wouldn't print without some of the seldom-used serial control lines functioning," explained Aaron) so the customer was asked to bring in the printer as well. https://www.theregister.co.uk/2020/02/24/who_me/

5 years, 4 months

1
0
0 0

Fwd: Old and Tradition was [TUHS] V9 shell

by Adam Thornton

---------- Forwarded message --------- From: Adam Thornton <athornton(a)gmail.com> Date: Tue, Feb 11, 2020 at 5:56 PM Subject: Re: [COFF] Old and Tradition was [TUHS] V9 shell To: Clem Cole <clemc(a)ccc.com> As someone working in this area right now....yeah, and that’s why traditional HPC centers do not deliver the services we want for projects like the Vera Rubin Observatory’s Legacy Survey Of Space and Time. Almost all of our scientist-facing code is Python though a lot of the performance critical stuff is implemented in C++ with Python bindings. The incumbents are great at running data centers like they did in 1993. That’s not the best fit for our current workload. It’s not generally the compute that needs to be super-fast: it’s access to arbitrary slices through really large data sets that is the interesting problem for us. That’s not to say that that lfortran isn’t cool. It really really is, and Ondrej Cestik has done amazing work in making modern FORTRAN run in a Jupyter notebook, and the implications (LLVM becomes the Imagemagick of compiled languages) are astounding. But...HPC is no longer the cutting edge. We are seeing a Kuhnian paradigm shift in action, and, sure, the old guys (and they are overwhelmingly guys) who have tenure and get the big grants will never give up FORTRAN which after all was good enough for their grandpappy and therefore good enough for them. But they will retire. Scaling out is way way cheaper than scaling up. On Tue, Feb 11, 2020 at 11:41 AM Clem Cole <clemc(a)ccc.com> wrote: > moving to COFF > > On Tue, Feb 11, 2020 at 5:00 AM Rob Pike <robpike(a)gmail.com> wrote: > >> My general mood about the current standard way of nerd working is how >> unimaginative and old-fashioned it feels. >> > ... >> >> But I'm a grumpy old man and getting far off topic. Warren should cry, >> "enough!". >> >> -rob >> > > @Rob - I hear you and I'm sure there is a solid amount of wisdom in your > words. But I caution that just, because something is old-fashioned, does > not necessarily make it wrong (much less bad). > > I ask you to take a look at the Archer statistics of code running in > production (Archer large HPC site in Europe): > http://archer.ac.uk/status/codes/ > > I think there are similar stats available for places like CERN, LRZ, and > of the US labs, but I know of these so I point to them. > > Please note that Fortran is #1 (about 80%) followed by C @ about 10%, > C++ @ 8%, Python @ 1% and all the others at 1%. > > Why is that? The math has not changed ... and open up any of those codes > and what do you see: solving systems of differential equations with linear > algebra. It's the same math my did by hand as a 'computer' in the 1950s. > > There is not 'tensor flows' or ML searches running SPARK in there. Sorry, > Google/AWS et al. Nothing 'modern' and fresh -- just solid simple science > being done by scientists who don't care about the computer or sexy new > computer languages. > > IIRC, you trained as a physicist, I think you understand their thinking. *They > care about getting their science done.* > > By the way, a related thought comes from a good friend of mine from > college who used to be the Chief Metallurgist for the US Gov (NIST in > Colorado). He's back in the private sector now (because he could not > stomach current American politics), but he made an important > observation/comment to me a couple of years ago. They have 60+ years of > metallurgical data that has and his peeps have been using with known > Fortran codes. If we gave him new versions of those analytical programs > now in your favorite new HLL - pick one - your Go (which I love), C++ > (which I loath), DPC++, Rust, Python - whatever, the scientists would have > to reconfirm previous results. They are not going to do that. It's not > economical. They 'know' how the data works, the types of errors they > have, how the programs behave* etc*. > > So to me, the bottom line is just because it's old fashioned does not make > it bad. I don't want to write an OS in Fortran-2018, but I can wrote a > system that supports code compiled with my sexy new Fortran-2018 compiler. > > That is to say, the challenge for >>me<< is to build him a new > supercomputer that can run those codes for him and not change what they are > doing and have them scale to 1M nodes *etc*.. > > _______________________________________________ > COFF mailing list > COFF(a)minnie.tuhs.org > https://minnie.tuhs.org/cgi-bin/mailman/listinfo/coff >

5 years, 4 months

14
20
0 0

2025

2024

2023

2022

2021

2020

2019

2018

COFF