[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Y2K Question



PureBytes Links

Trading Reference Links

Dennis Holverstott <dennis@xxxxxxxxxx> notes:
>JimO wrote:
>> But barring that, I think my trusty TS 3.5 may well keep
>> right on working through the rollover into Y2K.
>
>Errrrr.... maybe but I think you'd better do some testing. Wait for the
>weekend and set your computer clock ahead to 2000. Start the offline
>server and try plotting some ascii data with 99 and 00 dates. TS4
>chokes. Maybe 3.5 will do better.

Ah, that'd be a problem with the ascii date format; I'd bet 3.5
would fail the same way.  Hmmm... come to think of it, how would
you represent data from, say, 1979 through 1991, in ascii?  Can
you put four-digit years in ascii files; if not, what's the difference
between 1980 and 1990?

In any case, real-time data, or broadcast-collected EOD data gets
its dates represented in a mock-Julian format: n-days since 12/31/1899,
so it should be OK.

BTW, does the offline server care what the computer clock date is?
Will it fight me if I try to load data newer than its idea of "now?"
That'd be unfriendly, but not the first case of such stupidity...

FWIW, here's the format of the server's daily data:

-----------------------------------
Format of Omega's daily data file:
Sun Mar 15 14:54:11 1998
All multibyte fields are reverse ordered.
Date is same as tick file, ndays since 12/31/1899

Header - 32 chars, as below, for each symbol in file:
-----------------------------------------------------
01 00
exchange code (1 char) (CME = 0x05)
symbol (12 chars) (Upper case, padded with nulls)
scaling code (1 char) (1/100 = 0x6D)
number of days of data for this symbol (at least two chars, maybe four?)
pad - nulls (to a total of 32 chars from start)

Data - one block as below (26 chars) for each day:
--------------------------------------------------
date  (2 chars)
open  (4 chars)
high  (4 chars)
low   (4 chars)
close (4 chars)
volume   (4 chars)
open int (4 chars)

-------------------------------------------------------

Maybe someone can hack some post-y2k data together and try
it.  I'll do it myself someday, when I get a few other things
out of the way, but if someone beats me to it, let us know
the results.

Jim