Thursday, June 28, 2007

Ottawa Linux Symposium, Day #1

Day #1 at OLS started off nicesly with Jonathan Corbet and his "State of the Linux kernel" talk. I just recently also saw this talk given at the Linux collaboration Summit. Jonathan updated the slides a bit but also included pointers to sessions related to key technologies making their way into the Linux kernel.

I next sat in on Paul Menage's talk on containers. While Paul is clearly a sharp guy, unfortunately, the big picture of his talk and goals weren't super clear in his talk. Paul has the leading implementation for something called containers, but somewhat unfortunately at the momemt, there are several things known as containers. He made an attempt to show that he was integrated with all of them by the end of his talk, although, I think the level of integration with some of them is pretty weak. I think the code and paper are probably better referencdes than his slides & talk.

The next session I attended was near and dear to my heart - I'm a big fan of appliances as the next wave of deployable solutions. However the work that David Lutterkort in Red Hat's Emerging Technologies group was doing was new to me. Of course, being on the program committee, I did have *some* preview of his high level goals, but the tools, such as puppet, that he's working on where completely new to me. He also mentioned Kronolith which I have yet to research. But the talk was good and the technologies are definitely cool for rapid deployments, pre-tested configurations, and easy of management for systems administrators deploying
similar workloads on lots of similar machines.

More to come...

Wednesday, June 27, 2007

LIVE! from the Ottawa Linux Symposium!

Well, it would be live if my wireless cards worked better with Linux. I might be able to fix that. My primary wireless card is an Atheros and I hate it with a passion. It has sucked more useful life from my body than anything else in my standard operating environment (other key examples were early uses of apt-get upgrade on debian, and being an early adopter of Fedora 7).

Anyway, the opening/greeting/mingling location was at a bar called Vineyards in the Byward Market area and, whileonly a small percentage of the registered crowd showed up, the place was
very busy and the number of actual code developers and contributors was quite high. One of the concens going into this year is that the move of the Linux Kernel Summit to September in Cambridge UK would pull a lot of the core developers away and reduce the quality of the
networking opportunities at OLS.

However, the "newer" model of holding key mini-summits has substantially mitigated that issue, if, indeed it would have been an issue. There was a virtualization mini-summit, a power management mini-summit and, I believe, a third mini-summit that were all held prior to OLS. Those mini-summits pulled in deep experts in those areas, and several other development areas were well represented, including ext4 developers, various architecture maintainers & core developers, dm/md developers, storage folks, container folks,

Word from the organizers is that in excess of 660-670 people had registered for OLS this year, with a few stragglers still signing up at the last minute. There seemed to be a larger international component, with more people from Japen that previously, as well as a strong
European contingent.

There appeared to be a large amount of beer and a fair bit of wine consumed at the bar that evening, which seemed to enable a free flow of information and plenty of mingling. I had a chance to talk to people about everything from power management, containers & networking (Dave Miller hates container namespaces!), device mapper and support for EMC multipath hardware, a rework of the Linux sound subsystem, and a number of debates in several of those spaces, just to name a few The night ended a bit early for some when the bar started closing down a full 45 minutes before Ottawa's otherwise early 1 AM bar closure. However, all for the better since a 10 AM start was in the works for Wednesday morning.

Late note: these blogs may be more "as it happens" tomorrow since I think I figure out how to get the latest MadWifi and linux-2.6.22-rc6 to work just a bit ago!

Wednesday, June 13, 2007

End Users and Linux: What Works, What Doesn't?

Moderator: Tim Golden, Bank of America

  • Jeremy Allison, Google
  • Chris M
  • Randy Terbush, CTO of Infrastructure Architecture and Strategy at ADP
What factors led your company to adopt open source?

Randy: Started pressing at ADP about 4 years ago, wanting to provide a common abstraction above all hardware interfaces, driving towards a more utility computing model.

Jeremy: Google gets complete control of their environment. Jeremy likened it to Dr. Who's Tardis - weirder on the inside than the outside.

Chris: Has consulted for around 30 open source companies. Most started with the fundamental belief that open source was the better methodology in a philosophical sense.

Tim: Initially, for BofA - it was about the Money. In the core infrastructure space money was the initial issue although after a bit that changed to a more functional set of concerns. One key is a faster way to bring products to market.

Question: What does Linux have to do to win against other platforms when your company considers new deployments?

Randy: Constantly competing with knowledge and awareness around Solaris, and ensuring that Linux accellerates rapidly enough to compare well against Solaris.

Chris: Linux helps ensure that there is no single vendor lock in in the core portion of the stack for a number of European companies that Chris consults for.

Jeremy: Google is obviously highly Linux, but engineers can choose Linux, Windows, or a Mac box. Google uses Linux "when it makes sense" - it isn't a religious decision.

Tim: Maintaining a strong Linux presence on site is an ongoing challenge. There are constant competitive pressures to consider alternatives and repetitive TCO analyses are common.

Chris: Linux often enters "like a water leak from below" and rises until it gets some visibility - and that visibility is often initially negative from senior folks who are not familiar with Linux. Chris also pointed out the dynamics of departments who have a human desire to grow (or at least remain stable) and Linux plays against that desire by reducing costs - and reducing costs (being successful) implies that you can operate with less budget in most IT shops. This is not exclusive to Linux but a standard trend in deployment of platforms that increase or improve efficiency.

Jeremy: provided an anecdote of selling servers with Samba as a way to combat that thinking - $600 per server (for open source Samba!) helped make people think they were getting monetary value from their cost reductions (using open source).

Question: Why do you like open source?

Jeremy: "Control of our Destiny"

Randy: It's about standardization. They aren't really seeing cost benefit, but the real value is in standardization across the board. Sun is basically giving away Solaris as well, so on a price comparison Linux has no strong lead. Commonality of OS is a real key.

Tim: Went after cost savings, with a full, multi-year TCO assessment to show that value. There is probably a stack of benefits in moving to Linux. Enabling Linux helped bring in a bunch of other open source projects since it paved the way in terms of risk and license assessment, general comfort with open source, etc.

Question: What are the biggest challenges in deploying Linux within your organization?

Tim: This community has been a challenge. Is it a "big C" or "little c" community? Differences of opinion often lead to isoluation of communities and seperation or ostracizing a member of that Community. [The debate between whether various people are part of "the community" has come up a couple of times during the day --gerrit].

Jeremy: in talking with Andrew Morton about this question, the answer seemed to be "The engineers keep fiddling with it!" This winds up being a pain point because things are constantly changing, and not always for a good reason.

Question: If you were Linux God for a day what would you do?

Tim: Make a full ecosystem around SystemTap

jeremy: Make all distros release the same Linux kernel and glibc across the board.

Chris: Make the power management stuff work better. Greg KH: "It works for us." If it doesn't work for you, let us know, send us your laptop and we'll fix it for you. There are two ACPI implementations - one for Linux, one for Windows. ACPI is not power management and not suspend/resume.

Randy: Would like to see a single, unified, certified Linux distribution. Clarified: Ideally, a single source from which Red Hat and SUSE would pull from, where the single source was was the certification point for applications.

Greg KH: One Linux Distro, the Uber Distro: There some very large

Jeremy argues that the LSB is the real answer to that. [I don't believe that - LSB is such a teensy portion of a full distribution that it isn't practical to make an LSB large enough to cover the real issues, although LSB is a good step in the right direction.]

Question: As an end user, are you part of the Linux community?

Randy: I think we are - in part because we support the community with dollars.

Chris: Most companies employ individual contributors today - there are very few programmers living in their parents basement. As such, most companies are in some sense part of the community.

Jeremy: Yes, there are some

James Bottomley: The Linux Foundation now has a mechanism for accepting dollars to help direct those funds towards programs, coding, projects, etc. Projects can be proposed to the Linux Foundation as well for development and focus.

Powered by ScribeFire.

The Legal Protection of Linux - Patents and Licensing, GPLv3 and the Future

Moderator: Andrew Updegrove, Partner, Gesmer Updegrove

  • Mark Radcliffe, Partner; DLA Piper US, LLP
  • Jason Wacha, Vice President, Corp. Affairs; General Counsel, MontaVista Software
  • Karen Copenhaver, Partner; Choate, Hall & Stewart
  • McCoy Smith, Senior Attorney, Intel Corporation
  • Abdy Raissinia, IBM

Mark Radcliffe:  Has been practicing for over 25 years.  Has 40 lawyers working in Intellectual Property issues.  Ran Committee C for GPLv3.  Impressed by the flexibility of the group crafting the GPLv3 and converting the license into something that people will actually use and accept.  This is not the way that licensing is typically done!

Jason:  working with large companies on using Open Source on hand held devices, etc. including building their knowledge and comfort of working with open source, open source licensing and intellectual property concerns.

Abdy:  Referenced OIN, patent attorney for IBM Linux Technology Center

Andy Upgrove:  Worked with the X Consortium ages ago.

Karen:  First ran into GPL in 1994 on a library.  Found it to be "really werid".  Probably too weird.  She asked the developer to use something else, anything else.  The developer over time pointed out that the math library was so highly examined (open source participation) that there was no other viable option.  Karen re-examined and found the GPL to be "really weird".  She then spent some quality time with the license as her first foray into open source and open source licensing.

Question:  How might open source and open standards play together in the future?

Jason:  this will be pushed by commercial forces.  Linux started as professionals playing as amateurs, but now people are getting paid as professionals to work on Linux.  That commercial force will drive the future directions and leverage the deep & broad aspects of the operating system.  Andrew: Do more developers need to be involved in the process?  Should the community be more involved and activist in the standards (and licensing?) process?  Jason:  the community is less clear since it includes individual developers, commercial entities and many others.  Mark:  as more applications and environments mix open source and proprietary applications and uses in a heterogenous environment, their will be more focus on licensing and related issues.

Mark/Andrew:  In the US, corporations are pushing open source, in Europe, developers/linux community are the primary force, in Asia, the government is the primary force.  Addition:  places like Brasil also have a very strong government support situation, where in the US other proprietary concerns are preventing a strong US government push toward open source.

Jeff Liquea - works for LF, also part of debian (which is well known for "amateur laywering"  - speckling of laughter) - question:  what do you think of amateurs writing licenses, giving interpretations of licenses, etc.?

Karen:  Lawyers probably wouldn't have had the creative capacity to have this vision - without those amateurs, at some level the open source movement wouldn't exist.  Lawyers are now trying to help mold the existing work into something that has a solid legal foundation.

McCoy:  This question came up in part during GPLv3, amateurs may not truly capture the goals that they want to achieve.  However, communities have an understanding of their guidelines and are self-policing to some level, which is a good thing.  Of course, there is some potential for problems down the road when amateurs are driving.

Abdy:  More and more, attorneys are actively reviewing these licenses and there is more overview in a legal sense of the new directions.

Jason:  "amateurs" or the developers who have a vision for what they want to do and can set the boundaries work well with lawyers or attorneys to help craft better licensing.

Karen worked with a mock court with Federal Circuit Court Judges as a way to evaluate some of these licenses, and the effort was definitely challenging.  The judges frequently referred back to the definitions of "derived works" because the area is very complex.  There is a definite concern that a bad court case could set a bad precedent.

Mark:  as the amount of revenue pressure from Open Source on proprietary companies increases, the amount of pressure on open source licenses and potential legal challenges is likely to rise.

General comment:  Several pointed out that lawyers mostly don't understand open source and related licenses, and even the panelists took a while to embrace it.

Bernard Golden:  What is the likely uptake of GPLv3?

Karen:  Things look much improved and there looks to be some significant potential for uptake.

Mark:  the user level community is likely to accept it more readily than some communities, and the VC group looks at the license as pretty reasonable based on his experience with VCs.

Karen:  Clearly the uptake will be gradual and thoughtful over time as well.

Mark:  A number of ISVs are going to find GPLv3 more palatable than GPLv2 in many cases.

Abdy:  The process of evolving GPLv2 to GPLv3 and the presence of many large vendors in that upgrade process strongly suggests that open source and related licensing is here for the long term.

Powered by ScribeFire.

How Do We Get More Apps on Linux?

Dan Kohn started with the question:  Is Web 2.0 really important?  Isn't a browser on your desktop good enough?

Web 2.0 should solve a bunch of low hanigng fruit, but phones are not yet addressed in that environment.

Legacy code won't be reengineered as Web 2.0 apps - but many have been ported to Linux already.

Kay Tate - we have seen a lot of legacy apps being ported, but more so the small and medium sized business (SMB) are porting to Linux.

Web 2.0 should solve a bunch of problems but clearly won't be the be-all and end-all for software.  The further you get away from the silicon valley VC environment, the less powerful the Web 2.0 message is.  What will inspire people to port more easily to Linux?  A premise in the question is that Eclipse is part of an answer.

Adobe does have web based version of Adobe Photoshop.

Question:  why does Windows Vista have 3400+ certified apps?

Brian Aker:  Many of those apps are clones of programs which deal with zip files.  Linux has one [actually, I think Linux has as many as Windows, most likely ;-) --gerrit].  New applications should be encouraged to port to Linux primarily.

#2:  Yes, new applications should be ported to Linux first when possible.  Customers look first to solve a business problem, then looks for applications which solve that problem.  We need to focus on new applications which fill the current business needs.

Kay:  Distros are well certified with LSB, etc. so that is where we are doing well.

Mike: How do we get more *applications* that run on Linux - not just programs and products, but focusing more on users like bank tellers and the applications that they need.  We are sort of in the space of crossing the chasm - the easy, early adopters have been convinced, the next step is to cross the chasm.  Using gcc & gdb is obviously not sufficient, so we need a strategy, philosophy, culture change to convince people to change.  We need to provide developers with an environment that provides a full development environment - this is a place where Microsoft has done well.  Eclipse helps in this space although Eclipse is not as ubiquitous today in its adoption as any Microsoft development environment.

Ed  Real Player is small and available, tools are relatively easy and convenient to create media that plays in this environment.  Application writers what to make sure that their applications are ubuiquitous - and the Linux market is still relateively small compared to the Microsoft environment.  Having an application environment like the LSB should help address an ISV concern by showing that their application, once built, will run on many distributions, thus increasing the ubiquity of the application market.

Question:  If I were an ISV who has decided to target Linux, what problems would be challenging to address?

LSB helped consolidate the development environment, so it is good.  One of the challenges is the many installers on the various distributions.  [Dan points out that the LSB has been around for 8 years but the number of applications certified under LSB is still relatively small].

Flash is supported on lots of platforms, now certified on RHEL3/RHEL4, Novell and one or two others, and do minimal acceptance testing on a few other distributions.  However, the support matrix is huge already with old versions of Windows, etc., so adding lots of Linux distributions is hard and expensive today.

Mike:  He goes back to the economics of the problem.  Specifically, the number of Eclipse downloads have been 80% windows, 14% Linux, 3% Mac and hasn't changed in three years.  One good factor is that the distros now include Eclipse, which might mean that the number of Linux uses in the field is actually larger.  So the global question is:  has anyone made any money with LSB applications on Linux?  So what is the ROI model for software which provides incentive for people to port to Linux.

Kay: Most of the ISV's that IBM works with are already on Linux.  However, for those that aren't, being able to explain to them that a single source will run on multiple platforms, the source base will be more stable, and they'll invest less across the board, primarily because of the LSB and IBM's chiphopper program.

Darren:  Pointing people to the LSB site has helped people understand how to simplify their porting environment.  Most people tend to port to one or two distros and the remainder of the distros get left out in the cold.

Kay:  thinks the market has gotten better, but Brian disagrees, primarily because of the differences of kernels in each of the distributions.  The differences between distributions are just enough to cause real bugs in MySQL lately, and things have gotten worse lately.  As an example, linuxthreads is still an ongoing nightmare for MySQL.  Tools market is not a lucrative market, so implementing new tools is not a reasonable business model.

LSB application toolkit manager, version 0.1 was just released, which provides a wrapper and front end to the test infrastructure that has existed for a while.  This should simplify application certification/testing.

Cadence SW - apt and yum are a pain for their software - they are having problems with defect support on the variety of kernels and environments - it is too expensive to support their customers on anything other than a small set of platforms.  So, requested that Eclipse help in automatically building applications that are LSB compliant.  The observation is that Linux is not *one* distribution but 500 different distributions where each one expects to be different and yet expects applications to run.  Cadence is not as interested in the desktop but more interested in the server side and con

Dan Kegel - Picasso would like to be added to the distribution but it is a binary only application.  It would be nice if there were an xdg add-trusted-repository [I have no idea what this means, so I may have transcribed incorrectly --gerrit].  Dan would love to see that Wine improved to the point that most applications could run in wine without a port, yet potentially running slowly.  Then market numbers would help demonstrate the need for a native Linux port.

Gordon Hopper from Motorola:  loves apt and yum, having all these utilities in apt and yum seems to be a bad idea since it excludes all of the binary applications.  But having more of the open source applications become LSB compliant and shake the bugs out of the LSB tools and such would be good. 

Pacific NW startup, Linux ISV - as a startup they are required to make money and thus be profitable.  Answer:  LSB is really for the middle tier - big ISV's throw bodies at the problem and solve it however.  However, the LSB is more likely to address the SMB market.  Application was built on .NET originally, then ported to Mono.  The business model for pure Linux ISV's is hard - the comment made was "it sucks to be you!".    The only two technologies that work that way are .NET/Mono, rcp, wine, (and one other was mentioned by the audience).  The ISV model enabling applications first on Mono and porting to .NET seems to be viable.

Anyway, my summary would be that this session had lots of questions and discussion, but fewer answers than most might have hoped.  Hopefully, LSB helps, but it clearly doesn't solve everything.

Powered by ScribeFire.

Mark Shuttleworth Keynote at Linux Collaboration Summit

I'll skip Mark's intro - he's easy to google (Thawte, Verisign, Space Shuttle, Ubuntu).  He gets the keynote address at the LF Collaboration Summit.

How to collaborate.  Non-traditional folks are starting to hear about Open Source software.  But headlines come from "contests of will" such as Red Hat vs. Novell, MS vs. Ubuntu, etc.  However, innovation works better when people are inspired, when they have an itch to scratch, and you provide an environment for them to scratch their itch.

Community requsts:  pass stuff upstream.  Distro:  we didn't know where to send it (not all projects are well maintained).

Loudest ideas tend to bubble up, look to talk on Poisonous people, how to make sure best ideas win, as opposed to loudest ideas.  We have lots of collaboration tools - wikis, mailing lists, etc, which focus on collaboration within a project.  However, collaboration between projects is weaker.  Translations rarely move upstream.  Can distros create a "star topology" for passing fixes between distros, thus facilitating upstream adoption.

(Battery dying here, time to swtich).

Back now - Mark has been talking about Federating our interactions, with a focus on several facts, including multiple distributions working on the same projects, with similar inputs/bugs/needs, developers working cross projects, e.g. accessibility and translation being places where non-experts contribute to many projects, so many projects on different schedules being collated into distributions, bugs spanning multiple products for a single user-experience flaw, etc.  Mark used the example that creating an account on every bugzilla, sourceforge site, etc. would be prohibitive for any individual who just wanted to contribute translations, for instance.  The model today dos not span multiple projects, although the collaboration models work quite well within individual projects.

Ubuntu used Mozilla and Gnome as models (Mozilla - roadmapping; Gnome, committment to release cycles) as influences.  Each feature develops a micro-community, a branch, etc. in some projects which seems to work well.  IETF has a similar model for creating an idea and building a consensus around that idea.  How do we generally lower the barriers around development and contribution?  Why can't people do an apt-get of the source of a project, create a branch, allowing everyone to develop like the core developers, being able to easily contribute changes back to where they can be reviewed, integrated, etc.

Ubuntu uses bazaar (a brach of arch) for source control.  Matt Mackall has convinced them to look at Mercurial, Git is fairly common, Mark mentioned a couple of others.  However, each source control system raises the barrier for participation by the average developer.  His major point seems to be to advocate a standard mechanism for bug tracking, source management, etc., not by having everyone use the exact same tool set, but instead defining some basing API's etc.

In conclusion, Mark encourages greater collaboration across projects (and no, I didn't state this as well as he did, sigh).  But I, for one, agree.

Powered by ScribeFire.

Enterprise Panel at Linux Foundation

Jim Zemlin asked what the most exciting things are going on now in the Linux space.

Dan Frye took the first answer, focusing on Real Time Linux, working with the Defense industry and various Financial industries, where Linux has taken a clear lead over some of the competition, e.g. Solaris.

Christy Wyatt is focused on the mobility space, with over 3 million mobile users using Linux today.  For Motorola that create excitement and opportunity in the market space.

Chris DiBona is excited about Containers in Linux and resource containment in Linux.

Jim asked how the vendors work with the community and what the opportunities are to work with the community.

Chris talked about how open source usage is very integrated into their workload and how they avoid the redistribution clauses of open source simply because of their typical usage/business model.   Chris also suggested that he could trivially get Speak Up's for Greg KH to do testing based on the accessibility needs brought up in the previous session.

Christy pointed out that Motorola's interactions with the community are broad - 70,000 employees, 1/3rd of which are software developers at an 80 year old company, the culture change is a definite challenge.  However, mobile linux is becoming more defined and is making it easier for Motorola to engage with the communities.

Dan pointed out that IBM uses open source throughout the company, but the LTC's role is to be members of the community - not just "work with the community".  IBM's goal is to remain trusted, valued peers in the community, helping to make Linux better.

Jim asked:  why are you all here?

Of course, three of these are board members who were in for a board meeting the day before, and Chris pointed out that, well, it's here, in my building!  All tongue in cheek, of course.

Dan pointed out that this is where the industry comes together, handling legal, technical, business issues.

Christy pointed out that LF is the center for all things Linux.  She also reiterated that large companies don't want to work *with* the community, but want to learn to be *part of* the community.

I missed the next question, but Chris pointed out that there is more early release hardware available to the community, more interactions early with the community.  There are some people who are combatting the Linux community but the community is also strong enough that those attacks are often ignored.

Christy pointed out that there is no longer concern about whether Linux is a viable choice for the mobile space - with 6 million devices already shipped, that decision is clearly decided.

Dan pointed out that adoption by enterprise of Linux at the "edge" (infrastructure) portions of the business were trivial at most companies - they were less regulated, easier to enter.  Now Linux is being debated for various workloads closer to the center of the workload, multi-million deals in place with enterprise corporations.  There is more to go but Linux has penetrated into the center of the enterprise at this point.

Common Linux complaints today are about ease of use, level of integration, "Fit and Finish" - easy to use.  Uli ??? is a long time AIX/Linux/Mainframe user, how does Linux at IBM compete with IBM?  Dan's answer:  it doesn't.  The markets don't compete, AIX and Linux, for instance are distinct market segments for IBM, the solutions have different strengths.  Dan manages AIX, Linux, VM, VSE, etc. and the channel conflicts are practically non-existent.  Customer needs drive the type of solution that they want and their needs drive their choice of platform, operating system, etc.  Uli points out that internally, people are looking for ways to consider how to move from Unix to Linux and he expected some channel conflict.  Dan says that IBM rarely sees that from the customer - the customer weighs their needs and moves to the appropriate solution.

Uli points out that some of his people want things to work "just like they always did" as opposed to "just work".  Jim points out that IBM is one of the organizations that has long been involved with Linux and the potential for conflict and that there are many IBMers in the audience (/me ducks) who have been dealing Linux and IBM products for a very long time and can provide some insight and guidance for others.

Christy pointed out that Motorola's first product in 2001 was done in China as a way to enter that market.  Christy points out that there is a need to innovate at all layers of the stack to compete, and doing so with 3rd party products is very difficult; open source provides a mechanism for competing, especially in an industry where the software is becoming even more important than the shape and form factor of a hand help PDA/phone.

Eric Wickadella (sp?) from Business Week - is there anything about Linux and Motorola that competes well with the iPhone.  Christy points out that the iPhone capabilities can already been done on many platforms with the possible exception of iTunes (others have multimedia capabilities).  Motorola sort of welcomes the iPhone because it opens up the rich experience and shows where the industry is going.  However, Motorola is likely to drive its strengths based on voice and data plans, software capabilities, etc.  Christy clarified that Motorola doesn't intend to "dominate" the industry/stack - obviously like any company they will compete on strengths in that space.  Eric also asked about Safari - Christy thought it was useful, but it doesn't really cover 3D performance/gaming.  Messaging being the #1 capabiity and gaming be #2.

James Bottomley wants to collaborate:  what is the top 2 list of what the community could do for each of the vendors and the top 2 list of things where the vendor can contribute back to the community.

Chris wants to contribute back to the Pacific rim (Google helping the community, the community helping Google).  Also, he'd love to see the community create drivers for things like nVidia & ATI drivers.  The proprietary drivers are one of the significant thorns in their side.  [See the nouveau project, the vesa and, um, mesa? projects for work in this space --gerrit].

Christy points out that most mobile users don't realize that their phone is running Linux.  Consolidation and collaboration across the ecosystem to simplify and make more common the user experience on cell phones.  Motorola can offer back training of the mobile operaters about how Linux and open source works and is acceptable for use in that space.

Dan doesn't usually rely on the community to do things for us, but prefers to contribute directly and work with the community when we need something.  Request #1 would be:  when GPLv3 comes out, just "chill"!  A bit of laughter on that one.  ;)  Power management - that is where the community needs to do better.  Patience and persistence on the device driver front, helping to train and work with the stubborn, proprietary device owners.  Christy seconded the power management issue.  Chris wants to second the GPLv3 "chill" request.  Chris quoted Eban, roughly transcribed as:  "GPLv2 wasn't perfect until we started GPLv3 development."

Jim:  what do the panelists expect out of this meeting?  Dan:  continued collaboration, hopefully some specific work actions out of the meetings with a focus on making Linux better and building the industry.  This is important to the industry and we need to keep at it.

Christy:  reemphasized that the US/Europe locality comment from earlier in the day was important - most of their developers are actually in Asia.

Chris:  we take great benefit from Linux - when you leave here, have something in your mind to *do*, be it publishing, using, developing.

Powered by ScribeFire.

Linux Foundation Collaboration Summit

An exceptionally full house at the Linux Collaboration Summit with nearly all the names that matter.  Current panel has Andrew Morton, James Bottomley, Chris Wright, Ted T'so and Greg KH, moderated by Jonathan Corbet.  See the web site for the full agenda for details.  Jonathan's first question was related to the stability of Linux, which led to a fairly lively discussion discussion.  Even Christoph Helwig had some heckling, er, good points to make, including the fact that the rate of change has led to less community review, and the fact that there is still no regression test suite.

There are several forces pulling in different directions, leading to a lot of tension on the core kernel, including support for new hardware, stability for old hardware, support for new features, stability for commercial end users, continued subsystem clean up, purity and simplicity of design, ease of mainenance, etc.  Also, the "scratch your own itch" philosophy of code development leads to a lot of new and conflicting technologies moving in parallel to mainline.

Mark Shuttleworth started and engaged in a discussion where Andrew Morton requested more detailed input from the distros, including problem areas, proposed release dates, staying closer to mainline (perhaps), and better dealing with API stability (and how that causes divergence from mainline and related stability).  There was some applauding of the fact that distros were pushing back on accepting changes that were not also and already upstream.  That helps maintain a single source for bug fixing, stabilization, and a location for hardware enablement.

Don Marti from asked about Andrew's State of the Union comments about Linux's lack of Power Management and for an update on the state of Power Management and what help others could provide.  Today most power management in linux is binary - on or off, where many vendors are now providing low power or reduced power states.  Andrew strongly requested access to specifications to devices in general, specifically here for power management APIs but generally the community has requested APIs for devices in general to help make the devices operate optimally on Linux.

Jonathan pointed out that a recent driver he had written did not have suspend/resume capabilities and the community did not point out the need for power management as a basis for accepting a fully functioning driver (good point) but Greg KH pointed out that the community will accept drivers at any stage of development, assuming that they will evolve over time to have more complete functionality.  Jesse Barnes from Intel pointed out that there is not a robust, complete API or environment for people to program to.  Also, Jesse reiterated that specs for devices are good - but he would also like to see vendors write their own Linux drivers for their hardware.  Greg KH pointed out that he has 85 driver writers standing by, ready to write drivers for devices that vendors contribute and provide specifications to.

David Slessinger from Axxess pointed out that mobile/open source activites are more prolific now but Chris Wright countered that the mobile folks are not very visible to the kernel community.  More engagement, visibility and interaction there would be helpful.  Greg KH pointed out that servers have the same power management needs as the mobile community and the two communities need to work together better (server and mobility) with respect to power management.

Ted suggested that an interface to broadccast that a disk was just spun up could be used to communicate to other subsystems that the disk is available, power has been expended, why not opportunistically do any other current activities that need access to a spun up disk?  Andrew pointed out that he has not seen requests, requirements from the mobile community and really doesn't know what capabilities that they really need.

Accessibility workgroup chair, Janina Sajka is present and mentioned that a screen reader patch is available for the kernel but Greg KH pointed out that the patch needs a lot of work and it needs some testing - however he doesn't have access to the hardware to test.  Janina promised to work with him to get a coder and tester to get the code into the kernel.

Darren Dagless (sp)? from Novell brought up ATI & nVidia proprietary drivers.  Novell has a lot of proprietary driver partners and Novell does not support those drivers directly.  Christoph suggested to tell them to "go away" - however we do want access to the hardware that some of those vendors provide.  This area is as contentious as always - with the primary viewpoint of the kernel community is that vendors need to provide open source drivers or users need to avoid the use of those hardware.  Ted also pointed out that both individual users and system vendors integrating hardware into their systems need to be conscious of the open source driver availability of the hardware components that they use.

Greg KH pointed out that the Linux Foundation has a mechanism for allowing vendors to provide specifications to the Linux Foundation under NDA and still get a driver written and supported by the community.

Stephen Von Nicholson from Ziff-Davis talked about how ATI has now offered up specifications even if not drivers.  On June 29th 2007 the GPLv3 will go live.  He asked what the kernel community now thinks about GPLv3 (earlier some had claimed that the GPLv3 was the end of free software as we know it).  Greg KH still believes that the GPLv3 has not addressed the primary problems and is not a fan of it for use with the Linux kernel.  James pointed out the the logistics for converting to GPLv3 are probably also prohibitive because there are no localized ownership for copyright, etc., and the benefits of GPLv3 are not likely to be compelling for a change that would be logistically overwhelming at best.  Ted pointed out that the GPLv3 has improved dramatically over the past year or so, and he is not as opposed to the use of GPLv3 for user level applications as appopriate.  James and Ted point out that Eban Moglen was very responsive to the problems pointed out by the kernel community (among others).  And, there is now one more license in the ecosystem and they will all over time co-exist.

Someone asked if the kernel could be converted to GPLv2/GPLv3 dual licensed.  General answer was "not easily" and "not likely".  Don Marti pointed out that 1/3rd of the kernel is already licensed (in source code comments) as GPLv2 or later, as opposed to GPLv2 only.  Also, much of the kernel is dual licensed today, especially in the driver space (e.g. drivers used in both Linux and BSD is pretty common, with different licenses).

Jonathan posed a question about contributors - lots of US and Europe centric contributors, lots of desktop and server vendors, but not much from Asia, not much from mobile and embedded companies.  His question focused on how to expand the contributor community to include a more representative, diverse set of companies and interests.  Andrew pointed out that some vendors have not seen a need to interact with the communiy, in part because of the high rate of change of their products, and the need for a longer term involvement with the kernel community.  James pointed out also that the discussions are all in English, so non-English speakers are at an inherent disadvantage.  There is an LF collaboration summit in Japan already, James suggested one in India or Africa would be possible.  Also, the existance of a "universal translator" might help, with a couple of chuckles aimed at the Google folks present since they seem to be active in spaces like this.  Chris pointed out that the developer community has not been getting as much "new" membership lately, especially in the core contributor space.  Ted pointed out that the OSDL/LF sponsored Japanese collaboration summits provided simultaneous translation to/from Japanese/English which substantially helped collaboration at those summits.

That ends the first session, and I'll do more if my battery, fingers, and attention span all hold out.

Powered by ScribeFire.