microsecondClockValue

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

microsecondClockValue

Eliot Miranda-2
Hi All,

    I see that access to the microsecond clock is called "microsecondClockValue".  This name is wrong. It derives from millisecondClockValue which was indeed a value. The millisecond clock started from zero on every image startup and wrapped around every 45 days or so.  The microsecond clock is quite different; it is an absolute clock whose zero is midnight on January 1st 1901, the start of the 20th century (damn those monks). So "Value" should be omitted from the selector.  In Squeak we provide both Time utcMicrosecondClock and Time localMicrosecondClock which provide the number of microseconds since the start of 1/1/1901 in utc and local time zones respectively.  It wraps round no sooner than 24/4/20168 (/not/ 2168).  Time millisecondClockValue is provided for backward compatibility.

_,,,^..^,,,_
best, Eliot


Reply | Threaded
Open this post in threaded view
|

Re: microsecondClockValue

Tobias Pape

On 23.05.2016, at 19:37, Eliot Miranda <[hidden email]> wrote:

> Hi All,
>
>     I see that access to the microsecond clock is called "microsecondClockValue".  This name is wrong. It derives from millisecondClockValue which was indeed a value. The millisecond clock started from zero on every image startup and wrapped around every 45 days or so.  The microsecond clock is quite different; it is an absolute clock whose zero is midnight on January 1st 1901, the start of the 20th century (damn those monks). So "Value" should be omitted from the selector.

Isn't it still a value? I frankly do not understand the distinction here..
Even more, "microsecondClock" doesn't hand me a "clock" object… :)
 
Best regards
        -Tobias


>  In Squeak we provide both Time utcMicrosecondClock and Time localMicrosecondClock which provide the number of microseconds since the start of 1/1/1901 in utc and local time zones respectively.  It wraps round no sooner than 24/4/20168 (/not/ 2168).  Time millisecondClockValue is provided for backward compatibility.
>
> _,,,^..^,,,_
> best, Eliot



Reply | Threaded
Open this post in threaded view
|

Re: microsecondClockValue

Eliot Miranda-2
Hi Tobias,

On Mon, May 23, 2016 at 11:58 AM, Tobias Pape <[hidden email]> wrote:

On 23.05.2016, at 19:37, Eliot Miranda <[hidden email]> wrote:

> Hi All,
>
>     I see that access to the microsecond clock is called "microsecondClockValue".  This name is wrong. It derives from millisecondClockValue which was indeed a value. The millisecond clock started from zero on every image startup and wrapped around every 45 days or so.  The microsecond clock is quite different; it is an absolute clock whose zero is midnight on January 1st 1901, the start of the 20th century (damn those monks). So "Value" should be omitted from the selector.

Isn't it still a value? I frankly do not understand the distinction here..
Even more, "microsecondClock" doesn't hand me a "clock" object… :)

The Smalltalk-80 code has always maintained the distinction:

Time secondClock
Time millisecondClockValue

 The former is a clock, measuring time from a fixed point in history.  the latter is a counter starting from an arbitrary point.  I find it a very useful distinction.  Our current microsecond time is a clock, not a value.


Best regards
        -Tobias


>  In Squeak we provide both Time utcMicrosecondClock and Time localMicrosecondClock which provide the number of microseconds since the start of 1/1/1901 in utc and local time zones respectively.  It wraps round no sooner than 24/4/20168 (/not/ 2168).  Time millisecondClockValue is provided for backward compatibility.
>
> _,,,^..^,,,_
> best, Eliot






--
_,,,^..^,,,_
best, Eliot


Reply | Threaded
Open this post in threaded view
|

Re: microsecondClockValue

Tobias Pape

On 23.05.2016, at 21:43, Eliot Miranda <[hidden email]> wrote:

> Hi Tobias,
>
> On Mon, May 23, 2016 at 11:58 AM, Tobias Pape <[hidden email]> wrote:
>
> On 23.05.2016, at 19:37, Eliot Miranda <[hidden email]> wrote:
>
> > Hi All,
> >
> >     I see that access to the microsecond clock is called "microsecondClockValue".  This name is wrong. It derives from millisecondClockValue which was indeed a value. The millisecond clock started from zero on every image startup and wrapped around every 45 days or so.  The microsecond clock is quite different; it is an absolute clock whose zero is midnight on January 1st 1901, the start of the 20th century (damn those monks). So "Value" should be omitted from the selector.
>
> Isn't it still a value? I frankly do not understand the distinction here..
> Even more, "microsecondClock" doesn't hand me a "clock" object… :)
>
> The Smalltalk-80 code has always maintained the distinction:
>
> Time secondClock
> Time millisecondClockValue
>
>  The former is a clock, measuring time from a fixed point in history.  the latter is a counter starting from an arbitrary point.  I find it a very useful distinction.  Our current microsecond time is a clock, not a value.
>

This may be due to English not being my native language, but "clock" for me has always
the "feel" of a thing/device/object, not a number/value I could read off that thing.

So asking for
        Time microsecondClock
sounds to me like "give me that clock thing that counts in microseconds". I might ask
that to halt, rewind, or tell me its value….

But if the value-able interpretation is more common than I had though, why not? :)

Best regards
        -Tobias