Re: Re: zLinux/s390 Linux on Hercules

Some of what you say is true, but I don't think you understand just what PL360 is: It's an assembler-level language with the syntax (mostly) of Algol-60. As a result, there's mstly a one-to-one correspondence between constructs such as, e.g., R0 := R0-R0; which merely puts out the machine code for SR  0,0. Now, as to "real" code, Mike Green wrote an entire OS for the 360 in the late 1960s, fool thing is 34K, has automatic/automagic overlays of program segments, etc. etc. Dr. Robert McClure used the fire out of the thing generating automatic tests, where possible, of digital circuitry, which is certainly non-trivial work.

An interesting note about the language: I had occasion to write some code with a big "CASE" statement. I looked at the resulting  object code and discovered that the compiler had generated, as expected, a branch table; what blew me away was the fact that the branch table was half the size of what I would have generated in ALC. 'nuff said.From: hercules@...
  To: hercules-390@...
  Sent: Friday, February 11, 2011 5:30 AM
  Subject: Re: [hercules-390] Re: zLinux/s390 Linux on Hercules


   
  > Oh, for crying out loud! Please go back and read Prof. Niklaus Wirth and
  > associates' paper in the Journal of the ACM aboit why ALC is such a
  > piece of s... and why PL360 was invented.

  Wirth had a nefarious agenda and couldn't live with the realities of
  production code or solving real-world problems. Look at Pascal, it's a
  one-pass compiler designed for academia. Sure he's brilliant on paper but
  what has he *delivered* Not much but conflict and generations of zombie
  "coders" who think GOTO is evil and can't live without do while. Har de har.

  The IBM system IS assembler. The hardware and software are tightly coupled
  and all of the system interface is assembler (or PL/AS) only. I have seen
  most of the PL/AS variants and worked on a project using PL/X and the only
  thing I saw was it's harder to get people to code in PL/X than it is to find
  assembler programmers. I didn't see the advantage to PL/X but I'm sure if
  you're a big company who wants to make sure your guys don't have exportable
  skills (ie reverse job security) then PL/X is great.

  You can't say OS/360...z/OS aren't among the best OS (IMHO they are the best
  OS) ever written and they're entirely in assembler and PL/AS, PL/X.
  PLanything after all is *only* an HLA is sheep's clothing. Don't have a cow,
  dude!

  > Once I was exposed to some REAL code in PL360, I swore off that piece of
  > s... forever.

  Am I the only one who has no idea what this sentence means given your next
  sentence?

  > I don't have a truly reliable benchmark on compiler speeds, but the PL360
  > compiler appears to be forty to fifty times faster than asssembler F or H,

  Yawn. That was then, this is now. Nobody codes in PL360, not even IBM. And
  nobody has access to PL/X anymore (that I know of) now that IBM turned off
  the tap on the PL/X toolkit for vendors. So if you wanna write systems
  software on the best platform and OS in the world you either learn assembler
  and code like a man or get off the pot.

  > doesn't know what to with more than 64K for compilations unless one
  > requests a cross-reference listing, and the bloody source code is
  > readable evem by someone who knows little or nothing about S/3x0.

  So what? You're still living in 1970. Plus you don't seem to be able to tell
  the difference between a language and its implementation. Anyway arguing
  against assembler in an IBM mainframe list is going to get you invited out
  of the house post haste.

  Don't let the door hit ya in the ass on the way out!

  !!! A S S E M B L E R R U L E S !!!


 

[Non-text portions of this message have been removed]