Re: Anti-Ada FUD (rant)

[Per Sandberg <per.sandberg@sandat.dyndns.org>]

On Fri, 27 Sep 2013 23:21:19 -0700 (PDT)
Shark8 <onewingedshark@gmail.com> wrote:

> On Tuesday, September 24, 2013 5:18:33 PM UTC-6, Dennis Lee Bieber
> wrote:
> > On Mon, 23 Sep 2013 23:38:32 -0700 (PDT), krfkeith wrote:
> > >
> > >I love the laughable claim that WinNT is poorly designed. While
> > >some of the userland has left some to be desired, the core of the
> > >NT systems is *very* competently designed. The NT team was lead by
> > >Dave Cutler, of DEC fame. 
> > 
> > 	The internals may be good -- though I could wish for even
> > more of VMS to have been "ported"... Like (common) event flag
> > clusters, IPC via mailboxes, and maybe even the entire QIO
> > system...  Somehow the objects that WinNT+ uses for "wait" don't
> > feel natural to me, and the divorced socket I/O system show a few
> > warts (select only works on sockets, need a different API for
> > files, etc.)
> > 
> > 	But the failure to provide a decent command shell scripting
> > capability (a la DCL) saddled it. PowerShell finally makes an
> > apologetic entry (Pity my main use is a glorified "find":
> > get-childitem -path "xxx" -filter "*.ext" | select-string -pattern
> > "yyy" )
> 
> You know, I've heard a lot of good stuff about VMS (and DEC's Ada
> compiler) -- kinda makes me wish I had first-hand experience with
> them. (And, obviously, the R-1000, that sounds like a *very*
> interesting setup to me... and I'm not convinced that having "the
> library" [and source, and objects {.o/.obj}] stored in a database
> isn't "the right way" to do things.)

The R-1000:s where wonderful machines of their time, unfortunate they
got some drawbacks, memory and cpu.
I recall that rebuilding a system from scratch was a five week exercise
week 1)      Prepare carefully.
             All the sources in place and quad check the build scripts
             and sources.
weeks 2..4 ) Have the machine to compile then sources.
             The reason for a quad check was that this was time for
             summer leave.
week 5)      Clean up and validate the build.
             Here the quad check paid off.
The core idea of the R-1000:s was that the source was just a rendering
of the "model".
My guess is that if the common terminals of the time has been graphical
there has been a graphical view of the model as well as a source view.

/Persan.