ABC-TV "Grantchester" (a detective-priest series[*]) featured someone who
died from mercury poisoning, when someone opened the tanks. Apart from
the beautiful Teletype in the foreground, there was a machine with lots of
dials in the background; obviously some early computer, but not enough
footage for me to tell which one, but is a British show if that helps.
[*]
I suppose that I need not remind any Monty Python freaks of these lines:
"There's another dead bishop on the landing, Vicar-Sergeant!"
"Err, Detective-Parsons, Madam."
Etc... Otherwise I'd go on all day :-)
-- Dave
[ -TUHS and +COFF ]
On Mon, Jun 8, 2020 at 1:49 AM Lars Brinkhoff <lars(a)nocrew.org> wrote:
> Chris Torek wrote:
> > You pay (sometimes noticeably) for the GC, but the price is not
> > too bad in less time-critical situations. The GC has a few short
> > stop-the-world points but since Go 1.6 or so, it's pretty smooth,
> > unlike what I remember from 1980s Lisp systems. :-)
>
> I'm guessing those 1980s Lisp systems would also be pretty smooth on
> 2020s hardware.
>
I worked on a big system written in Common Lisp at one point, and it used a
pretty standard Lispy garbage collector. It spent something like 15% of its
time in GC. It was sufficiently bad that we forced a full GC between every
query.
The Go garbage collector, on the other hand, is really neat. I believe it
reserves something like 25% of available CPU time for itself, but it runs
concurrently with your program: having to stop the world for GC is very,
very rare in Go programs and even when you have to do it, the concurrent
code (which, by the way, can make use of full physical parallelism on
multicore systems) has already done much of the heavy lifting, meaning you
spend less time in a GC pause. It's a very impressive piece of engineering.
- Dan C.
> I've never heard of the dimensionality of an array in such a context
> described in terms of a sum type,[...]
I was also puzzled by this remark. Perhaps Bram was thinking of an
infinite sum type such as
Arrays_0 + Arrays_1 + Arrays_2 + ...
where "Arrays_n" is the type of arrays of size n. However, I don't see
a way of defining this without dependent (or at least indexed) types.
> [...] but have often heard of it described in terms of a dependent
> type.
I think that's right, yes. We want the type checker to decide, at
compile time, whether a certain operation on arrays, such as scalar
product, or multiplication with a matrix, will respect the types. That
means providing the information about the size to the type checker:
types need to know about values, hence the need for dependent types.
Ideally, a lot of the type information can then be erased at run time,
so that we don't need the arrays to carry their sizes around. But the
distinction between what can and what cannot be erased at run time is
muddled in a dependently-typed programming language.
Best wishes,
Cezar
Dan Cross <crossd(a)gmail.com> writes:
> [-TUHS, +COFF as per wkt's request]
>
> On Sun, Jun 7, 2020 at 8:26 PM Bram Wyllie <bramwyllie(a)gmail.com> wrote:
>
>> Dependent types aren't needed for sum types though, which is what you'd
>> normally use for an array that carries its size, correct?
>>
>
> No, I don't think so. I've never heard of the dimensionality of an array in
> such a context described in terms of a sum type, but have often heard of it
> described in terms of a dependent type. However, I'm no type theorist.
>
> - Dan C.
> _______________________________________________
> COFF mailing list
> COFF(a)minnie.tuhs.org
> https://minnie.tuhs.org/cgi-bin/mailman/listinfo/coff
[-TUHS, +COFF as per wkt's request]
On Sun, Jun 7, 2020 at 8:26 PM Bram Wyllie <bramwyllie(a)gmail.com> wrote:
> Dependent types aren't needed for sum types though, which is what you'd
> normally use for an array that carries its size, correct?
>
No, I don't think so. I've never heard of the dimensionality of an array in
such a context described in terms of a sum type, but have often heard of it
described in terms of a dependent type. However, I'm no type theorist.
- Dan C.
Moving to COFF where this discussion really belongs ...
On Sun, Jun 7, 2020 at 2:51 PM Nemo Nusquam <cym224(a)gmail.com> wrote:
> On 06/07/20 11:26, Clem Cole wrote (in part):
> > Neither language is used for anything in production in our world at this
> point.
>
> They seem to be used in some worlds: https://blog.golang.org/10years and
> https://www.rust-lang.org/production
Nemo,
That was probably not my clearest wording. I did not mean to imply either
Go or Rust was not being used by any imagination.
My point was that in SW development environments that I reside (HPC, some
startups here in New England and the Bay Area, as well as Intel in general)
-- Go and Rust both have smaller use cases compared to C and C++ (much less
Python and Java for that matter). And I know of really no 'money' project
that relies yet on either. That does not mean I know of no one using
either; as I know of projects using both (including a couple of my own),
but no one that had done anything production or deployed 'mission-critical'
SW with one or the other. Nor does that mean it has not happened, it just
means I have not seen been exposed.
I also am saying that in my own personal opinion, I expect it too, in
particular in Go in userspace code - possibly having a chance to push out
Java and hopefully pushing out C++ a bit.
My response was to an earlier comment about C's popularity WRT to C++. I
answered with my experience and I widened it to suggest that maybe C++ was
not the guaranteed incumbent as the winner for production. What I did not
say then, but I alluded too was the particularly since nothing in nearly 70
years has displaced Fortran, which >>is<< still the #1 language for
production codes (as you saw with the Archer statistics I pointed out).
Reality time ... Intel, IBM, *et al,* spend a lot of money making sure
that there are >>production quality<< Fortran compilers easily available.
Today's modern society is from Weather prediction to energy, to param,
chemistry, and physics. As I have here and in other places, over my
career, Fortran has paid me and my peeps salary. It is the production
enabler and without a solid answer to having a Fortran solution, you are
unlikely to make too much progress, certainly in the HPC space.
Let me take this in a slightly different direction. I tend to use the
'follow the money' as a way to root out what people care about.
Where firms spend money to create or purchase tools to help their staff?
The answer is in tools that give them return that they can measure. So
using that rule: What programming languages have the largest ecosystems
for tools that help find performance and operation issues? Fortran, C, C++
have the largest that I know. My guess would be Java and maybe
JavaScript/PHP would be next, but I don't know of any.
If I look at Intel, were do we spend money on the development tools: C/C++
and Fortran (which all use a common backend) are #1. Then we invest in
other versions of the same (GCC/LLVM) for particularly things we care
about. After that, it's Python and used to be Java and maybe some
JavaScript. Why? because ensuring that those ecosystems are solid on
devices that we make is good for us, even it means we help some of our
competitors' devices also. But our investment helps us and Fortran,
C/C++ is where people use our devices (and our most profitable versions in
particular), so it's our own best interest to make sure there are tools to
bring out the best.
BTW: I might suggest you take a peek at where other firms do the same
thing, and I think you'll find the follow the money rule is helpful to
understand what people care the most about.
Hi
I'm wondering if anybody knows what happened with this? (Besides the
fact it crashed and burnt. I mean, where are the source trees?)
lowendmac.com/2016/nutek-mac-clones/
It was a Macintosh clone that reverse-engineered the Mac interface and
appearance, using the X Window System as the GUI library - what in
Macs of that era was locked into the Mac system BIOS.
As such, it's an interesting use of the X Window System.
Wesley Parish
Cc: to COFF, as this isn't so Unix-y anymore.
On Tue, May 26, 2020 at 12:22 PM Christopher Browne <cbbrowne(a)gmail.com>
wrote:
> [snip]
> The Modula family seemed like the better direction; those were still
> Pascal-ish, but had nice intentional extensions so that they were not
> nearly so "impotent." I recall it being quite popular, once upon a time,
> to write code in Modula-2, and run it through a translator to mechanically
> transform it into a compatible subset of Ada for those that needed DOD
> compatibility. The Modula-2 compilers were wildly smaller and faster for
> getting the code working, you'd only run the M2A part once in a while
> (probably overnight!)
>
Wirth's languages (and books!!) are quite nice, and it always surprised and
kind of saddened me that Oberon didn't catch on more.
Of course Pascal was designed specifically for teaching. I learned it in
high school (at the time, it was the language used for the US "AP Computer
Science" course), but I was coming from C (with a little FORTRAN sprinkled
in) and found it generally annoying; I missed Modula-2, but I thought
Oberon was really slick. The default interface (which inspired Plan 9's
'acme') had this neat graphical sorting simulation: one could select
different algorithms and vertical bars of varying height were sorted into
ascending order to form a rough triangle; one could clearly see the
inefficiency of e.g. Bubble sort vs Heapsort. I seem to recall there was a
way to set up the (ordinarily randomized) initial conditions to trigger
worst-case behavior for quick.
I have a vague memory of showing it off in my high school CS class.
- Dan C.
Hi all, I have a strange question and I'm looking for pointers.
Assume that you can multiply two 8-bit values in hardware and get a 16-bit
result (e.g. ROM lookup table). It's straightforward to use this to multiply
two 16-bit values:
AABB *
CCDD
----
PPPP = BB*DD
QQQQ00 = BB*CC
RRRR00 = AA*DD
SSSS0000 = AA*CC
--------
32-bit result
But if the hardware can only provide the low eight bits of the 8-bit by
8-bit multiply, is it still possible to do a 16-bit by 16-bit multiply?
Next question, is it possible to do 16-bit division when the hardware
can only do 8-bit divided by 8-bit. Ditto 16-bit modulo with only 8-bit
modulo?
Yes, I could sit down and nut it all out from scratch, but I assume that
somewhere this has already been done and I could use the results.
Thanks in advance for any pointers.
Warren
** Back story. I'm designing an 8-bit TTL CPU which has 8-bit multiply, divide
and modulo in a ROM table. I'd like to write subroutines to do 16-bit and
32-bit integer maths.
On Sun, May 17, 2020 at 12:24 PM Paul Winalski <paul.winalski(a)gmail.com>
wrote:
> On 5/16/20, Steffen Nurpmeso <steffen(a)sdaoden.eu> wrote:
> >
> > Why was there no byte or "mem" type?
>
> These days machine architecture has settled on the 8-bit byte as the
> unit for addressing, but it wasn't always the case. The PDP-10
> addressed memory in 36-bit units. The character manipulating
> instructions could deal with a variety of different byte lengths: you
> could store six 6-bit BCD characters per machine word,
Was this perhaps a typo for 9 4-bit BCD digits? I have heard that a reason
for the 36-bit word size of computers of that era was that the main
competition at the time was against mechanical calculator, which had
9-digit precision. 9*4=36, so 9 BCD digits could fit into a single word,
for parity with the competition.
6x6-bit data would certainly hold BAUDOT data, and I thought the Univac/CDC
machines supported a 6-bit character set? Does this live on in the Unisys
1100-series machines? I see some reference to FIELDATA online.
I feel like this might be drifting into COFF territory now; Cc'ing there.
or five ASCII
> 7-bit characters (with a bit left over), or four 8-bit characters
> (ASCII plus parity, with four bits left over), or four 9-bit
> characters.
>
> Regarding a "mem" type, take a look at BLISS. The only data type that
> language has is the machine word.
>
> > +getfield(buf)
> > +char buf[];
> > +{
> > + int j;
> > + char c;
> > +
> > + j = 0;
> > + while((c = buf[j] = getc(iobuf)) >= 0)
> > + if(c==':' || c=='\n') {
> > + buf[j] =0;
> > + return(1);
> > + } else
> > + j++;
> > + return(0);
> > +}
> >
> > so here the EOF was different and char was signed 7-bit it seems.
>
> That makes perfect sense if you're dealing with ASCII, which is a
> 7-bit character set.
To bring it back slightly to Unix, when Mary Ann and I were playing around
with First Edition on the emulated PDP-7 at LCM+L during the Unix50 event
last USENIX, I have a vague recollection that the B routine for reading a
character from stdin was either `getchar` or `getc`. I had some impression
that this did some magic necessary to extract a character from half of an
18-bit word (maybe it just zeroed the upper half of a word or something).
If I had to guess, I imagine that the coincidence between "character" and
"byte" in C is a quirk of this history, as opposed to any special hidden
meaning regarding textual vs binary data, particularly since Unix makes no
real distinction between the two: files are just unstructured bags of
bytes, they're called 'char' because that was just the way things had
always been.
- Dan C.
