16 Oct
2011
16 Oct
'11
1:23 p.m.
On 10/16/2011 12:32 AM, Chuck Guzis wrote:
On 15 Oct 2011 at 22:51, Eric Smith wrote:
Not just bit-aligned data: some processors from that same broad era (e.g. iAPX432)
didn't even require that instructions be byte or word aligned.
I think the point folks are missing is that it isn't the processor that defines the
standard data types, but the language.
I used to be more condescending towards the attitude that
the-whole-world-is-a-stream-of-bytes (having worked with many I/O devices and languages
and computers that had far more evolved record concepts) but here i am in 2011 and if I
can coerce anything into a stream of bytes - I've got a lot of tools for working with
that.
Tim.
Yes. All C types, except bitfields within a
structure, must have
sizes that are a multiple of the size of the char type.
Wasn't it presumptuous of K&R to assume that the smallest native
datum was a char? At the time the spec was being written, there did
exist bit-addressable machines, so directly-addressed bit arrays were
certainly possible on some hardware. 
16 Oct
16 Oct
1:59 p.m.
On 16 Oct 2011 at 13:23, Shoppa, Tim wrote:
I used to be more condescending towards the attitude
that
the-whole-world-is-a-stream-of-bytes (having worked with many I/O
devices and languages and computers that had far more evolved record
concepts) but here i am in 2011 and if I can coerce anything into a
stream of bytes - I've got a lot of tools for working with that.
Unfortunately, our tools constrain our thought--and that's the evil
in a "universal" tool such as C.
Program in APL, for example, for six months and I guarantee that your
way of thinking about things will begin to change dramatically. I'm
not saying that "that Iverson language" (as one of my co-workers used
to call it) is the be-all, but its constraints on thinking are very
different from most traditional languages.
--Chuck
2:25 p.m.
New subject: languages and patterns of thought - Re: C trivia - was Re: Dennis Ritchie has died
On 16/10/11 1:59 PM, Chuck Guzis wrote:
On 16 Oct 2011 at 13:23, Shoppa, Tim wrote:
Absolutely!!! This is a very important point.
Alan Perlis:
"A language that doesn't affect the way you think about programming,
is not worth knowing."
This should be stenciled on the door of every programming office.
Language constrains and moulds thought - as George Orwell also knew and
wrote about.
via http://en.wikiquote.org/wiki/Alan_Perlis
I used to be more condescending towards the
attitude that
the-whole-world-is-a-stream-of-bytes (having worked with many I/O
devices and languages and computers that had far more evolved record
concepts) but here i am in 2011 and if I can coerce anything into a
stream of bytes - I've got a lot of tools for working with that.
Unfortunately, our tools constrain our thought--and that's the evil
in a "universal" tool such as C.
Program in APL, for example, for six months and I guarantee that your
way of thinking about things will begin to change dramatically. I'm
not saying that "that Iverson language" (as one of my co-workers used
to call it) is the be-all, but its constraints on thinking are very
different from most traditional languages.
--Chuck
17 Oct
17 Oct
6:22 p.m.
New subject: languages and patterns of thought - Re: C trivia - was Re: Dennis Ritchie has died
From: Toby Thain
Sent: Sunday, October 16, 2011 11:25 AM
Language constrains and moulds thought - as George
Orwell also knew
and wrote about.
As did Benjamin Lee Whorf, whose weak form of the hypothesis
("languages influences the way we can think") became a stronger claim
in the hands of his mentor, Edward Sapir. It is the latter, more
disprovable form which influenced Orwell, Vance, Delany, and Elgin.
Rich Alderson
Vintage Computing Sr. Server Engineer
Vulcan, Inc.
505 5th Avenue S, Suite 900
Seattle, WA 98104
mailto:RichA at vulcan.com
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/
2:27 p.m.
On 16/10/11 2:15 PM, Dave Caroline wrote:
It's a bit more subtle than that. The *physical* int depends on the
machine (this is pragmatically important). The *conceptual* int is
architecture independent. (Applies to all C types.)
If you regard these as the same, you are not writing portable C.
--T
I think the
point folks are missing is that it isn't the processor that defines the standard data
types, but the language.
it started out quiet the inverse.
K&R2 page 9 defines an int as depending on the machine see page 32 too.
Dave Caroline
2:46 p.m.
Shoppa, Tim wrote:
Not just bit-aligned data: some processors from that
same broad era (e.g. iAPX432) didn't even require
that instructions be byte or word aligned.
Though C does not have any constructs that involve pointers to
instructions. There are pointers to functions, but those do NOT
necessarily point to a machine instruction.
I think the point folks are missing is that it
isn't the
processor that defines the standard data types, but
the language.
Yes. The language committee tried to accommodate reasonable variations
in machine architecture, up to a point. For instance, they did not
require that integers be stored with a two's complement representation,
even though that is the case for >99.999% of the processors used to run
C code.
4663
days inactive
4664
days old
6 comments
6 participants
participants (6)
-
cclist＠sydex.com -
dave.thearchivist＠gmail.com -
eric＠brouhaha.com -
RichA＠vulcan.com -
toby＠telegraphics.com.au -
tshoppa＠wmata.com