|
Beta, IB Refresh Data Different, Sync Time, Transfer Study, Max Bars Recalculation CPU |
June 27 , 2007
16:44 Ensign Howard: Welcome to today's Wed training class with Ensign.
16:45 Ensign Howard: Over the past 6 weeks we have been in beta test phase for changing the program because I was migrating the component library used for Internet connect to a different provider. However, after getting the task finished and some users on the beta, the feedback was more error messages where there were problems with the Internet. The new components did not seem as gracefully tolerant of Internet problem as the original components being used. So a week ago I changed everything back to the original. I still managed to consolidate the 3 vendors used for Internet components down to 2, and that did accomplish a program size reduction of 85K, or about 2% of the program size. So the effort has had a benefit, though it sure has been an exceeding amount of work to get it accomplished.
16:49 Ensign Howard: If you are using a beta version, please check the version date and have it be 06-24 or later so you are off of the earlier versions that attempted to use the Indy library, which proved to be more problematic that the original. Also, beta version has the check box for the IB users who have upgraded TWS to version 873.3 or later, which version changed the way the IB refresh was timestamped.
16:50 Ensign Howard: The change at IB was to timestamp when the period ends whereas in earlier versions they timestamped when the period began. So, on a 15 minute refresh bar, the earlier version timestamp could be 09:00 and the new timestamp for the same bar is now 09:15. The check box on the Setup | Connection form for the Use TWS 873.3 permits Ensign to handle either the older TWS or the current TWS. IF using 873.3 version, then check the box on the Setup | Connection form so Ensign adjusts the refresh timestamps as needed. I will probably continue on beta status for another week or so, and is feedback dies down or is positive about the beta, then have an official release.
17:30 @Buffy2: Howard just want to let you know we appreciate the Topic Roster for newsletters.
17:31 Ensign Howard: Great buffy, glad it gets used. I really appreciate all that is done on the dacharts site to support the Ensign community. Thank you very much.
17:34 @Buffy2: Welcome from Scylnx and I
IB Refresh Data Different
17:36 bogeybunky: Howard when I refresh IB data because my data seems different from others it often does not seem to refresh. So I delete all bars and then refresh maximum. That seems to do okay except for last few bars. Then I do refresh Today and everything seems okay. Am I doing this the correct way?
17:38 Ensign Howard: When you Refresh Maximum the source auto switches to use DTN Market Access because we can get more data from that source, However, that source is through the delay requirement, so you lack the last 10 to 20 minutes of being current. When you refresh Today, you are then refreshing from IB and they will be through current tick.
17:39 Ensign Howard: If this is a time based chart, then having Preserve Existing unchecked and refreshing from IB should have been sufficient and eliminate the extra steps of erasing and refreshing Max and Today. Is it a time based chart?
17:40 bogeybunky: I do not have Preserve checked. it is 5 min chart
17:40 Ensign Howard: Let me make commentary on this 'difference' you perceive.
17:41 Ensign Howard: IB does not time stamp their ticks, so we have to rely on your computer system clock to time stamp the ticks when they are received. If, your computer system clock has migrated ahead, say 30 seconds, then you will be ending a bar and starting a new one 30 seconds sooner than everyone else does, or than IB does.
17:42 bogeybunky: It happens when some in CCI room get a signal and I do not have one. Sometimes if I refresh I then have the same signal, but did not have it before. So I need to make sure I have correct time on my computer clock. I use atomic clock to try to keep it correct, but maybe I should have it sync more often
17:42 Ensign Howard: Best I can suggest is to try to keep the computer clock in sync with exchange time, and I offer the following work around.
7:43 Ensign Howard: Click menu Setup | Connection and check the box for the FXCM feed.... this will activate that free feed and it contains a clock that Ensign can use to sync your computer clock.
17:44 Ensign Howard: Click menu Setup | Computer and check the option to Set Computer Clock from Data Feed. Now the Ensign program will sync your computer clock once per hour to the exchange clock seen on the FXCM feed. Try it, and see if your collected bars then match more closely with others and with the IB refresh.
17:44 bogeybunky: Ok Howard, that is simple enough. I will do that. Again thank you very much and please go work on your project. I really appreciate all you do. You need your free time.
Transfer Study
17:59 MakesOwnWeather: Ensign Howard, I have a challenge with the redrawing of charts. sometimes, just the slightest "jarring" of Ensign Sofware by a CRTL-P on some other chart, the chart I am looking at refreshes and I look real-time indicators. Once refreshed, they are gone and the chart does not reflect what I have built with DYOs. Any suggestions on how to avoid this? I Loose real-time indicators.
18:02 Ensign Howard: MOW, the DYO that is reading GV or Quote page values have this issue that in a recalc they only see a current value. They cannot see any history at a particular bar position. This is like unto the issue discussed on our web page for Study Transfer.
18:04 Ensign Howard: One DYO line that is useful to tolerate this behavior is the Global Action | Plot Study Transfer in GV. This special DYO line does not clear itself or recalculate because it is assumed it has collected momentary data that cannot be replace in the recalculation.
18:03 MakesOwnWeather: Yes, EH, I understand fully. I am just wondering what "things" cause the redrawing of charts besides the CTRL-P on a specific chart.
Max Bars Recalc
18:05 Ensign Howard: Besides the visit to a study or chart property form, the chart is collecting bars and can reach the Maximum Bars setting during the day. At which point the chart will resize by dropping a set of oldest bars, and then recalculate based on the changed data set. So there is a recalculate from the resize activity.
18:06 Ensign Howard: one way to minimize that auto resize and recalc is to position the cursor earlier in the file and delete bars leftward. So you have pre deleted enough bars so that today's collection will not reach the Maximum Bars setting
18:07 @Buffy2: If you are near max bars at start of day then the drop and recalc is CPU burden? Would be wise to have days worth of bars left before max bars at start of day?
18:08 Ensign Howard: yes, the drop and recalc does require CPU time/burden that could be avoided by pre-deleting with pop-up menu Delete | Delete Left
18:08 @Buffy2: Thank you
Cont.
18:08 MakesOwnWeather: Ensign Howard, if I am using a Global Flag from another timeframe in indicator logic how can Global Action - plot ... be used to preserve status?
18:09 Ensign Howard: You can read the GV you want to preserve and it can write to save GV, which helps in a recalc because for each bar position, it will not read. Read is restricted to working with real-time. Yet I think it can write GV at each bar position which would update the GV with the intended value for that bar position. So that might be your solution MOW. Try it.
18:26 MakesOwnWeather: Thank you for your help, Ensign Howard.
18:20 Ensign Howard: OK, thanks for the class attendance... see you again next week.