ObjectAda - no clock drift! 
Author Message
 ObjectAda - no clock drift!

Quote:

> The Ada rm in D.8 paragraph 41-43 requires vendors to give "an upper bound on
> the drift rate of Clock with respect to real time." I was curious what my
> Aonix ObjectAda compiler said, so I looked it up.

> The documentation Requirements section of the ObjectAda 7.1 NT docs (p 2-9)
> has the following statement:  There is no software clock drift.

> What is meant by this? Surely not what I think it's saying!

Since you express surprise, I would concur that it must not be
what you were thinking, whatever that is.

I believe that Aonix is saying that the software they supply
introduces no drift, and if you see drift you should take it
up with Intel (or Cyrix, or AMD).

Since Aonix does not sell hardware, I don't know how much
further they can go (aside from saying you might also take
it up with DEC/Compaq or Samsung, but they don't support
Alpha the last I heard (drift-free or not)).

Larry Kilgallen



Sun, 03 Dec 2000 03:00:00 GMT  
 ObjectAda - no clock drift!

Quote:

>Apparently, Aonix has made atomic clocks obsolete with the simple use of a
>Wintel PC.

>The Ada rm in D.8 paragraph 41-43 requires vendors to give "an upper bound on
>the drift rate of Clock with respect to real time." I was curious what my
>Aonix ObjectAda compiler said, so I looked it up.

>The documentation Requirements section of the ObjectAda 7.1 NT docs (p 2-9)
>has the following statement:  There is no software clock drift.

>What is meant by this? Surely not what I think it's saying!

>T.E.D.

I *think* what this means is that there is nothing in ObjectAda to make the
clock drift on "Wintel" hosts any worse than it already is.

If clock drift with respect to real time is important to you, you should not
be relying on the unaided operating system clock of a "Wintel" host, or any
product that derives its clock from such a clock.

NTP has been ported to Windows NT.  I do not know how effective it is at
disciplining the clock on NT; it is generally effective on Unix hosts.
Visit http://www.eecis.udel.edu/~ntp for details.

--

Advanced Technology Center                   Phone (949) 583-9119
22982 Mill Creek Drive                                   ext. 220
Laguna Hills, CA 92653                       Fax   (949) 583-9213



Sun, 03 Dec 2000 03:00:00 GMT  
 ObjectAda - no clock drift!

Quote:

>Apparently, Aonix has made atomic clocks obsolete with the simple use of a
>Wintel PC.

>The Ada rm in D.8 paragraph 41-43 requires vendors to give "an upper bound
on
>the drift rate of Clock with respect to real time." I was curious what my
>Aonix ObjectAda compiler said, so I looked it up.

>The documentation Requirements section of the ObjectAda 7.1 NT docs (p 2-9)
>has the following statement:  There is no software clock drift.

>What is meant by this? Surely not what I think it's saying!

I think it's saying that the way the software is written, it will not cause
the current
time to drift as a side-effect of the way it keeps time the way some
algorithms do.  However, if the hardware clock drifts, it has no control
over that.

- Corey



Sun, 03 Dec 2000 03:00:00 GMT  
 ObjectAda - no clock drift!

Quote:

> I guess I was mostly posting this to see if this is a just a farcical non-
> compliance with the documentation standard, or if I'm misunderstanding the
> standard.

I think the Documentation Requirements in the RM shouldn't be there at
all.  You can't legislate good documentation.  If you try, you get the
sort of documentation you're complaining about -- totally useless
information, intended purely to satisfy the letter of the law.

Note that the validation process makes no attempt to verify that
Documentation Requirements are met properly.  If the vendor had written,
"Humpty Dumpty sat on a wall" as the answer to that question, they would
still be just as validated.

- Bob

--
Change robert to bob to get my real email address.  Sorry.
--
Change robert to bob to get my real email address.  Sorry.



Sun, 03 Dec 2000 03:00:00 GMT  
 ObjectAda - no clock drift!

Quote:

> In all fairness to Aonix, there's no way they could know what the clock drift
> rate for the hardware in any particular PC is. But they should have said that,
> rather than answering half the question in an incredibly terse manner.

Well, Aonix was describing a compiler, not a computing system.

If the RM applies to compilers rather than computing systems,
_it_ should provide the wording that excludes drift beyond the
control of compilers.

On the other hand, some of the responses in this group have
indicated that the intentional style of the RM is terse. If
so, Aonix would seem to be fully justified in responding in
the same style.

Larry Kilgallen



Mon, 04 Dec 2000 03:00:00 GMT  
 
 [ 8 post ] 

 Relevant Pages 

1. Clock drift in gnat 3.13p on Windows

2. clock scan [clock format [clock seconds]] fails

3. TS Drifting downSTREAM

4. Button Drift

5. Voltage drift in an analog data acquisition

6. wm geometry drifts up and left on linux

7. change TOD clock with set and store clock

8. How to do IP clock balance with external clock

9. How to Generate 1MHz clock from 44MHz Clock??

10. Problem with clocking logic usign a derived clock

11. Gated Clock(double rate clock)?

12. clocking on rising AND falling edge of a clock

 

 
Powered by phpBB® Forum Software