search

van-smith / OPBM

Open Productivity Benchmark
1 stars 0 forks source link

Spelling error / inconsistencies: uninstall #21

Closed van-smith closed 13 years ago

van-smith commented 13 years ago

"Uninstall" is misspelled / inconsistently spelled throughout the Atoms. All spellings should be "uninstall".

ghost commented 13 years ago

Fixed.

van-smith commented 13 years ago

The spelling errors still exist in the Atom list window.

ghost commented 13 years ago

It was Un/Install? Fixed.

van-smith commented 13 years ago

Yes.

van-smith commented 13 years ago

The spelling errors persist as of 8.26.11. I will correct them and upload the fixes this weekend.

Please spell "uninstall" correctly for all future edits.

van-smith commented 13 years ago

I corrected all of these spelling errors and pushed scripts.xml.

ghost commented 13 years ago

The changes made to scripts.xml include not only spelling corrections, but also a restructuring at the suite, scenario and molecule levels. These changes have not been discussed with me and are not consistent with the OPBM's design, such as calling a suite "Official Run".

We need to discuss these changes so they operate within OPBM's intended design.

van-smith commented 13 years ago

Rick, in addition to the persistent (and new) spelling errors, there was no obvious order to the scripts.xml file.

As a first order need for AMD, we have to provide a simple way for them to easily run all of the scripts. You eliminated the entry for "run all with install" and I didn't see a way to accomplish this otherwise. You have also not reported progress on the Official Run and there is no support for iterations yet.

I created structure that should generate results somewhat consistent with my intentions for the moment. The "Official Run" suite will, if you are calculating scores as I have specified, calculate results by geometric mean. This is not ideal since the arithmetic mean is more appropriate for iterative runs, but until this support is added to the harness, this method is an acceptable stand-in.

I created an issue several days ago discussing molecule/scenario/suite scoring structure:

https://github.com/van-smith/OPBM/issues/34

It is not yet possible to generate scores within OPBM as I have defined -- notice that the scoring aggregation differs from the execution aggregation. It is not a high priority to accomplish this. Scoring as is generated currently is acceptable (if it is being carried out to my original spcifications -- geomean to combine atoms, then geomean to combine molecules, etc.), at least below the Official Run level.

I am closing this issue again, since this discussion is unrelated to the original topic. If you feel that I have introduced a bug with my changes, open a new issue.