Difference between revisions of "Milliseconds"
(New page: '''Status: Requirements Complete; Estimate Needed''' == User Story == As a researcher in a language domain, I want to see milliseconds in the timestamps, so that I can understand better ...) |
(→Notes/Comments) |
||
| Line 6: | Line 6: | ||
== Notes/Comments == | == Notes/Comments == | ||
| − | * Phil has started asking for this. Email 3/12/2008. ET Mtg 4/9/2008. | + | * Phil Pavlik has started asking for this. Email 3/12/2008. ET Mtg 4/9/2008. |
* Connie Guan is asking for this too. 4/2008. | * Connie Guan is asking for this too. 4/2008. | ||
| − | * Previously asked for by Ying back in 2005 | + | * Previously asked for by Ying Liu back in 2005. |
* Maintain order of events which occur within same second. [Anders, email 10/3/2007] | * Maintain order of events which occur within same second. [Anders, email 10/3/2007] | ||
** "I realize this is a very fine point, but we thought we'd mention it anyway: In E08E7-Jan24-10-04-23.xml there is a deletion event followed immediately by a hint request. Both these events occur within the same second, so carry the same time stamp to 1-second granularity: 2006-01-24 10:58:41.:In the exported log, it turns out the DELETION event is listed AFTER the hint request and before the hint response. This sequence is confusing to look at, because it suggests the student did something in the middle of the hint sequence, and that is not supposed to be possible on our interface. Because our logs are sequentially generated within a single thread of execution, we know the order of events in the logs corresponds to the order in which the events occurred in time, even if timestamps do not have sufficient granularity distinguish them. So it would be nice if that order could be preserved in the exported data." | ** "I realize this is a very fine point, but we thought we'd mention it anyway: In E08E7-Jan24-10-04-23.xml there is a deletion event followed immediately by a hint request. Both these events occur within the same second, so carry the same time stamp to 1-second granularity: 2006-01-24 10:58:41.:In the exported log, it turns out the DELETION event is listed AFTER the hint request and before the hint response. This sequence is confusing to look at, because it suggests the student did something in the middle of the hint sequence, and that is not supposed to be possible on our interface. Because our logs are sequentially generated within a single thread of execution, we know the order of events in the logs corresponds to the order in which the events occurred in time, even if timestamps do not have sufficient granularity distinguish them. So it would be nice if that order could be preserved in the exported data." | ||
| − | * Developer Note: Scarab Issue: DS65 Need to store microseconds in the action time stamps. | + | * Developer Note: Scarab Issue: DS65 Need to store microseconds in the action time stamps. [Aug 2005] |
Revision as of 14:23, 1 September 2009
Status: Requirements Complete; Estimate Needed
User Story
As a researcher in a language domain, I want to see milliseconds in the timestamps, so that I can understand better where the subject is having some trouble and to get better order of events that occur within the same second.
Notes/Comments
- Phil Pavlik has started asking for this. Email 3/12/2008. ET Mtg 4/9/2008.
- Connie Guan is asking for this too. 4/2008.
- Previously asked for by Ying Liu back in 2005.
- Maintain order of events which occur within same second. [Anders, email 10/3/2007]
- "I realize this is a very fine point, but we thought we'd mention it anyway: In E08E7-Jan24-10-04-23.xml there is a deletion event followed immediately by a hint request. Both these events occur within the same second, so carry the same time stamp to 1-second granularity: 2006-01-24 10:58:41.:In the exported log, it turns out the DELETION event is listed AFTER the hint request and before the hint response. This sequence is confusing to look at, because it suggests the student did something in the middle of the hint sequence, and that is not supposed to be possible on our interface. Because our logs are sequentially generated within a single thread of execution, we know the order of events in the logs corresponds to the order in which the events occurred in time, even if timestamps do not have sufficient granularity distinguish them. So it would be nice if that order could be preserved in the exported data."
- Developer Note: Scarab Issue: DS65 Need to store microseconds in the action time stamps. [Aug 2005]
See complete DataShop Feature Wish List. [[Category:DataShop]
