I have just checked the Trickle info for the model that I am currently running. The leftmost column in the table is headed \"Time Sent (UTC)\" and the most recent entry is for 09 Apr 2008 09:24:16. As I look at my clock the time is 09 Apr 2008 09:39:00 BST which equals 08:39:00 UTC. It appears that I am reporting in the future (or could the server be reporting receipt in local time??)

F.

---

Well spotted, Fred. This problem has been noticed before.

The simplest solution, I think, would be to remove the UTC text from the heading in the trickle \'time sent\' column, since the times seem to be in the user\'s locale (BST for you and me). The logic for that is that the information is for the user and should be in terms familiar to the user.

An oddity is that, last time I looked at this, the \'last contact\' times were in UTC - perhaps for the convenience of users with an international spread of PCs! Actually, the computer page can be sorted by contact time, so the time ought to be on a sortable scale (e.g. UTC).

As a matter of interest, is this a cross-project problem? I know that CPDN and the BBC-CCE both had it.

I think we need to wait until until someone in a time-zone far, far away can check their trickles and report how they match (or otherwise) their local computer clock.

I\'ve noticed it before at CPDN, and my personal take on it was:

BOINC is an international platform, and is pretty consistent in showing every time reference in UTC format.

None of the other projects I run uses the \'trickle\' mechanism, so we have no general point of reference for that particular page. I think it\'s a CPDN homebrew lash-up.

We\'ve talked often enough on these pages about trickles arriving late, long after our computers have reported them, when the CPDN servers are under stress.

Ergo, I reckon that what we\'re seeing is a timestamp in the server\'s local (Oxford or Milton Keynes) wall-clock time. If someone can match up their message log \'trickle-up\' entries with trickle timestamps, we could even work out where the server is (Oxford\'s are properly sychronised with an NTP time server: Milton Keynes, notoriously, is not, and the php board is currently about seven minutes slow).

Edit - the latest trickle for result 7177229 was initiated at 09/04/2008 11:22:38, local computer time. The clock on that host is currently 01:59 slow, so the trickle was actually sent at 11:24:37, and acknowledged at 11:24:47. It\'s shown on the website as 11:25:10.

Oxford!

<snip>
Ergo, I reckon that what we\'re seeing is a timestamp in the server\'s local (Oxford or Milton Keynes) wall-clock time. If someone can match up their message log \'trickle-up\' entries with trickle timestamps, we could even work out where the server is (Oxford\'s are properly sychronised with an NTP time server: Milton Keynes, notoriously, is not, and the php board is currently about seven minutes slow).
</snip>

That is what I surmised was happening, rather than displaying in users\' local time. In which case, someone needs to do something more fundamental to get the timestamp to match the column header rather than vice versa.

F.

---

<snip>
Ergo, I reckon that what we\'re seeing is a timestamp in the server\'s local (Oxford or Milton Keynes) wall-clock time. If someone can match up their message log \'trickle-up\' entries with trickle timestamps, we could even work out where the server is (Oxford\'s are properly sychronised with an NTP time server: Milton Keynes, notoriously, is not, and the php board is currently about seven minutes slow).
</snip>

That is what I surmised was happening, rather than displaying in users\' local time. In which case, someone needs to do something more fundamental to get the timestamp to match the column header rather than vice versa.

F.

It\'ll be a relatively trivial mod to the query invoked/defined by the php code driving the page, and it\'ll be easy to copy the relevant function/format flags from the BOINC source for equivalent pages. But I\'m afraid I don\'t know php well enough to re-write the query at short notice. If someone could forward me a copy of the php source for the page, I could have a look at it.

If someone can match up their message log \'trickle-up\' entries with trickle timestamps

Trickles for a slab model I\'m running are reported here.

I am in Alberta, Canada which is UTC-7 (MST) though we are currently on DST so it\'s UTC-6.

The last 4 trickles in my log (order reversed to match order on trickle report):

Wed 09 Apr 2008 01:37:34 PM MDT|climateprediction.net|Sending scheduler request: To send trickle-up message.
Wed 09 Apr 2008 07:20:05 AM MDT|climateprediction.net|Sending scheduler request: To send trickle-up message.
Wed 09 Apr 2008 01:14:20 AM MDT|climateprediction.net|Sending scheduler request: To send trickle-up message.
Tue 08 Apr 2008 06:56:16 PM MDT|climateprediction.net|Sending scheduler request: To send trickle-up message.

The above show on the web page as:

 
Time Sent (UTC)  	Host ID  	Result ID  	Result Name  	Phase  	Timestep  	CPU Time (sec)  	Average (sec/TS)
09 Apr 2008 20:37:52 	844627 	7378595 	hadsm3fub_0363_005940979_0 	2 	216,040 	969,602 	2.0400
09 Apr 2008 14:20:49 	844627 	7378595 	hadsm3fub_0363_005940979_0 	2 	205,238 	947,172 	2.0392
09 Apr 2008 08:15:09 	844627 	7378595 	hadsm3fub_0363_005940979_0 	2 	194,436 	925,296 	2.0395
09 Apr 2008 01:56:34 	844627 	7378595 	hadsm3fub_0363_005940979_0 	2 	183,634 	902,796 	2.0385

The times show 7 hours difference so it seems the web page is showing proper UTC times.

---

... The times show 7 hours difference so it seems the web page is showing proper UTC times.

Thanks, Dagorath, that\'s really helpful.

The seven hours is more likely BST - MDT = (UTC + 1) - (UTC - 6) = 7.

So that separates the theories: it\'s server local time that\'s being used and not the user\'s locale. So the fix is either to remove the UTC from the heading and leave the user to work out that it\'s UK civil time, or fix the times to UTC (which is more consistent and fairer [Brits will be confused as well as everyone else]).

Or change Sent to Received.

Or change Sent to Received.

Oh I do like a bit of lateral thinking ;)

F.

---