Friday Mar 23, 2007

OGB Elections: Contributors and Voters

After some discussions on #opensolaris IRC today, I downloaded the lists of contributors from poll.opensolaris.org and ran some stats. Each OpenSolaris community was asked to name both Contributors and Core Contributors over the months leading up to the election - the Core Contributors are those eligible to vote. By my count, there's 268 unique Core Contributors across the communities.

First, the breakdown of Contributors by community - “unique” refers to those in each category which are not listed as either a contributor or core contributor by any other community:

CommunityCore ContributorsContributors
TotalUniqueTotalUnique
Academic and Research 3 2 3 1
Appliances 0 0 0 0
Approachability 1 0 14 8
Architecture Process and Tools 9 5 4717
BrandZ 4 1 4 3
Chinese Users 0 0 0 0
Community Advisory Board (CAB)14 8 3 0
DTrace 5 1 0 0
Databases 0 0 0 0
Desktop 8 8 6 4
Device Drivers 0 0 0 0
Documentation 5 2 12 8
Fault Management 6 1 3 0
GNU Solaris 0 0 0 0
Games on OpenSolaris 0 0 0 0
HPC Developer 0 0 0 0
Immigrants 2 0 1 1
Installation and Packaging10 6 4 1
Internationalization and Localization 5 2 11 8
Laptop11 4 0 0
Marketing 6 0 3 1
CommunityCore ContributorsContributors
TotalUniqueTotalUnique
Modular Debugger (MDB) 5 0 0 0
NFS 3 2 0 0
Networking 8 2 15 2
OS/Net (ON)4816 8 0
Observability 0 0 0 0
OpenSolaris Printing 0 0 0 0
Performance 8 4 1 0
PowerPC 4 2 11 5
Security 8 4 4 0
Service Management Facility (smf(5)) 8 1 16 8
Solaris Volume Manager 0 0 0 0
Storage 0 0 0 0
Systems Administrators 0 0 0 0
Testing 9 5 2 1
Tools22 8 20 6
Unix File Systems (UFS) 0 0 0 0
User Groups8866 1 1
X Window System 3 3 0 0
Xen15 7 1 0
ZFS14 8 1 0
Zones 7 2 3 1

As you can see 13 communities have named no contributors, which will probably result in the new OGB taking a close look at them to determine if they are really dormant or just need leaders to step up and organize them.

I also did counts of how many contributors were recognized by multiple communities as either a contributor or core contributor - these aren't necessarily the biggest contributors to OpenSolaris, just those with the widest breadth of contributions that were recognized:

# of Communities# of peopleContributors
7 2Adam H. Leventhal, James D. Carlson
6 2Casper H.S. Dik, Mike Shapiro
5 3Daniel B. Price, Eric Schrock, Stephen Hahn
4 9Bart Smaalders, Ben Rockwood, Bill Sommerfeld, Dan McDonald, Darren Reed, Kais Belgaied, Keith M. Wesolowski, Peter Memishian, Richard Lowe
3 25
2 68
1245

If you're active in one of the communities and didn't get named as a contributor, it's too late for this election, but for future status you should ask the leaders of the community what criteria they used to choose the contributors and how you can be considered for recognition.

If you were named a core contributor and haven't voted yet - hurry up! You've got a little less than 72 hours left, and while we've just passed 100 votes cast, approval of the constitution requires 50% of the eligible voters, so we need at least 30 more (and that's assuming most people voted yes on ratification). Instructions on voting are at http://opensolaris.org/os/project/website/poll_instructions/. Two tips from debugging problems reported by a pair of voters today:

  • The machine name is poll.opensolaris.org, not polls. Four engineers on IRC could not figure out why none of us could ssh to polls.opensolaris.org and were just getting mysterious connection closed messages, until stevel piped up that we were all misspelling the hostname and hadn't noticed, because the ssh/socks5 proxies we were all using didn't actually report hostname unknown errors.
  • If you copied and pasted your ssh key from a terminal window that caused line breaks to be inserted in the middle of all the ascii hex dump portion of the key, remove them - the web form allows you to put them in and doesn't warn you that it's invalid. (Again, thanks to stevel for coming to the rescue to solve that user's problem.)

Sunday Mar 11, 2007

OGB Elections: Proper Positioning

I haven't had a chance to write full position papers like some of the other candidates, so for my positions, you'll mainly have to rely on these answers I wrote to some of the questions submitted via e-mail to the candidates:

I did vote in the poll being used to test the OpenSolaris elections system, so will share my vote there, and explain some of my choices to help further show my positions on the issues facing the community.

The poll asked (and the order of these is the one I was shown - the order is randomized for each voter to reduce bias if people just start voting "1 2 3 4..."):

QUESTION 1.1: ("Priorities") Which of the following items, presented
in a randomized order, should be prioritized by the Governing Board in
order to promote OpenSolaris and increase its developer mindshare?

1 - Deploy a public defect management system ('Defect_Mgmt')
2 - Remove inactive Communities or Projects ('Remove_Inactive')
3 - Provide an x86/AMD64-based kernel/project build facility
    ('X64_Build')
4 - Deploy a public Request To Integrate (RTI) system ('RTI')
5 - Provide a SPARC-based kernel/project build facility
    ('SPARC_Build')
6 - Create an infrastructure project to run opensolaris.org
    ('Infrastructure')
7 - Replace or remove Jive forum interface ('No_Jive')
8 - Reorganize the existing Community/Project organization
    ('Reorganize_Community')
9 - Replace opensolaris.org tools with an open source CMS or wiki
    ('New_CMS')
10 - Deploy a public wiki on opensolaris.org ('Wiki')
11 - Deploy a public code review facility on opensolaris.org
    ('Code_Review')
12 - Eliminate reliability issues with opensolaris.org web pages
    ('Reliability')

My vote was:

RESPONSE 1.1 --> 1 6 9 8 4 5 3 11 10 2 7
VERIFY:  Your ballot is:
Priorities: [Defect_Mgmt Infrastructure New_CMS Reorganize_Community
 RTI SPARC_Build X64_Build Code_Review Wiki Remove_Inactive No_Jive ]

So why these?

Deploy a public defect management system

Right now bugs.opensolaris.org is not truly a two-way communication system - for people outside Sun, it's Write Once Read Many (WORM) - they can file a bug, search for bugs, and read bugs, but not update them or have them assigned to them to work on. Additionally many of the bugs and bug fields are still restricted to internal users only, though that's a harder nut to crack due to various legal and customer privacy issues.

This means projects with external contributors are at a gross disadvantage, and has led to projects such as ksh93 using a separate bug tracking system (Steve Lau graciously set up a bugzilla for them on his system), which segregates their bugs from the rest of the community, so bugs.os.o no longer has all the relevant data. This needs to be solved before non-Sun employees can be first-class members of the development community.

Create an infrastructure project to run opensolaris.org

Right now, too many improvements are stalled because the team running opensolaris.org can't keep up with the everything they need to do to implement them. Open it up now, let the community help you, and let us all move forward.

Replace opensolaris.org tools with an open source CMS or wiki

The current infrastructure is customized and weird - Project & Community Leaders are the only ones who can edit pages in their areas, so to add someone to edit anything, even just one project subpage, they have to be made a leader, even if they're not doing any leadership. The syntax is also strange and confusing to those used to just about any other wiki-like system in the world.

Reorganize the existing Community/Project organization

This has been discussed to death elsewhere already, and a big part is just cleaning up the original set of communities created before projects were done. The election confused many people since those doing the most work (projects) are not given voting rights, while those just interested in the technology (communities) are.

Deploy a public Request To Integrate (RTI) system

This is specific to the ON family of consolidations that use the RTI system (Desktop consolidations, including X, don't - I don't know about all the others), but once the code repository is outside, this will be the biggest obstacle to letting non-Sun people commit code on their own without any internal sponsor.

Provide a SPARC-based kernel/project build facility
Provide an x86/AMD64-based kernel/project build facility

SPARC is higher priority than x86, since more people have x86 system access on their own, but commits to most OpenSolaris consolidations require you to build and test on both platforms first, and that's hard to do if you aren't working at Sun or another place with both kinds of machines around and installed with the build environment.

Deploy a public code review facility on opensolaris.org
Deploy a public wiki on opensolaris.org

For both of these, while it would be nice to have on opensolaris.org, we've got community ones already that work fine - for instance, I'd be happy just making wiki.opensolaris.org redirect to http://www.genunix.org/wiki/ and declaring that one done.

Remove inactive Communities or Projects

These clutter up crowded pages and cause confusion, but don't stop any real work from being done.

Replace or remove Jive forum interface

Again, as much as it annoys me, it's not preventing real work, and can be mostly ignored if you truly hate it. (It would be nice if it didn't break threads and cross-posting and such for people who use the mailing lists instead.)

Eliminate reliability issues with opensolaris.org web pages

As much as I was tempted to vote for this just to tweak stevel, I just haven't seen enough of a reliability issue to declare this a major problem that needs attention, so I just left it off my choices completely.

Saturday Mar 10, 2007

OGB Elections: Biography of a Candidate

As you may have seen, I am a candidate for the OpenSolaris Governing Board, and as the other candidates are doing, I'm writing a bio to give background on myself. I've previously posted a more personal summary of my background, so I'll focus here on topics I believe are more relevant to being an OGB candidate. If you have no interest in the OGB elections, or in long rambling non-linear self-promoting stories, you can just move ahead to the next entry in your RSS feed reader now.

First the basics that all the candidates are stating: As you may have guessed from the URL this is posted at, Sun Microsystems, Inc. has been my employer for the past 8 years, but I have never built my own kernel. (I have designed, implemented, and integrated my own DTrace provider though.)

While attending the University of California, Berkeley I became involved in the running of Berkeley's Open Computing Facility, a somewhat unusual student group who provided access to Unix workstations and servers in a time when that was hard to come by for a non-engineering student. When I first started volunteering there in 1991, it was the only place most students on campus could get e-mail and Usenet access. Later it became the first place most students could host web pages, and continued to evolve over time. It was run by a completely volunteer student organization, in which I served first as a member of the system administration staff (their Apollo Domain/OS workstations were the first Unix machines I got root on, and the first time I got to have the account name alanc instead of just a class account name like cs60a-ej). I later joined the OCF's Board of Directors, who set the policies for the machines, and served a semester each as Site Manager (head sysadmin) and General Manager (head of the organization).

It was also at the OCF that I first started contributing to Open Source projects. My first major contributions were to the Gopher client/server software for Unix platforms from the University of Michigan, an early precursor to the world wide web. Of course, my most notable open source work has been through my job at Sun, contributing to the X Window System software that forms the foundation of all major Unix and Linux GUIs today. I participated in the old closed-membership industry consortium X.Org Group on behalf of Sun, and contributed to their irregular public code drops of the very closed development tree, and also contributed some of our code to the XFree86 Project, including the support I wrote for IPv6 in X11. When we were looking for ways to revitalize the stagnant X.Org Group at the same time that a group of current and former XFree86 developers were looking to create a more open development model to replace XFree86, I got to be one of the developers in at the very beginning of the transformation of both groups into the X.Org Foundation and it's Xorg software releases. I was elected to the short-lived X.Org Architecture Board, and was the release manager for the X11R6.9 release and one of the leads of the modularization project that resulted in X11R7.0. Through this I've experienced first hand how much more vibrant and productive a fully open development community can be than one in which a small core team holds the leash to all access, and all changes have to be funneled through them. (OpenSolaris isn't quite as bad, as there are several dozen “request sponsors” who can check in code from community members, but we still see requests that take much longer than they should because sponsors simply get too busy and become an unnecessary bottleneck. We've got several external contributors who have clearly proven they deserve commit access, but we haven't gotten the master repositories for anything besides JDS outside the Sun firewall yet so they can exercise that.)

When I took some time off from college, the Unix administration and end-user communication skills I got at the OCF, combined with the connections I had from that community, got me a position as a contractor in Sun's technical support center (then in Mountain View, California) answering Solaris system administration questions. Since I was more familiar then with other forms of Unix, and already had been addicted to Usenet for years, I looked for help in learning the Solaris specific methods to the comp.unix.solaris Usenet newsgroup, and returned the help I got by answering the questions that I could. (And even though I haven't posted as much lately, I just noticed Google still shows me in the top ten all-time posters to comp.unix.solaris, under my alum.calberkeley.org address.) It was through comp.unix.solaris that I first participated in the Solaris community, and first came in contact with current OGB members Rich Teer and Casper Dik, and of course, ksh93 for OpenSolaris project leader Roland Mainz.

After joining Sun I also started participating in the Solaris x86 YahooGroups mailing list, sticking with it even during the dark days in which Sun had declared Solaris 9 for x86 “indefinitely delayed.” After the “Secret Six” from that group convinced Sun to restore Solaris x86, one of the projects Sun started to restore relationship with the community was monthly meetings with a “Cabal” of community leaders. When we decided to ship the Xorg server in Solaris, I was asked to come to one of these meetings to talk about our plans, and given my involvement in the community already, was invited to become a permanent part of the group. That group didn't last long, as it ended when the participants, including me, became some of the original members of the OpenSolaris Pilot Program, so I've been part of the OpenSolaris community since the very beginning. Since the public launch of OpenSolaris I've served as a leader of the OpenSolaris X Window System Community and the representative of the X Consolidation to the OpenSolaris Program Team. (The OpenSolaris.org forum software says I've posted 1,190 times to OpenSolaris forums since the launch, but I don't know how many of those are duplicates from cross posts to multiple forums.)

At Sun, I've also served for about 6 years now on the Desktop C-Team, which oversees the various Desktop consolidations (X, GNOME, Mozilla, CDE, etc.) - approving integration of new projects, setting and enforcing various policies, and coordinating joint projects with other consolidations. For the past three years I've been part of the Layered Software Architecture Review Committee, first as an intern, and now as a full member. LSARC is the ARC which reviews software above the core platform - including the desktop environments, web browsers, compilers and developer tools, and various management tools. X itself is usually straddling the boundary between LSARC and the core platform ARC (PSARC) - cases interfacing with the hardware or kernel go to PSARC (including most updates to the X servers themselves), while client libraries that are mostly used by the desktop environments go to LSARC, so that they fit into the larger picture of those desktops. As such, even though I'm part of LSARC, from time to time I submit cases to PSARC and participate in relevant PSARC reviews.

So that's the long boring story of how I got to where am I today (if you've made it this far, thanks for sticking through it, and be glad that I didn't make it even longer and more boring with recounting my experience as editor of the high school newspaper or how bad I was at memorizing my lines in high school Drama Club productions). Now that we've covered the past, I'll save the topic of the future and my positions on various OGB issues for another post...

About

Engineer working on Oracle Solaris and with the X.Org open source community.

Disclaimer

The views expressed on this blog are my own and do not necessarily reflect the views of Oracle, the X.Org Foundation, or anyone else.

See Also
Follow me on twitter

Search

Categories
Archives
« April 2014
SunMonTueWedThuFriSat
  
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
   
       
Today