Tools to audit our study efficiency.

Questions about MultiCharts and user contributed studies.
Posts: 2152
Joined: 21 Jul 2007
Has thanked: 212 times
Been thanked: 424 times

Tools to audit our study efficiency.

Postby bowlesj3 » 20 Nov 2009

I copied this directly from another post (skip it if you have read it). It is a request for two new reserve words. The first one might be called, TickArrivalTime_s. It would be used to track how well our studies are doing relative to how they were doing before we made a recent change (as one example). If I (or anyone) gets any ideas related to auditing our own studies maybe they could dump them in this thread. Here are the details for this one.

The MC missing ticks issue (and the theory I had that maybe MC was smart enough to selectively bypass ticks during heavy load) has got me thinking that it would be handy if MC could somehow analyse heavy loads and help the user detect studies that they should focus on for the purpose of going back to the study and figuring out how to make them work faster (or shutting them off if they just are not needed at specific times).

I am thinking along the line of MC providing tools to help us audit our own studies for efficiency (database programs have these features but I am not thinking that MC provide tools this complex). More realistically if MC marks ticks as they come in with an arrival time stamp and MC provides a command allowing the user to access this using EL code within each study their study could detect if the study is processing it excessively late.

Here is some sphedocode for it if there was a reserve word to get at it.

Code: Select all

If CurrentTime_S minus TickArrival_S > AcceptableUpperLimit then
{We may have a problem in our study}
do whatever

Another idea would be to write out the difference "CurrentTime_S minus TickArrival_S" to a log along with the study name and current time so the user could analyse the report at days end to see if the times were too large in places and at what time of day it ways.

A command returning a master tick number would be handy for helping us audit our studies as well. If you wish to understand why I think such a master tick number exists in the central tick buffer then study the very first post in this thread Such a command would allow us to easily track the order our studies execute (chart order, workspace order) and may have some debugging uses.

I am at the point where I am reluctant to give MC too many more studies. It amazes me that it can get through all of them before the next tick arrives. Tools like this would allow us to maximize our use of MC by knowing when we are pushing it too hard (realistically no TS/MC type program is infinite).

Return to “MultiCharts”