Skip to main content
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
[technology-pmc] Re: [Fwd: Re: EPF reviews]

Bjorn,

please see repsonse below after PKR:.

Thanks

Per Kroll
STSM, Manager Methods: RUP / RMC
Project Lead: Eclipse Process Framework
Rational Software, IBM Corp
408-342-3815


Bjorn Freeman-Benson <bjorn.freeman-benson@xxxxxxxxxxx>

09/18/2006 01:47 PM

To
technology-pmc <technology-pmc@xxxxxxxxxxx>
cc
Per Kroll/Cupertino/IBM@IBMUS
Subject
[Fwd: Re: EPF reviews]





Technology PMC members,
Per has sent the attached draft slides for the potential EPF 1.0 Release.  We should all review them before giving our approval for EPF to go ahead with a 1.0 graduation release - we want to avoid a repeat of the ALF situation.

Some specific comments:
  • All slides: the footer is incorrect because the Eclipse Foundation is not publishing this material
  • Slides 2 and 19: according to the commits explorer, there are a maximum of 6 companies (IBM, Telelogic, and four individuals assuming the individuals work for four different companies) instead of the 11 claimed in the slides with 17 committers instead of the 27 listed in the slides. And IBM employees are doing 94% of the commits. This doesn't strike me as a particularly diverse committer community.
    PKR:
    a) 100% of committs for tool component is by IBMers, which was expected. However, there are some indications that this could change and LogicaCMD has a complete working Wiki extension to EPF they want to contribute. We need to resolve some issues such as usage of commercial components. See http://lcmglab.xs4all.nl/wikis/openup/
    13% of committs for content component are from non-IBMers. Over the last 4 months, that figure has been 22%. If you count percentage of people doing committs on the Content component, 50% are non-IBMers. As discussed below, the actual contributions from non-IBMers are much greater than these figures indicate. Also, it was discussed during the Acceptance Review that we should expect to see other companies to primarily be interested in the content component, not the tool component.
    b) The companies that have done committs are: IBM, Telelogic, Xansa (Mark), University British Columbia (Steve), Ambysoft (Scott), and Number Six.
    c) These figures do not fairly reflect actual contributions. Typically a guideline is written up, or content is discussed, and many people are involved in shaping the content. Then one person (often an IBMer) does the actual committ. One reason for this is that you cannot mark edits with change bar within the tool, so it is more practical to have a discussion in bugzilla / epf-dev using Word, and then take the end result and committ. As an example, Chris Armstrong, Armstrong consulting, has commented on 95 bugs, but has made zero committs. <I have a separate issues to deal with that some valuable team members should committ their work and not work through others, but it does not mean that they do not add value or are not involved>
    d) Of the last 200 epf-dev mails, 45% was sent by non-IBMers
e) OpenUP has the following content leads that drive development in each area:
=> Requirements - Chris Sibbald, Telelogic
=> Project Management - Chris Armstrong, Armstrong Process Group
=> Architecture - Mark Cickson, Xansa
=> Development & test - Brian Lyons, NumberSix
=> General (anything else) - Steve , University British Columbia
These are the people that drive the discussions and meetings in each content area. We then have 4-5 IBMers core to the content component. There you got the current kernel of 10 committers for the content team...
  • Slide 19: who are the non-committer contributors?
    PKR: Please see project/IP log, http://www.eclipse.org/epf/project-info/project_log.csv
  • Slide 20: "Several major organizations..." and "Many organizations..." - I'd like to see a concrete list.
    PKR: I could rapidly find 13 companies that add value to EPF. I think this is just the tip of an iceberg... Every time I go to a conference, a new company tells me that they are using EPF one way or the other...Also note that that I do not have all companies listed that are supporters of EPF....
  • Slides 9-11: the outreach to the user community appears to be above average for Eclipse projects
In general, however, I am concerned that the committer community is not diverse because over 90% of the work is still being done by IBM.

PMC members,
Your comments?

Per,
Can you answer the three issues (diversity, non-committer contributors, concrete lists) raised above? Thanks.

- Bjorn

Attachment: EPF 1.0 Review v0.8.ppt
Description: Binary data

Attachment: EPF Value Add.doc
Description: Binary data

Back to the top