[Tfug] A sense of time

[Bexley Hall]

> This is probably the most interesting post I have 
> seen on this list :)

Wait until next week when I ask the proverbial
question, "Why is there air?"  :>

> The question appears to be a blend of technical and
> philosophical thought.

The *technical* is easy (or, rather, just an exercise
to
be cranked through).  I won't stake a claim to the 
philosophical aspect  :>  but, rather, to the
*personal*
aspect of "Time" that each individual defines for
himself
(along with a myriad of possible REdefinitions!)

> The primary question is "What is time?". If we could
ever
> really get a definitive answer to that we may be
able to
> find a technical solution. Time is not what is on
our 
> clock, on our calendar or a relationship between our

> planet and its physical position in the universe.
All 
> of those devices are man-made creations that help us

> measure the effect of Time.

Time is *always* "relative" (think about it...)

> For example Time exists without the definition of a
second.

Of course.  As does mass without the definition of a
gram/pound/etc.

> It exists on planets that have no rotation, thus no
"Day".
> It exists in the absence of anything to measure it.
We 
> have only been able to recognize that it exists at
all 
> because we see it as a phenomena. As such we have
studied
> it and come up with terms to help us deal with it.
Terms
> such as second and minute and year. These terms try
to
> explain the phenomena of time but do not define it.
For
> example: I can say it is 09:29:07am but I can also
say 
> it is 10:29:07am and still be correct.

You can also say it's MILLER time!

> It happens all the time, when I call somebody to the
east 
> of me. Time is not the terms we use to describe it.

> If I read the post correctly, you are looking for a
way 
> for us to mark our position in time consistently and

> accurately. In your mind the technology of "keeping
time" 
> should reflect more the reality of time and not the
terms
> we use that relate to it. Monitored and recorded
events 
> should always relate in a consistent way to previous
events.
> If one event (file creation) precedes another event
(file 
> update), the latter should reflect the delta between
the 
> two events accurately. There should not be any way
that 
> the latter should precede the former.

Yes, but that is the "easy" part of the problem.  I
can
ensure that all events/files (a file timestamp is just
an event in another guise), etc. bear correct
relationships
to each other such that I can *always* order them in
their
correct chronilogical order regardless of the user's
"current" sense of time.

The PROBLEM is that the user's *brain* is part of the
equation.  A *key* part!  *THE* key part!  And, the
user
sees all "times" in relationship to something in *his*
context -- which may *vary*!

> All of the tools we use operate on establishing a
point 
> in time and marking time following that event. The
month
> is August: 8 months from the beginning of a new
year. 
> It is 9am: 9 hours from the start of a new day. It
is 
> 2007: 2007 years from the birth of Christianity.
When we

Well, not *really*...  :>

> set our clocks on our PC's we establish the starting
point
> and the hardware tries its best to keep an accurate
count 
> of crystal oscillations so as to measure the time
that has
> transpired between the setting of the start point
and "now".

> The problem is that we have the ability to set that
start
> time to be anything we want it to be. There is
nothing 
> that keeps us from lying to the hardware. What we
need 
> is the ability to force a starting time on a
computer 
> that is predefined and immutable. This has been
solved in
> other technologies. I think of the GPS systems and
the 
> various public atomic clocks that are available. I
am 
> wearing a watch that picks up the time from the air
and 
> is highly accurate.

No!  This is NOT the problem!  :>  I can put a
timekeeping
device into one of my designs (I design the hardware
on
which my software runs) *and* prevent *anything* from
modifying that time -- even the OS!  (i.e., never
implement a "write" to that device... just "reads").
And, I can *chose* to define a mapping between that
hardware timekeeping device and "time of day" that
follows
whatever rules I decree...

*But*, if I have to let the user define his own "sense
of
time" (even if I let that "float" wrt my own
hardware's
"sense of time"), then these problems arise (assuming
I let
the user set his sense of time MORE THAN ONCE in the
life
of the product).

> Because our marking of time is our own creation we
will 
> always be dependent on some device of our making to
set 
> the starting point for other devices.

Moreover, the choice of *which* device the user wants
to
use AT THAT TIME to define his "sense of time" may
vary.
"A man with two clocks is never sure of the RIGHT
time"

> There must exist some "gold standard" for getting
that 
> initial starting place. It is not rational for every

> device to have the ability to watch the stars,
calculate
> our position in the universe and report a calculated

> time reference. We must depend on something else to
> do that for us. In developed countries we currently 
> we have access to these types of sources via the
air, 
> internet, cell towers, satellite and phone. The
choice 
> of which technology to use is based on the resources

> we have when developing the devices that depend on
it.

> Clearly, as Ben has stated, we are not there yet. It

> is not a problem in search of a solution. It is a 
> solution in search of an implementation. We can not 
> continue to use current solutions because the
initial 
> starting point for time calculations are under our 
> control. The solution will have to involve taking 
> that control away from us. Perhaps future hardware 
> can have have components that can extract the time
from
> various sources.

> If you have a network card, maybe it could have a 
> component that taps the internet source for time. A 
> radio on the mother-board that picks up time form
the
> air. As computers and wireless networks such as G3
> become more ubiquitous they could be imbued with the
> ability to take time from cell towers. I believe the

> days of setting time via the land-line phone are 
> limited but a huge percentage of people still
connect
> to other networks via a modem. Modems could easily 
> be made to deliver correct time.

There are a lot of (inexpensive) schemes for keeping
quite good time without resorting to "high tech"
approaches (e.g., I implement an LFC in every mains
powered device that I design so that the local
oscillator
can be tweeked (long term) to keep "good time".  This
is just a convenience optimization (folks get pissed
if they have to reset the clock on their Gonkulator
every 3 days because of creep!  :>  )

[extra credit:  ever wonder how those cheap digital
wrist watches can keep such good time??]

> So, there are solutions. They are just not
ubiquitous 
> enough to be depended on. What we need to do is to 
> communicate our desires to the engineers of today 
> and tomorrow so that the solutions become reality 
> instead of wishes,

(sigh)  You've still missed the last link in the
chain:
the user's brain!  How do we get user's to think of
time in a monotonic continuum *despite* the devices
he/she/we employ to keep track thereof?  How do we
get people to "fix" (as in, "make immutable") their 
points of reference so that *any* notion of time
can be expressed consistently?

I.e. "we" (as *a* society -- not true of all
societies)
may have adopted the date-/time- keeping scheme you 
outlined above.  It enables us to interact with each
other describing events in time in a meaningful way
without ambiguity (to *any* degree of precision
based on the needs of those engaged in that
discourse).

*But*, there is nothing that "forces" a user to
exercise the same sense of discipline in his
interaction
with a *device*!

By way of analogy, imagine two *people* discussing
temporal events when each individual is free to
define his/her own sense of time!  ("Now, when you
said you went to the doctor at 3PM, was that *before*
or *after* I stopped by your house at 9:30PM?  I
noticed that the clock on your wall said 11:15 at
that time but I recall you said you had turned it
backward an hour sometime that day...")  :<


       
____________________________________________________________________________________
Take the Internet to Go: Yahoo!Go puts the Internet in your pocket: mail, news, photos & more. 
http://mobile.yahoo.com/go?refer=1GNXIC