The NetBSD Foundation 2010 Annual Report

The NetBSD Foundation held its Annual Group Meeting of all members of the NetBSD Foundation. It took place from 22:00 UTC to midnight on May 21st 2010. This is a report from the presentations of the various groups within the NetBSD project about their past achievements and future Goals. This report is mostly a verbatim version of the text presented in the online meeting, with a few places edited slightly.

Introductory address from Matthias Scheler on behalf of board:

	A welcome to all members of the NetBSD project who made it
	to this meeting at this late or early hour, depending on
	where you are.
	 
	Another year has passed and despite the usual rumor the
	NetBSD project still isn't dead. Quite to the contrary,
	several remarkable things have happened:
	- The NetBSD project took part in Google's Summer of Code 2010
	  for the sixth year in a row. We got awarded six slots and all
	  projects were completed successfully. Five of those projects
	  were already integrated into the source tree.
	 
	  This year the NetBSD project did even better. For our seventh
	  consecutive participation Google assigned nine slots for
	  projects to us. The new projects will officially start
	  on Monday.
	  The board would like to thank all the mentors and
	  administrators who made Google Summer of Code a big success
	  story for the NetBSD project.
	 
	- NetBSD 5.1 was released which is probably our best update
	  release ever with e.g. new Xen features, RAIDframe parity
	  maps, new drivers and a lot of other improvements.
	 
	- While NetBSD 6.0 wasn't finished a lot of interesting work
	  has happened on it. New kernel modules, TLS, Rump, NPF, LVM,
	  MIPS64 and a lot of other features have already been
	  completed. More exiting changes like Xen SMP Support are
	  still being worked on.
	 
	- Thanks to a massive effort led by Antti Kantee NetBSD now
	  has an impressive set of regression tests managed via ATF.
	  Andreas Gustafsson created an infrastructure which
	  periodically builds, installs and tests NetBSD/current on
	  a virtual machine. The results of these automated
	  test runs are available as a web page. They provide a
	  chronological view of the quality and an early warning
	  system for regressions.
	 
	  A lot of  bugs have already been caught early by our new
	  test driven development methodology. This will greatly
	  benefit the quality of future NetBSD releases.
	 
	  Even better, other BSD projects are starting to see the
	  benefits of continuous automated testing.  For example,
	  Junos recently adopted ATF and FreeBSD is interested in
	  doing the same. We are glad to be leading the efforts in
	  this area.
	 
	- The pkgsrc project is thriving. MirBSD is the third BSD
	  operating system which adopted pkgsrc as their primary
	  package platform.
	 
	- The Board of Directors and the Core Group have started to
	  use the foundation's money to fund targeted development.
	  NPF, TLS and Xen SMP Support are examples of such projects.
	  While the underlying process still needs to be improved the
	  results are already encouraging.
	 
	- NetBSD can now be run on Amazon's Web Services. There is
	  ongoing work to make the necessary images available. And
	  the completion of the Xen SMP Support project will greatly
	  improve NetBSD's performance on this platform.
	 
	Of course the project suffered setbacks as well. The
	unexpected death of Martti Kuparinen in June last year was
	a particularly sad moment for all of us.
	 
	The board's plans for the future are:
	 
	- Make sure that the project keeps running smoothly.
	 
	- Improve the process for funded development and spend more
	  of the project's money on improving NetBSD. Suggestions
	  for funded development are very welcome.
	 
	- Provide administrative and financial resources to improve
	  the project's build infrastructure. The board would in
	  particular like to see frequent pkgsrc binary package
	  builds for all supported NetBSD releases.
	 
	 
	At this point we would like to take the opportunity to thank
	Alistair Crooks and David Maxwell who leave the board of
	directors after many years of dedicated service. We would
	also like to welcome our new board members S.P. Zeidler
	and Julio Merino.
  

Alistair Crooks presented some thoughts as leaving president of TNF:

	I'd just like to add a few words
	Just a short few sentences from me. I promise.
	I have been a director of the NetBSD
	Foundation since 2003, when it was
	established.  For the last 6 years, I have
	been the President of the NetBSD Foundation.
	I have now decided that we have many good
	young people who are now developing software
	for NetBSD, and that it is time to give them
	a chance of introducing new ideas, and trying
	new ways to manage things.
	I'd like to thank you all for your support
	over the last 9 years of being a director.
	We wouldn't be where we are now without you,
	the developers, who put it all together, and
	keep it running, and put so much hard effort
	into making NetBSD relevant in the future..
	The last 9 years has seen many, many
	transformations; we will see many more over
	the next few years.  That is good.  If we
	stop developing, we will die - there is no
	comfortable middle age for an operating
	system, and its packaging software, and we
	have to keep moving to stay in the same
	place.  I know we will, and I know we are in
	good hands.
	I'd like to thank you all once again, and
	look forward to the next few years immensely.
  

Martin Husemann presented the report from the membership committee:

	Good evening!
	 
	The main responsibility of membership-exec is handling
	new member applications for the NetBSD project. Since
	the last AGM we handled 16 new developer applications,
	and got 15 new members. Unfortunately at the same time
	15 developers left the project. For the election,
	membership-exec updated the list of active and inactive
	developers. As of today, the NetBSD Foundation, Inc.,
	has 190 active and 9 new members.
	 
	Last year we reported that we had to reject one application.
	The good news: the formerly reject applicant has worked
	with his sponsors and the community and, after another
	application round, now has joined the NetBSD foundation.
	 
	Sadly one developer rejected the invitation after
	his application got accepted. He was intimidated by
	some legalese in the membership agreement, and asked
	for changes. We do believe the same rules should apply
	to everyone (and board concurred), so a personalized
	agreement was rejected.
	 
	The applicant is still contributing code, and we hope
	he will change his mind or find good legal advice (part
	of the problem might be the difficult translation of
	the legalese in the agreement to foreign - and very
	different - law systems).
	 
	The other responsibility of membership-exec is handling
	developer issues and disputes. We are happy to say that
	last year was quiet in that regard.
	 
	The current members of membership-exec are:
	  * Christos Zoulas <christos>
	  * Ken Hornstein <kenh>
	  * Martin Husemann <martin>
	  * Matthias Scheler <tron>
	  * Lex Wennmacher <wennmach>
	  * Thomas Klausner <wiz>
	 
	That is all from us, thank you.
  

Christos Zoulas presented the report from the finance committee:

	Financially 2010 was a robust year with moderate
	income and expenses. Most of our income came from
	a 5 large donations ($35,500), but we've received
	many (123) smaller ones for a total of $11,759.35.
	We again spent less ($16,650.58) than we made
	($47,493.24), adding to our total balance ($223,892.72)
	a good $30,842.69. Our donations are still balanced
	well enough for us to maintain our 501(c)(3) status.
	-
	This year board has created a budget for 2011 based
	on the expenses from 2008 and 2009. This budget allows
	for total expenses of $77,541.00 broken down as:
	-
	-   - Standard (hardware, legal, marketing)     $20,447.00
	-   - Consulting                                $42,094.00
	-   - Developer (hardware/docs/xen smp2)        $15,000.00
	-   - Misc (travel, events, t-shirts)           $12,450.00
	-
	The standard category is generous and allows for
	more than $10,000.00 in new hardware and $5,000.00 in
	marketing. The consulting expenses are already committed
	and are for XIP, NPF, TLS, XEN/SMP1 development. We don't
	expect to spend all that in developer expenses, and we
	can decide what to do with the optional ones item by item.
	-
	We have not started a fundraiser for two years now
	so perhaps we should consider doing so after the 6.0
	release. If things keep as they are we can assume
	to have a stable income of approximately $20,000.00
	per year, but our fixed expenses (legal, fees, hardware)
	are a quarter of that and if we are going to keep
	wanting to pay for consulting projects we need to
	do better than that.
	-
	This year I ended up doing all the tax work, which
	is not too much fun. It is fun if you do it for
	a few times, but I've been involved in it for
	many years, so any help is appreciated. The good
	thing is that this year I spent some time building
	on Lex's work on documenting and automating the
	process so it is not going to be too tough in the
	years to come.
	-
	The merchandise sold at Cafepress is down again this
	year to $233.89 (from $258.44 last year and $381.24
	the year before and $696.86 the year before that).
	We need someone to take this over, get some new art
	and products in and promote it better.
	-
	The breakdown of our expenses in 2010 amount to
	$16,650.58 (down from $17,716.84 last year and
	consist of hardware ($468.76), bank and paypal fees
	($445.42), travel reimbursement ($460.40),
	consulting fees ($10,000.00), conference fees
	($1,500.00), and legal expenses to run the
	Foundation ($3,776.00 USD).
	-
	The initial balance on Jan. 1, 2010 was $193,050.06
	and the final balance on Dec. 31, 2010 was $223,892.72.
	-
	I'd like to return to the donations we received in
	2010 and provide you some statistics. As I said,
	the number of individual donations was 123 (down from
	130 last year). Excluding the 5 large donations:
	The average amount donated was $77.94 (down from
	$87.58 last year and $456.00 the year before), while
	the median value was $18.92 (down from $48.25 and $57.36
	last year and the year before, respectively).
	-
	The split of the donations by value is:
	Value                   Number of Donations     USD       
	< 10                                     56     101.05
	>= 10 && < 100                           41     2609.45
	>= 100 && < 1000                         24     6885.28
	>= 1000                                   0     0
	-
	As always the financial results for each year are available
	in http://www.NetBSD.org/foundation/reports/financial/
	-
	that's it for now
  

Alistair Crooks presented the report from the core team:

	The core team consists of agc, christos, chs,
	matt, mrg, and yamt. For most of the year, we
	also consisted of pooka, who recently decided
	to stand down for personal reasons. As you'll
	see from what happened in core-land, pooka's
	influence was inspirational, energetic and
	very effective.
	Core is the technical management for the
	NetBSD project. Some of the achievements we
	have mentioned here are not down to core's
	involvement -- however, they are notable and
	might not get mentioned elsewhere, and they
	are part of NetBSD. We're proud of them, and
	wanted to share them with you.
	A quick review of the year finds these
	highlights:
	+ copying pkgsrc policy on including a
	summary of useful changes when importing code
	into src and xsrc - thanks, mrg
	+ NetBSD/EC2, which has had an interesting
	history, but is now very much a reality -
	jym, jans and others were the main force
	here, and david even facilitated a meeting
	with the FreeBSD developer to fix some
	miscommunication.  See jym's wiki on this:
	http://wiki.NetBSD.org/amazon_ec2/ Our
	virtualization strategy is leading the way
	folks; if you doubt this, then remember that
	FreeBSD are installing a NetBSD dom0 for one
	of their main clusters.
	+ there was a joint core and board meeting
	last august, which helped with a number of
	areas
	+ we gained lua in october, thanks to marc's
	perseverance.
	+ releng released 5.1 in december; a number
	of meetings and email threads have taken
	place about 6.0 - watch this space
	+ the modular/monolithic kernel debate comes
	up again every so often
	+ pooka delivered the tiering system, which
	documents the status quo that we've had for
	the last N years
	+ bounties for bug fixing are another area
	which pooka felt strongly about - we tried
	this year, and it was a great success. Thanks
	to all who participated, and, if you think
	you fixed something and didn't get your
	bounty, please send mail to core.
	+ there has been emphasis placed on funded
	development, and core has helped out on the
	technical side of this, with reviews and
	advice. When you see TLS, Xen SMP, and NPF
	mentioned, these were made possible by the
	funding of the Foundation. The bug bounties
	are another example. If you have more ideas
	for funded projects, please send them to
	board and core. We do not have endless funds,
	and want to fund work that is otherwise too
	large to be done by people in their spare
	time. In particular, core has project managed
	the following projects:
	++ NPF (rmind): This project is now complete,
	the code has been committed to the repository
	and work is beginning to support IPv6 as part
	of the SoC.
	++ TLS (joerg): Mostly done; support for most
	platforms has been committed.
	++ XENmp (cherry): In progress, 50% done.
	+ arguably foremost of pooka's achievements
	was the concentration on testing using jmmv's
	atf framework. Others are now following suit,
	and we have all seen how easy it is to notice
	regressions when they happen. Personally, I
	must admit to some skepticism when this was
	first suggested, as we'd tried the same thing
	at Amdahl in 1997/1998, and the results there
	were negligible. I was happy to be proved
	wrong by pooka, and think that he has
	achieved many things in his short time in
	core - he should be happy, and I am very
	thankful to him for so many things.
	+ we have seen some interesting new ports
	arrive in-tree, including the mips simulator
	from Microsoft, and other MIPS and arm ones.
	+ we are likely to see chuq's work on gcc 4.5
	in src/external soon, and joerg has been
	working on clang/llvm support, and plunky on
	pcc. Great work, guys!
	+ together with releng, we have been planning
	for 6.0 -- while it has not materialised just
	yet, the automated testing that is being done
	means that once we are ready to branch, we
	believe that the release process will go much
	more quickly than before.
	In summary, we believe that core has been
	more responsive to email than in previous
	years, that we've seen a large amount of
	innovation, and improvement from the whole
	community. Please keep this coming.
  

Soren Jacobsen presented the report from the releng team:

	We are:
	  abs, agc, bouyer, he, heas, jdc, jmc, martin, mkirby,
	  msaitoh, ober, phil, riz, sborrill, snj
	#
	Here's a brief summary of the past year:
	- Released 5.1 in November.  Good.
	- Processed approximately a zillion pullup requests.  Good.
	- Got back to using all of our build cluster's muscle.  Good.
	  (Thanks, admins!)
	- Didn't put out any 5.[01].x security/critical releases.  Bad.
	- Still didn't put out any further releases for 4.0.  Bad.
	#
	A word on security/critical releases: they don't work.  When
	we moved to our new x.y.z numbering scheme, we thought we
	could push out security/critical releases quickly and easily,
	but experience shows that this is much easier said than done.
	There is a lot of hassle surrounding releases, whether
	the release is major, minor, or teeny.  Having these
	security/critical branches is great, but we need to start
	thinking about faster ways to get security fixes to our users.
	In short, it would be great if you all could consider this a
	kickstart to yet another discussion^Wflame war about binary
	updates.
	#
	Looking forward, we need to branch 6.0 soon.  There are a
	couple things we'd like to see hit the tree first, but we're
	getting very close.
	#
	I'd like to close with a couple gentle reminders about pullup
	requests.
	First, please remember that changes submitted for pullup
	should be grouped logically.  As an example, if you update a
	driver to support a new chip, and the changes were committed
	to HEAD in multiple steps, all of these changes should be
	submitted in a single pullup request.  That includes any
	documentation updates.
	Second, please remember that security/critical branches
	(e.g., netbsd-5-1) are for security fixes and critical
	bugfixes only.  In other words, if your pullup request is not
	clearly marked as a security fix, your request should include
	a justification of why the change should be applied to a
	security/critical branch.
	#
	That's all I've got.  Thanks again to all of you developers
	who take the time to think of our release branches and submit
	pullup requests.
  

Alistair Crooks presented the report from the security-officers:

	The security-* role has been reorganized
	lately.  The NetBSD security officers are now
	the people who have the security officer PGP
	key and handle communication with "the
	outside". Please remember not to use the old
	1024 bit key for communication with them any
	more, but to use the new 4096 bit key.
	Current security officers are agc, christos
	and tonnerre
	security-alert is a list consisting of
	security-officer, core and david. Its role is
	to see that incoming issues get a reaction.
	security-team, which existed previously as
	well, is supposed to do the body of the work,
	adopting an issue, finding (someone to
	create) a fix, writing an advisory if
	appropriate, and correcting each others
	advisory drafts before handing them over to
	security-officer (who gives it a final sanity
	check and publishes it).
	security-team has now a rotation, martin and
	spz are currently on it.  The main task of
	the people on rotation is to make sure issues
	do not get lost (ie, be reminder service).
	security-team currently are billc, cyber,
	drochner, joerg, martin, mouse and spz.
	security-officer and core are also members of
	security-team by function inheritance.
	Thanks to all the people who were active on
	security-* in the past and have now moved to
	other tasks, especially dan and jnemeth.
	releng plays a vital role in keeping our
	release branches safe, thanks!
	There have been 15 security advisories since
	the last meeting.  Advisories get written for
	security issues that are not a local DoS nor
	a "root can shoot themselves in the foot" and
	are present in a release branch.
	We currently have 48 open tickets where there
	is no conclusion either if that is a
	vulnerability in the first place, or if the
	problem is fixed.  We currently only have 5
	tickets younger than a year, of which three
	are likely to be closed with 'not a
	vulnerability'.
	This is likely not due to there being no
	issues, but due to security-officer /
	security-alert / security-team not being made
	aware of them. If you fix something, please
	don't assume that it will be picked up
	automatically.
	security-team could use additional members of
	a more diverse subject coverage who actually
	have time to dig through old issues, and to
	adopt new ones.
  

Amitai Schlair presented the report from the pkgsrc project management committee:

	If pkgsrc did not exist, it would be necessary to
	invent it. Fortunately, it exists. Therefore we
	merely need to continually reinvent it.
	        -
	We've made our usual efforts to update packages,
	address security issues, and maintain stable and
	useful quarterly branches. We'd particularly like
	to thank the people who do all the addressing and
	maintaining; you'll be hearing from them next. We've
	also managed to hoodwink a handful of new folks into
	becoming developers, which does wonders for our
	ability to make the usual efforts.
	        -
	pkgsrc's infrastructure sets it apart from other
	package systems. Our team is relatively small, but
	that infrastructure multiplies our total work by a
	large factor -- and for a package system, whose job
	it is to reduce sysadmins' work by a large factor,
	it lets us tell a compelling story. Our teamwork
	makes it possible for anyone to manage thousands of
	third-party applications on a variety of Unix systems
	and in a variety of real-world scenarios. Like all
	pkgsrc developers, I'm also a pkgsrc user. At work
	I'm stuck with Linux and no root access, but so what?
	I get to keep my package-management habits, I get
	to easily install the tools I need, and I get to do
	my work much more productively as a result. pkgsrc
	doesn't judge, it just works.
	        -
	Modernizing the infrastructure that makes this
	possible is a difficult and ongoing task. Some steps
	that are being taken:
	        -
	* The wrapper scripts aren't far from being replaced
	  with more readable and performant C implementations.
	        -
	* Packages will soon be required to support DESTDIR-style
	  staged installation, conferring a number of practical
	  advantages.
	        -
	pkgsrcCon usually falls in the spring, and those of
	us who attend find that it reenergizes our thinking
	about where pkgsrc needs to go. We seem to have
	missed the chance for this year, so let's please
	start making plans for next. Any volunteers to host?
	        -
	This completes pkgsrc-pmc's presentation. Now for
	the presentations from the pkgsrc groups that actually
	do important work. Back to you, moderator.
  

Matthias Scheler presented the report from the pkgsrc releng team:

	Greetings, programs!
	 
	at the end of every quarter the pkgsrc developers create a
	new stable branch within "pkgsrc". This process was started
	a few years ago to provide a reasonably stable platform for
	using the NetBSD package collection on production systems.
	 
	The maintenance of these branches is handled by the
	Pkgsrc Release Engineering Team. Based on tickets submitted
	by pkgsrc developers the team integrates changes from the
	HEAD branch or applies patches to fix critical bugs, close
	security vulnerabilities or address build and
	packaging problems.
	 
	The current team members are:
	 * Fredrik Pettai <pettai>
	 * Tyler R. Retzlaff <rtr>
	 * Lubomir Sedlacik <salo>
	 * Steven Drake <sbd>
	 * Eric Schnoebelen <schnoebe>
	 * S.P. Zeidler <spz>
	 * Matthias Scheler <tron>
	 
	During the last year Geert Hendrickx left the team after
	many years of service. The team would like to thank him
	for all his hard work in the past. Fortunately Fredrik Pettai,
	Steven Drake and Eric Schnoebelen joined the team which now
	has more active members than in the past few years.
	 
	The development of a new script ("pkgsrc-pullup.py" available
	in "localsrc") also helped to speed up the handling of
	pullup requests.
	 
	Thanks to the tireless efforts of the admin team the NetBSD
	project's own pkgsrc build infrastructure now covers
	NetBSD/amd64 and NetBSD/i386 for two releases (NetBSD 4.0.*
	and 5.0.*).
	 
	One problem however remains: a lot of pkgsrc developers
	forget to submit pullup requests. Without pullup requests
	important bug fixes especially for security vulnerabilities
	are often missing from the stable branch.
	 
	The team's list of wishes for the future is:
	- A few more active volunteers for the team
	- A stronger commitment by the pkgsrc developers to keep
	  the stable branches in a working and secure state.
	- Better (faster, more releases) build infrastructure for
	  binary packages
	 
	We would like to take this opportunity to thank all
	the developers who make the effort and submit pullup
	requests, especially Takahiro Kambe. We couldn't do our
	job without your support.
	 
	Finally some statistics:
	- 64 tickets were pulled up into the pkgsrc-2010Q1 branch.
	- 53 tickets were pulled up into the pkgsrc-2010Q2 branch.
	- 68 tickets were pulled up into the pkgsrc-2010Q4 branch.
	- 37 tickets were pulled up into the pkgsrc-2011Q1 branch
	  so far.
	 
	The list of packages for which the most pullup requests
	were submitted:
	- net/wireshark (11 pullups)
	- devel/xulrunner, lang/php53 (8 pullups each)
	- www/firefox (7 pullups)
	- www/typo3, databases/mysql51-server (6 pullups each)
	 
	End of File
  

Tim Zingelman presented the report from the pkgsrc security team:

	The mission of the pkgsrc Security Team is to ensure that
	the ever-growing ecosystem of third party software is
	either safe to use or at least be sure people are aware of
	the known vulnerabilities.
	Our members monitor publicly available vulnerability
	feeds such as Secunia, CVE and bugtraq. We aggregate
	received advisories believed to impact pkgsrc into the
	pkgsrc vulnerability list and notify individual package
	MAINTAINERs. When time allows we locate and commit patches
	to fix the vulnerabilities. What follows is a brief
	summary of last years' activities.
	-
	Our ticket handling crew is down to four and could use
	a few more folks to handle the one week rotations.
	Currently handling tickets are:
	- OBATA Akio
	- Guillaume Lasmayous
	- Fredrik Pettai
	- Tim Zingelman
	The other current members of the team are:
	Mihai Chelaru, Alistair G. Crooks, Ulrich Habel,
	Thomas Klausner, Tonnerre Lombard, Tobias Nygren,
	Adrian Portelli, and Joerg Sonnenberger.
	-
	The year in numbers:
	Since the last meeting, the vulnerability list grew by
	697 entries, for a total of 5253 known vulnerabilities.
	The ticket queue received 4394 new tickets of which
	over 99% have been acted upon.
	The current count of vulnerable packages in pkgsrc-current
	is 218.  See the periodic email to packages@NetBSD.org for
	the list.  We can always use help locating and committing
	security patches, in particular for the many (127) of
	these that are maintained by pkgsrc-users.
	-
	That's all from us, but before I hand off I'd like to
	encourage all developers to help us keep the vulnerability
	list up-to-date. If you become aware of a security issue
	or perform a security update in pkgsrc, do note it in the
	list. Developers don't need any special privilege for this.
	The team periodically signs off and uploads new revisions
	to ftp.NetBSD.org.  (If you are more comfortable asking
	us to make the edits, just provide the info.)
	-
	Thanks for listening! Over to the moderator.
  

Jan Schaumann presented the report from the www team:

	www apparently did not put together a presentation for this
	year's meeting, so I hastily made up this summary while the
	meeting was in progress.  It reminded me of school, somehow.
	Anyway, this is (a) not particularly troublesome and (b)
	somewhat indicative of the current status of the www team
	and our website.
	 
	(a) - because you can just look at the archives from last
	year and read that report.  Nothing much has changed.  I'm
	willing to bet that even the support lineup is the same.
	 
	We still just handle queries coming in to www@ and mirrors@
	and perform the various minor updates to the website.  It's
	not very difficult or demanding work, but it still needs to
	be done and I thank all participants in the rotation.
	 
	(b) - well... our website is... not all that good.  Our
	documentation is, in general, not kept up to date very well.
	News announcements are now made mostly via the Blog (to
	which I'm afraid I can't speak -- schmonz@ may be able to
	add a quick summary at the end), which unfortunately has the
	side effect of leaving our "news" page look rather stale.
	 
	Our framework continues to be the somewhat fragile and
	cumbersome translation of XML documents into HTML.  So far
	nobody has undertaken the herculean effort required to
	completely restructure all of htdocs.  For internal docs,
	though, we now have a Wiki, which is used to some degree, I
	believe.
	 
	mirrors@ finally is another orphan child.  We have a long
	list of mirrors worldwide, and we have a script that checks
	them regularly and attempts to ensure they're up to date.
	The majority are not.  Unfortunately we do not seem to have
	the manpower to follow up on this.
	 
	So while this sounds all very negative, I'd like to look at
	it this way: we have so much room to improve that things are
	only going to get better if we find volunteers with time to
	help; if we don't, well, then the website will still chug
	along as it did throughout 2010, which isn't the end of the
	world either.
	 
	I would like to encourage all developers, however, to
	consider updating htdocs (or requesting updates from
	provided plain text documentation) whenever they add new
	features and to contribute to the blog / news section.
	 
	My apologies for such a cynical take on the matter and the
	hastily slapped together report.  I was depressed not to
	have gotten raptured and have a broken thumb to boot.
	 
	Back to the moderator...
  

David Holland presented the report from Gnats:

  This year has been fairly routine for bug handling. We've again
  managed to hold the line and avoid going over 5000 open PRs.
  Here's the summary report since the last AGM:
   
  GNATS statistics for Apr 22 2010 to May 21 2011:
   
  New PRs this year: 1792, of which 769 are still open.
  Closed PRs this year: 1745. Net change: +47. 
  Total PRs touched this year: 2714.
  Oldest PR touched this year: 1722.
  Oldest open PR: 698; PR ignored for the longest: 3019.
   
  Total number open: 4887
  (In case anyone is wondering, PR 698 wants umount to support
  running external per-fs umount programs (mount has since ~forever)
  and PR 3019 is a complaint about NFS export attributes being
  per-volume instead of per-directory.)
   
  Here's a graph showing the last year in GNATS in more detail:
   
  Open PR count, April 2010 through May 2011, by week:
   
                                                      *         4950
                                             *   **********    
          *              ******            ********************
     **  **           *****************************************
     *******       * ******************************************
     ********************************************************** 4770
   
  The big drop last summer reflects a lot of work put in; much of that
  was done by mrg and obache, so a big thank you to them. The increase
  afterwards was (I believe) largely prompted by the 5.1 release; it's
  good that we got past that without a large net uptick in PRs.
   
  Even though this graph looks like it shows an upward trend, it's
  actually quite close to flat overall. And while we're up by 47 from
  last year, that isn't a lot. If we look at the last six years, it
  shows that recently we've done a pretty good job at stemming the
  usual long term upward trend.
   
  Open PR count, May 2005 through May 2011, by month:
   
                                                   ****                    5000
                                             ***********    **  *** ******
                                            ******************************
                *                         ********************************
               **                     ************************************
              ***              **    *************************************
             *****           **** ****************************************
           *******    *  *************************************************
          ********    ****************************************************
      ********************************************************************
   ***********************************************************************
 ************************************************************************* 3800
  
  Now we just need to increase speed a little to establish a sustained
  downward trend. :-)
   
  Just for fun, here are not the top five PR closers (since that's
  gnats administrivia) but the top five people who've committed to PRs
  over the past year:
    50  christos@NetBSD.org
    59  drochner@NetBSD.org
    91  dholland@NetBSD.org
   114  wiz@NetBSD.org
   169  obache@NetBSD.org
   
  And last but not in any way least, here are the top five people
  who've processed PR-related pullups for releng:
    12  spz@NetBSD.org
    17  snj@NetBSD.org
    27  tron@NetBSD.org
    28  bouyer@NetBSD.org
    46  riz@NetBSD.org
   
  (These figures count commit messages that make it to gnats, which is
  most but not always all of them, so the numbers aren't absolutely
  perfect.)
   
  Thanks to everyone for helping out.
  

S.P.Zeidler presented the report from admins:

	good $TZ all
	 
	- 162 Tickets resolved since 2010-04-24 (the last meeting)
	- 'lots' package  and some security updates
	- 15 new developers, 16 accounts observered,
	  4 accounts reanimated, 19 closed
	- Many acts of small-scale system administration
	Changes to servers since the last meeting:
	- homeworld is now an ex-build-box with 2GB RAM and SSD disk.
	  It has no swap, so its /tmp is a bit limited; please be mindful
	  of that (and use /var/tmp for larger temporary files).
	- babylon5 has been reanimated and is being prepared to run anita tests
	  it was used to build the stable pkgsrc branch on 5.1 amd64 and i386;
	  these builds currently have no server to run on.
	- b6 is back in business
	- cole.NetBSD.de replaced babylon4.NetBSD.org as off-site backup server
	 
	Thanks to tls and riz for their time and blood sacrifices. :}
	 
	We have a new service:
	- homeworld now runs spamassassin to add to the despamming rules,
	  tagging only
	 
	Todo:
	- a new USB-to-serial for the ISC rack needs yet to be bought
	- mollari should take over the rest of the duties of narn.
	- narn and a bunch of special-purpose hosts are still on NetBSD 4
	  and should update or be retired before we retire the branch
	 
	Known issues (most to least annoying):
	- morden suffers from tstile hangs roughly every 10-15 days
	- morden panics (less often since rmind cornered a bug, thanks)
	- blog occasionally suffers from tomcat indisposition
	 
	Other events:
	- power and network outages and maintenances at ISC and NYC
	 
	some statistics for the public servers:
	mail.NetBSD.org: up 34 days
		last reboot for reinstall
	www.NetBSD.org: up 168 days
		last reboot for a kernel security update
	mollari.NetBSD.org: up 114 days
		last reboot for a kernel security update
	ftp.NetBSD.org: up 12 days
		last reboot for tstiles
	anoncvs.NetBSD.org: up 170 days
		last reboot for a kernel security update
	nyftp.NetBSD.org: up 112 days
		last reboot for a kernel security update
	 
	and some non-public ones:
	cvs.NetBSD.org: up 170 days
		last reboot for a kernel security update
	build.NetBSD.org: up 112 days
		last reboot for a kernel security update
	pkgbuild.NetBSD.org: up 112 days
		last reboot for a kernel security update
	 
	Back to the moderator. 
  

Darren Reed presented the report from the marketing committee:

	A quick summary for communications (marketing/www).
	(Thank you to jschauma who gave more details on www
	earlier.)
	    -
	The marketing list is:  david, dsieger, gls, haad,
	imil, mishka, reed, rmind, sarah, uebayasi, and
	weinem.
	    - 
	The weekly www rotation members are: darcy,
	jschauma, reed, and weinem.  (david and zafer also
	helped during the past year.)
	    
	    -
	Remember if you don't want to hassle with htdocs,
	just send an email to the www team and we can do
	the docbook work.
	    -
	NetBSD was represented at some conferences.  TNF
	members gave lectures related to NetBSD at events:
	mbalmer at FOSDEM, jmmv at BSDCan and NYCBSDCon,
	ahoka at AsiaBSDCon, and agc at EuroBSDCon.  I
	gave a presentation about NetBSD at NYCBSDCon (and
	sold a few NetBSD System Manager's Manual books).
	kml and david had a NetBSD booth at SCALE.  NetBSD
	was a sponsor for MeetBSD and NYCBSDCon (who also
	shared back their earnings).  (Sorry if I missed
	some. Please let marketing team know.)
	    -
	Lots of work and articles on new Wiki by schmonz
	and others.
	    -
	Several NetBSD blog articles. Thank you!
	    -
	Consolidated two (of three) Facebook groups.
	    -
	Helped on literature about features in upcoming
	6.0.
	    -
	As for upcoming needs:
	    -
	+ Help with ready-to-use maintained brochures /
	  print literature. (Update what we have.)
	    -
	+ Represent NetBSD with booths or lectures;
	  marketing may coordinate funding of this.
	    -
	+ Updated website (dsieger has design).
	    -
	+ One place for news so blog and wiki feeds are
	  on main page (or main site replaced with wiki).
	    -
	+ Maybe use wiki for blog too? (any good reason
	  to have multiple infrastructures?)
	    -
	+ Continue to write Blog articles.
	    -
	+ Continue to advocate NetBSD publicly and internally
	  at your companies and organizations.
	    -
	TNF members are encouraged to promote NetBSD by
	running booths, giving NetBSD-related presentations
	at conferences and events, or simply by attending
	and representing NetBSD in other ways.
	    -
	If you'd like monetary assistance to help attend
	events as a NetBSD representative, please send
	your request explaining your plan and amount needed
	to marketing for consideration.
	    -
	If you have time to help promote NetBSD, please
	join marketing.
	    -
	That's all for my short marketing update.  ...