Kepler

I went to the kepler repository and did a "Browse all components" action. The website showed 373 components (on 37 pages, one page has 10 components). Then I ran the kepler and clicked the "CoreActors kar" tree.

It is very interesting that the components on website almost one-by-one map the ones on kepler "CoreActors kar" tree. For example, the first one on website is ANOVA. And first one on kepler is ANOVA as well. I downloaded the kar file from the website and tried to open it on kepler 2, but it failed.

Since they have the same name, i will assume they are same. So I will go ahead to remove those components on the repository which can be found on the kepler components tree.

Thanks Jing. Yeah, just go ahead and delete any actor that is in the local repository and the remote repository unless it is owned by someone other than the kepler user. They don't need to be in there anyway.

After compared, those actors are not found on the kepler local library:

Authentication
kar.454.1
actor.454.1

BlastOne
kar.34.1
actor.238.1

BlastTwo
kar.35.1
actor.239.1

Command Line Exec
kar.53.1
actor.185.1

File Stager
kar.107.1
actor.196.1

File Fetcher
kar.100.1
actor.195.1

GenBank
kar.131.1
kar.245.1

GriddlesFileCopy
kar.476.1
actor.476.1

Griddles Output File
kar.148.1
actor.444.1
Griddles Input File
kar.147.1
actor.443.1

Griddles Exec
kar.145.1
actor.442.1

GridReset
kar.144.1
actor.363.1

JCOGGridFTP
kar.451.1
actor.451.1

JCOGPROXYExec
kar.450.1
actor.450.1

Invoke Service
kar.164.1
actor.150.1

JCOGWorkflowExec
kar.452.1
actor.452.1

JGridlet Creator Jagan Kommineni
kar.502.1
actor.502.1

LocalBlastOne
kar.174.1
actor.246.1

Permanent String Constant Wibke Sudholt
kar.203.1
actor.290.1

Parameterized Globus Job
kar.201.1
actor.200.1

Port Line Reader
kar.211.1
actor.254.1

RExpression Dan Higgins
kar.216.1
actor.360.1

Run Job Grid Client Ilkay Altintas
kar.232.1
actor.202.1

Single Fire Constant Efrat Jaeger
kar.262.1
actor.128.1

Updated Grid FTP Ilkay Altintas
kar.314.1
actor.199.1

SST Data Access Yuhong Xiong
kar.524.1
actor.523.1

ChadsCompositeActor Yuhong Xiong
kar.519.1
actor.518.1

Current Time_sigsit jie liu and edward a lee
kar.571.1
actor.730.1

Most of these seem like variants on existing actors. You should check if they were simply uploaded for testing purposes, or if they really represent an adaptation of the functionality that is useful. If the former, they can be deleted too. Feel free to contact me if you have questions.

I saw several actors are in modules that are not in Kepler 2.0 release, such as Griddles* ones are in https://code.kepler-project.org/code/kepler/trunk/modules/griddles/ module and Grid FTP and Globus ones are in https://code.kepler-project.org/code/kepler/trunk/modules/globus3.0/ module.

wrote a program to delete the kar files except the following ones. I am waiting people's response.

Authentication zhije guan
kar.454.1
actor.454.1

BlastOne
kar.34.1
actor.238.1

BlastTwo
kar.35.1
actor.239.1

Command Line Exec Edward A. Lee
kar.53.1
actor.185.1

GenBank
kar.131.1
kar.245.1

Invoke Service Efrat Jaeger
kar.164.1
actor.150.1
LocalBlastOne
kar.174.1
actor.246.1

Permanent String Constant Wibke Sudholt
kar.203.1
actor.290.1

Port Line Reader
kar.211.1
actor.254.1

Single Fire Constant Efrat Jaeger
kar.262.1
actor.128.1

SST Data Access Yuhong Xiong
kar.524.1
actor.523.1

ChadsCompositeActor Yuhong Xiong
kar.519.1
actor.518.1

Current Time_sigsit jie liu and edward a lee
kar.571.1
actor.730.1

The files that have "jie liu", "edward a lee" and "Yuhong Xiong" are likely
duplicated file from the ptII repository.

See http://bugzilla.ecoinformatics.org/show_bug.cgi?id=4926

To my knowledge, Jie and Yuhong have never had write access to
the kepler repository. At best, they would be contributors.
I've never seen Edward write his name in lower case, so I suspect
this is also a cut and paste problem.

FWIW - the Ptolemy coding style is that if code is copied, then
the previous authors should be listed as contributors:

@author Tycho Brahe, based on The Almagest by Claudius Ptolemaeus

So, to track these down, you would need to look at the version control info.

CommandLineExec has a long and tortured history. I believe we
merged the Kepler version and the Ptolemy version and came up with
a version that is in the Ptolemy tree, but you'd have to look at
Kepler-dev.

Their names are listed for these KAR files because they were in the documentation of the uploaded kar. It may be that these kar files were just used in testing the kar upload, and thus the documentation didn't change, or if the actors were slightly specialized (i.e., a new port or parameter) and then uploaded, in which case the author may not feel that they deserved authorship, and left the documentation as it was. Most likely, the uploader didn't realize that they should or could change the author field when uploading a specialized kar. In any case, we should look carefully at these and determine if and how they are specialized beyond the original actors. If they are no different, then we can probably just get rid of them. 'ChadsCompositeActor' strikes me as very likely just for testing the upload system.

I tried to open those actors downloaded from the repository on Kepler 1.0. But none of them succeeded. I talked with Chad. Chad said he got a task to upload the actors in kepler library. But we didn't test it.

Since none of them works, I will go ahead to delete those actors. Before I do this, I will backup the kepler database again.

I backed up the kepler database as kepler_backup_05042010 under /var/lib/pgsql/backups. Then the left actors were deleted too.

Original Bugzilla ID was 4928