Monday May 21, 2007

I attended JavaOne this year for the first time. I went to CommunityOne sessions on Monday, prior to the main conference. Then, I worked the OpenSolaris pavilion booth the rest of the week. Following are my notes from the event.

Rich Green keynote (CommunityOne)

Why get together? Rich talked about the importance of face-to-face meetings now and then in our virtual world of work. JavaOne is a great time to gather and have verbal conversations.

Aside: Some folks are currently discussing this same idea WRT user groups on the OpenSolaris main list today. I agree with Rich and the reason I go to user group each month is to learn by having new conversations with folks interested in OpenSolaris. I have a much greater understanding of the documentation needs of the OpenSolaris community by going every month to meet and talk with folks about documentation. You cannot know unless you ask.

In addition to Rich Green's insights on live conversations, he said "simplicity and access is far more important than technical perfection". WOW. read that again everyone. This is big, McNealy was saying this years ago and I still don't think we get it a Sun. I still see a lot of complexity and secrecy, not as much, but we're hearing this from Rich because we've still got WAY too much. He went on to say that the power to share technological information overrides technological limitations. That is, the desire to communicate is the dominant force today. I think this is all very cool because communications is my trade.

Then Tim O'Reilly took the stage to moderate a panel discussion with Rich Green, Tim Bray and Ian Murdock. He started the discussion with a statement about 'How do we help developers to 1) harness collective intelligence, 2) create live software, and 3) think about open source and open standards in lock-in for the future (fuzzy notes on that last one).

Tim Bray complained about Tim O's use of terminology right off the bat. Which is right on target for a Sun conversation because we always nitpick terms first, get on same page, then have a discussion. O'Reilly did a nice job of heading off a full-blown terminology death spiral and focused on the real questions he'd posed: how do we get developers software toolkits to extract the data they need. How do we make it easy for developers to extract, implement and deal with all the data? Ian chimes in with commodity and expertise. Tim Bray mentions advent of Spring/Rails makes it easier to contribute. He talks about the convergence of human and compute communication.

Then, O'Reilly mentions web 3.0, there will be no typing, we'll use gestural interfaces, voice, movements instead. Tim Bray reacted with surprise to this idea. [I was relieved because my typing stinks, always has. Go W3.0!] Ian jumps back in with the importance of the operating system. ? Then, there was a discussion of the processes needed to grab all of what is written out there, or creation of the process to pull-through the right information, how to create more value than you capture. People contribute for their own self-interest, not to contribute to XYZ project.

O'Reilly and Bray are impressive in this conversation. I wish they'd discussed this topic a lot longer, as they only just scratched the surface and had to finish up just when it got interesting. I highly recommend taking any opportunity you have to hear those two talk, very exciting!

At the Silicon Valley OpenSolaris User Group April '07 meeting, Brendan Gregg talked about the DTrace Topics project. This is a collaborative documentation project for DTrace. So, of course, as a writer, doc community leader, and infogeek, I loved this talk.

Brendan went on for 3+ hours and I would have stayed longer yet. He is so keen on documentation, that I just want to soak up his experiences and knowledge and sensibilities until they are my own. He also has a great sense of humor and appreciates Star Wars, so that doesn't hurt.

I let him know about the latest update to the Solaris DTrace Guide and offered help to update his DTrace toolkit with link targets when he finds an appropriate SPARC box to hook up the HD where it is all patiently waiting for his attention post-overseas-move. Hopefully the docs.sun.com hardware update is complete by then because the slowness is insanity right now. But, we did see the first interface update of the site in 5+ years last week, so my hope are high, thank you dot sun!! Brendan then had some threads on docs-discuss on opensolaris, where we're getting started on some cheatsheet documentation, and we had a chance to talk again in person at JavaOne.

I mentioned that we should get the DTrace toolkit included on the Starter Kit in future. That would totally rock. As an aside, I'm all stoked because I initialized a Mercurial workspace for the Starter Kit project late on Friday, so I should have it populated as soon as I get a README together. Mercurial just works and that is very cool, being a program manager all year (read:slideware hell), it felt really good to have a live workspace even if the only output is 0changes 0files 0diffs.

Currently, Brendan, me and few others are having a separate thread about wikis and how to get all the good doc collaboration from solaris internals and genunix.org consolidated on opensolaris.org wiki. We'll have to start small with regard to features for this new wiki, but we'll have all the basics covered and we can build out the fancier conversion and processing functions as we grow it.

I was accepted into the SEED mentoring program at Sun for the Jan '06-June '07 term. I was matched with a truly awesome mentor, Lynn Rohrer, since then we've been working together regularly on business strategy, leadership, and open source.

In April, SEED had its mid-term meeting and following are my notes.

The most interesting talk was given by Jim Baty, the first Distinguished Engineer in Sun Service, on "Everything 2.0". Jim talked about historical Sun a bit:

1st ten years: workstation development
2nd ten years: servers, enterprise Sun-on-Sun
3rd ten years: services

He then asked us each this question: What Sun do you work for?

The important part of the question is 'what measure do you use to describe your Sun'. Is it your geographic location? Is it software? Is it hardware? Is it mobile? Is it your organizational location within Sun? I usually think 'software' first, then 'education', then 'open source'. The education piece is bi-directional, I see Sun as a learning place, it is where I learned UNIX and tons of other great stuff, it is also part of my role to educate developers and users about Sun products.

Jim went on to talk about 1.0 versus 2.0. In 1.0 we pushed information to customers, visited customers in person, and made users wait for information. In 2.0, the lines between employees and customers get blurred as we increase collaboration on products and recognize the programmer as the end-user with a broadening of their customer role as innovative contributors. We move away from the push model and begin viral marketing for the participation economy. Where read/write is mostly write and cultural changes re-invent 'your Sun'.

This really resonated with me because OpenSolaris is all about information earlier, participation of everyone on the things that matter most to the community of developers who innovate the most.

Next up was Roger Meike from Sun Labs with some interesting demos of SunSpots.

We all had lunch together that day, then visited the executive briefing center, followed by afternoon talks and a lovely tea party with great food and wine. It was really a wonderful break from the usual Friday and a great opportunity to rub-elbows with lots of technology gurus.

It is a real privilege to be a part of this mentoring program, I have many many other informal mentors and I recommend having multiple mentors throughout any career. I have never been a mentor myself, but I surely will become one. The impact of my mentors on my life and well-being has been huge and I would not be here without their help, advice, consistency, and challenges. I applaud the SEED program for making this a part of our official work, for so many who might not otherwise be able to make time for mentoring.

Monday Apr 09, 2007

What a year of music in San Francisco! Spring is really when things start to heat up, tours begin, new albums release, it is a joy to be alive.

So, it is at this time I feel that I must post a wrap-up of the concerts from 2006, before too many shows have piled up that I become blocked by my own procrastination. It is such a privilege to live here and have access to so many musical artists, here are the big shows I've had time to write up:

February generally starts the fire with Noise Pop festival in SF, but this year (2007) was a lineup that left me flat completely. I traveled abroad during this time, but for-the-record, only Erase Errata would have got my money had I been in town. Last year was not so, we saw 10 shows in 7 days, it was a blur, so I'll just begin with March:

March 14 2006, _Frank Black_, Cafe Du Nord. This night was exceptional because Du Nord is so small and intimate and he is so brilliant. Frank played three sets, one with electric, much acoustic, and one with accompaniment. He played everything that matters.

March 24 2006, _The Strokes_, Concourse. Probably the best show of the year, just following the release of their latest album, and everyone at the Concourse knew every lyric. After the first song, they were stunned by our adoration and said simply 'San Francisco, who knew?' We jumped and danced and sang the whole night with them and their indescribable guitars, I love them. I saw them once before and this was nothing like it, this was magical, drunken, loudness at it finest. And we were all soaking wet from the rain and did not care, I lost my polk-a-dot umbrella from childhood and it was worth it.

March 29 2006, _Heavenly States_, Rickshaw Stop. The Heavenly States are one of a few favorite local bands, their lead signer is divine and the band combines fiddle with rock with keyboards in a manner most cannot comprehend nor pull off without theatrics. I think it is because of the Telecaster ripping between all the sounds and brining them gracefully together and apart, but I'm probably biased. You might find a stop-action video of them on Havoc if you dig around.

March 31 2006, _The Cloud Room_, Slim's. I was not floored by this band really, we went to Von Iva show after this show and that is what I remember.

April 28 2006, _Yeah Yeah Yeahs_, Warfield. This is a close second to The Strokes show because she is just so amazing to watch and listen to. Vocally she really is the tops now and she is pulling through a giant mass of female vocalists that will change our music for a long time to come. She is gigantic. She bopped around a ton and said only on thing to the crowd, once she fell right on her ass, at start of like the second song in their set, and she just said 'Holy Shit!' and kept right on going. We had dueling tickets on this night for the Mates of State, but went for YYY, of course. I've seen Mates and love them, but they're no YYY, no? :)

May 2 2006, _Sleater Kinney_, Great American Music Hall. My one main brush with greatness, only major claim-to-fame, is that I opened for Sleater Kinney back in 1996 at the Fox Theatre in Boulder, CO. God those were the days, we even got to hang with them backstage. Anyhow, Sleater Kinney rocked like never before this 2006, I mean Carrie traded her trusty Rickenbacker for this huge guitar, a Thunderbird, I think, and played like she was channeling Jimi or something. Trees is the most enormous album of 2006, I really think, just like a Redwood. Just get it and One Beat too. I did not know it at that show, though we all felt it, that this would be their last, they have disbanded, and it is the end of an era for me, a decade of riot grrrl has finished and their mark on punk and rock and pop will always be felt strong by me.

May 26 2006, _The Ministry_, Fillmore Auditorium. This show was incredible, truly demonstrating that music defies age completely, these guys still bring it from every pore. San Francisco did give them a good reception, somewhat subdued as usual, and certainly not the moshing, spinning, scary-wonderful affair I experienced back in '94 when I saw their tour with Helmut, but still an awesome spectacle to enjoy.

June 5 2006, _Charletan's UK_, Fillmore Auditorium. My sweetie introduced me to this band, in great wisdom, this show was mind-blowing. This guy can sing like no other man can. He wales and croons and just slathers you with his voice and its beauty until you almost cannot take it. In so many differing styles, he sings effortlessly. This is to say nothing of the band, he can sing all day and only plays the occasional mouth-keyboard, but the band of Charletan's is truly exceptional rock band. Strong and tight in all of its parts, unfortunately a rare thing these days to hear a band this well-versed, trained together making such fantastic pop music with hard driving drums bass and guitars. Of all the bands this year, this one surprised me most because I did not know their superiority, or even a shred of it, but once I heard it I will always remember.

June 1 2006, _Les Claypool_, This Les show was my first on the main floor of the Warfield, and I'll never again sit in a seat upstairs. Why did I ever sit upstairs? My favorite of all women, Melora Craiger, divine beauty and queen of the Cello, creator of Rasputina, opened for Les. It was a crashing, attention-demanding opening act if there ever was one. Frustration Plantation was the setting with a bit of frontiering for good measure. Fewer vignettes than usual, but for the anticipation for the Les crowd, it was fitting. No corsets were worn and freedom from them was heard in each string plucked. Les had Gabby LaLa with him and to see her in this sold-out venue, after being schooled by Les for over a year, was to see her at her best, I am sure. She has grown before our eyes as a musician and on this night she played Sitar on-par with Les, trading licks and rifs with his genius, then jumping to accordion, then theramin, the yukelele, and back again. I will leave my ramblings about Les for the summary of a show at a different venue, he is in Bay Area, so we see him often...

July 8 2006, _Nine Inch Nails_, Shoreline Ampitheater. My first time having an actual seat at Shoreline, to be this close to Trent again was incredible. I stood on the grass for Radiohead two years back and the loudness beat on me until I was cured of all ills...that is another blog I suppose...if you ever might see or hear Radiohead, stand upon any surface to do so, and you will not be disappointed. Peaches and Bau Haus opened this show. Not much to write of these two acts really, Peaches is just too nasty for me to be interested and Bau Haus is just Bau Haus, but it was my first time to see them, so I'm glad for that, just to have seen a legendary band such as they are. Trent and NIN, well, if you read my blog, you know that we saw the first two nights of this tour at Warfield fall 2005. This Shoreline show was the final date to wrap it up. So, it was insanely tight, improved from repetition and rehearsal on-the-road all year long and summed up in this one night With Teeth. Lighting design was really stellar and highlighted nicely Trent's muscular frame, oy, I'll stop there. He was sad for this tour to be over, he said as much to us, he said he had to go back into the studio and sounded just so emotional about it, said that he would miss us while he was gone. He even tossed a guitar into the crowd at the end of crushing and banging the daylights from one another in a ritual that is uniquely NIN. They surely were on the brink of Survivalism and Year Zero at that time, and knew it. Recording and mixing is a test and grueling trial and Trent is no exception. He has now provided his Garageband files of Survivalism to the world for our use and edification, and I feel kindred to him in our delivery of sources to OpenSolaris. All of Year Zero will be made available from what I understand and the files for Survivalism are a joy to watch and listen to, we've been enjoying them all week. You can hear (pre-release) of the new album here.

September 17 2006, _Israel Vibration_, Independent. Another brilliant show from Israel, roots reggae at its finest that gives you joy and light in every verse.

November 5 2006, _Lee Scratch Perry_, Independent. My first introduction to Lee Scratch, also awesome reggae, this guy basically discovered Bob Marley FWIU. It was strange, the band that opened we felt was not very good at all, then as it turned out, they were backing Lee Scratch, but *with him* they sounded incredible, without him mediocre. It speaks volumes about his presence and voice and energy. He appears to be about 70 years of age, but this man sang and preached and danced until we were all tired, then danced with the women some more. He was inspirational in his energy, not to mention his mirrored cap.

Date? _Peeping Tom_, Great American Music Hall. Peeping Tom, uhg, now this is tied with The Strokes for best show of the year, actually. This was more fun than any other show at GAMH ever, Mike Patton is a living legend in the Bay Area and the crowd makes it known. It was a romp and rough evening of beats and just downright hard to match music.

Date? _Kool Keith /return of Dr. Octagon_, Mezzanine. Small very fun show, so fun I was hit right in the bridge of my nose with a flying LP record during the first act. But, I sucked it up folks, that's right, I needed to see the doctor! Dr. Octagon.

November 26 2006, _Alice in Chains_, Warfield. This is a legendary band, so we went even though, of course, there is a new singer who is so-so, and this tour is a reunion of sorts and well, we just never did see Alice in Chains back in the day and now we have, sort of.
The lead guitar player really led the show and made sure the new signer stayed in place. He had the vocals yes, but not the gruff of Alice and Chains that comes from that angst of the early 90s. They paid nice homage to the man who was the face of Alice in Chains, took time after the first set to remember him and reflect, then played an acoustic set, followed by another set. Overall a good solid show I would recommend if you never saw them ever.

December 2 2006, _Primus_, Berkeley Community Theater. For this show, we had seats in the second row! What a way to wrap the calendar year. The seats were so close we were stunned throughout the opening act, I could smell their body odor, that's how close we were, I could see the pores in their skin and the sound trembled across my arm-hairs it was so loud and rich. Also my first time at Berkeley Community Theater, which I highly recommend if you have insanely great seats. We went for sushi beforehand over on Telegraph, I think, and it was so fun and yummy. Little boats of fish and rolls swimming past you while the anticipation of sitting directly before, no two rows from, a virtuoso bass player of our times. We will go again to enjoy his unique mad-hatting, frogish friendly fun at end of June, hooray.

Now, we've had a rush of nice, small shows this month of March 2007, that I'll cover real quickly, as follows:

-_Octopus Project_: great Austin-based band at BOTH that was a thrilling wall of sound
-_Miho Hatori_: New York based Miho, formerly of Cibbo Matto, brilliantly with her new album Ecdysis
-_TV on the Radio_: good solid band, not my favorite style, if I want this I want Bloc Party.
-_Teddy Bears_: amazing tiny show, we touched them, great great Sweden pop right now!!
-_Whitey_: New York band, very great guitar rock all Dan Electro through Marshall half-stack, check it.

And this brings us up-to-date completely. I will not fall behind this year, I will not! Now, if I do, just read this stuff or order one of these or find me over here where I spend almost all of my writing juice.

Friday Mar 16, 2007

Background
----------
I have worked in software since 1996, started out shrink-wrapping and printing disk labels for a language translation software shop. There were two of us, smoking Marlboros and drinking Folgers working on Windows 3.x machines to parse bi-directional language dictionaries for word and phrase translation. I'm a horrible pack-rat, so I actually still have my 'I Hate DOS' book in the trunk of my car, just for a laugh when I load the groceries. You don't even want to see my home directory.

I worked up to print ad copy writing, WinHelp grammar file editing, and answering support calls. That is where I found great satisfaction, helping folks set up their keyboard for Icelandic or whatever and successfully getting them through a real software problem, from angry enough to call us, to happily translating phrases from an old letter their ancestor wrote.
I went back to school for technical communications, and worked at two other software companies before Sun in PR and tech pubs.

Sun Work
--------
I worked as a tech writer at Sun 4.5 years, quit smoking, found sumatra, and just became a Program Manager(mangler) last fall. I never wrote books or help for Solaris proper, only component products like Sun Cluster and N1. So, it is a great privilege for me to work on such a fantastic product. A product I've used everyday these last five years and I truly have never been so productive as I am on Solaris. That is part of what I like to share with folks most, great productivity when you learn even a tiny bit of UNIX and run on a solid platform. I've had to write content and build systems the other way and it sucks. I won't go back to using those sad sad tools.

I am extremely proud to be a part of the OpenSolaris community and consider it a real gift to contribute to its success. As soon as I heard about the Doc PM job, I could not stop thinking about open source for one second. I think it is just outer-limits to be making redistributable all these awesome documents, and I feel the power of having done this has not yet been realized by the larger market.

Position
--------
I just want to help create a place where folks want to come to innovate on this OS. I don't give nearly as much thought to how we rule ourselves as I do how to get things done. I care about how to conduct fruitful discussions, how to make needed connections, how to remove barriers, how to empower contributors, how to work on files and mail to make incremental progress everyday, every week, so we all can feel successful because we built something new and better and someone found use for it.

Wednesday Mar 07, 2007

Nexenta: Martin Man
-------------------

Martin's presentation was expert, had many demonstrations and highlighted Unix for humans concept. He talked about the Debian social contract: you can use it free and it will stay free. He mentioned the Debian Free Software Guidelines. He described the Ubuntu Code of Conduct: we value anything you contribute and we respect that you have no time to work on it.

Nexenta, OpenSolaris for humans, packaging and compile changes.

Issues: partitioning and installation
Sun compiler and GNU gcc compiler
linker bugs not being worked

http://gnusolaris.org

Debian->Ubuntu->Nexenta

Belenix: Moinak Ghosh
---------------------

Live media changes default to boot RAMdisk (this is main difference in addition to what Joerg discussed). Also, pre-built repository.db to save time at boot. You want this because Live media boot is like a 'first boot' every time. Pre-populate driver_aliases and order block files on CD. Moinak used DTrace to figure this out, to determine the order. Compress 1.8 GB to 700 Mb, then you read one block and get two blocks worth of data, uses HFSF for decompression. Uses I/O scheduler to make blocks consecutive (re-order and coalesce). Then make them adjacent to like blocks, and call them all as one request so that the disk head always seeks in the same direction.

Get
livemedia project details.

USB boot goes like this:
1. Insert thumb drive
2. Go to BIOS
3. Expand +
4. Switch boot order of HD and thumb drive

This is great because you can have and save state with USB drive. For example, you could add a user account and when you boot on another machine, that new user will exist. You can also have encryption if you lose the drive. It requires only a computer and BIOS support for USB boot.

LiveDVD will be in Caimen.

Comments: We need an online software repository for cooperation between distributions.

Schillix Distribution: Joerg Schilling
--------------------------------------

First community distribution of OpenSolaris, LiveCD with a simple GUI. It doesn't compile on OpenSolaris because OpenSolaris is still incomplete. Joerg has wanted LiveCD of Solaris since 2003. In September of 2004, CDDL was authored because OpenSolaris combines code that is open and code that is not open. March 2005, Jeorg achieved first pure OpenSolaris boot after much work on the missing parts, libm, the math library being the biggest amount of work. He did this by porting from FreeBSD, most of which code, ironically originated from Sun. But, it was an older version from 1992. Other missing pieces he worked on to get LiveCD of OpenSolaris operational:
-libXML2
-libz
-Netscape portable runtime
-bzip2/gzip
-Make
-ksh
-Parts of SMF
-NIC drivers

So, Schillix boots only on x86 using GRUB. How does it work?

1) Load RAMdisk image (compressed ISO or UFS) and start kernel

This is a chicken-egg problem to solve because you are trying to load RAMdisk before you have it and start kernel before you have it. So, you must put drivers and basic kernel modules into the BIOS.

2) Mount /usr from CD

3) Re-mount only read/write

4) Plumb network devices/DHCP

5) Enter normal multi-user startup

There is a kit provided by Schilling, so you can do this yourself, roll your own:

ftp://ftp.berlios.de/pub/schillix/README-kit

Issues:

To compile OpenSolaris on OpenSolaris, we need redistributable Java and some jar files the Sun compiler uses. Schillix needs your help! If you read German, Jeorg has co-written a document of 1,200 pages about OpenSolaris administration and internals all in German.

Questions:
Is there an update mechanism? No, must copy to hard drive, and is still difficult. Need free partition, then diff. What we need is stable S11, so we can develop upgrade.

Roy T. Fielding, Phd
Chief Scientist, Day Software
-----------------------------

The Keynote address from Roy really was a talk describing his extensive work on computing technologies since 1986. This was fascinating for me, a short history of UNIX from his perspective, and I'm so glad I was able to hear it.

Roy began work on WICAT minicomputer in 1986, with a 50Mb hard drive, which was very large at that time. He was an Apache founder in 1995. He is focused on 'engineering for serendipity'.

His work has been about a constant set of commands, simple construction of co-routines, that are reusable and extensible, pipe and filter.

This progressed to visualization of pipelines with images and image processing. In 1993 he designed protocols for Web architecture. And in 1994 he designed HTTP 1.1 components-based REST, representational state transfer.

Then he moved on to Apache, httpd2.x Filters and modules that could filter incoming and outgoing messages with Apple Automator. Apache had the framework to chain filters. Now, he's working on 'Plagger' for feed, these are pipelines that take blog entries, images, find where to take data, do operations (filter and feed through routines), then output: Yahoo Pipes.

He discussed then some deficiencies:

UNIX pipes: only single stream, no meta-data
HTTP: only one stream, each direction with meta-data up front
RSS/Atom feeds: go over HTTP, take meta-data in XML, still single-stream

Now he currently works on the 'waka' protocol to replace HTTP. With this protocol, the message has the entire request embedded, works over any transport mechanism. Imagine a hierarchical network of cache messages, client could broadcast message to cache and immediately take back best message without delay. Better interconnectedness. Imagine proxy that takes requests and at the same time sends request to the rest of the world, send two parallel requests and overlay the results when they return.

[note] This totally reminds me of Xenocide. So cool.

He then talked about what he means by engineering serendipity, the land of unexpected benefits. This means to intentionally design for things you don't expect. To him, this has been THE major component of success for all of his projects.

As far as OpenSolaris, he said to get involved in the Constitution. Sun needs help to give up control of the project. The existing Solaris processes do provide tight quality control ensuring deployed interfaces don't change. However, this results in lack of innovation and makes it hard to get changes into the system.

Engagement: his purpose is that UNIX can go on forever, by maintaining multiple active communities, sometimes working on competitive projects, to learn from each other.

Issues with the OpenSolaris community: knowledge of where decision-making happens.

[note] The context for the type of decision he is talking about is important here, he means decisions about whether we should release, or whether a certain interface should change. Not the regular and commonplace decisions that engineers make in their daily work.

There are two options:
1. individual developer makes decision
2. project groups make decisions

So, Roy feels that, once we have a newly elected OGB, we need to re-organize communities and projects so that they have the decision-making powers they need to be successful.

Elephant on OpenSolaris: Zdenek Kotala
--------------------------------------

There are now 8 USDT probes in PostgreSQL for DTrace, Zdenek provided a demonstration using Chime, the DTrace visualization tool. This was extremely interesting and exciting, since the Chime project was one of the first projects on OpenSolaris, is quite mature and now being seen in presentations of other technologies as demonstration material. It was cool. You can go here for example PostgreSQL DTrace scripts.

There is also now a Sun Cluster data service for PostgreSQL, it is only failover data service, not scalable type, but this is also quite exciting to me because I worked quite a while on Sun Cluster data services documentation and man pages. To now see this product reaching out and being further developed in open source is wonderful to hear about. The PostgreSQL dataservice is only for Sun Cluster 3.2.

Web2.0, Thorleif Wiik, PixelPark
--------------------------------

Thorleif's presentation was the most corporate-type presentation of the two days that I attended. It was so interesting to hear from him about the business model they support and how they do it with Solaris technologies. PixelPark supports web sites which serve, for example, auto sales industry. So, they have the task of supporting storage of 400k images for one such site and serving them up across the world, for example. They used to do this with many small boxes and had high overheads of system administration and space. Today, they put all layers of the stack onto one box, separated by Solaris Zones, like this:
-Network
-Security
-Load Balancing
-Servers: Apache
-Database

Much to my pleasure, he uses the blog of Eric Kustarz, a good friend of mine in San Francisco, to find his file system tuning solutions and he referred it in his presentation.

It is just a totally small world these days.

Ksh93: Roland Mainz
-------------------

Motivations of the ksh93 project for OpenSolaris:

-It is a modern shell that provides portability for writing small applications and increased compatibility for OpenSolaris.

-Students with terminal need working cursor keys by default, arrow keys are so very important to new users, they also need input/editor mode by default (or they just start kicking the machine with their foot).

[note] Roland's presentation incorporated more humor and garnered more laughter and fun than any of the presentations. He had us all in stitches and went over his available time, (which is pretty strictly enforced in Germany, except at dinner) but I think we all could have listened to Roland for another hour and enjoyed it a lot, particularly commentary on the ARC process might have been a real gas.

-Name and variable command expansion will also be possible and is very useful to new users because it limits extra typing by using tab keys and menu lists.

-Performance is better because of built-in commands with the modern shell which prevent a fork() for subshell and thereby limit overhead.

-Administration is simplified with global configuration file for interactive sessions using /etc/ksh.kshrc for global changes without touching /etc/profile or home directory of users, which is not always permitted.

Features provided by the ksh93 implementation:

-Associative arrays, having non-numeric is important for those who script extensively.
-Locales, prevent manual work needed for gettext.
-Built-in floating point math.
-Networking for client-side and server-side operations, unlimited complex applications.

The future work:

-Complete the first putback
-Complete L10n message catalogs
-Migrate default shell to ksh93, with respect to Makefiles
-DTrace support

The following are my notes of a discussion with Perez, an attendee at both FOSDEM and OSDevCon who is interested in documentation and currently works as a chip designer, but hopes to enter into software development because of the excitement around open source.

There should be a front-end for doc content entry that is easy to use. We discussed this point and how to maintain trust of the reader. He himself had not considered differing levels of trust of documents per se, and made me think that we might also think of this problem less. That is, it might hold us back less from moving forward with unmoderated wiki development of documentation.

Perez shares the feeling that we espouse in Sun software documentation regarding a focus on increased communication outweighing any focus on formatting. Focus should be totally on the content, let the formatting be automated. He asked what I bring to the conferences and also to my user group SVOSUG. I replied that I come to ask what documentation developers and system administrators need and what they might contribute. He replied that is quite a luxury to be able to provide for an open source project, doc support that is. This echoes a comment on my blog recently about the importance of documentation of open source products and the common lack of documentation in many projects.

We talked about excitement of open source software with respect to having information freely available on the web, not just proprietary information held closely by corporate engineers. This is a problem in his industry, much like what we have in commercial software, that results in difficulties, as follows:
-Training of new individuals is time-consuming and slow because information is held by a small number of people and cannot be easily shared or transferred because those individuals are so busy and because the information they have is competitive.

We discussed how open source is a better and overall excellent model because it allows information to be shared, communities can help one another freely across corporate boundaries and a willingness to do so is encouraged by the model.

Perez asked about the Starter Kit and wondered where they are and if the left hands simply knows not what the right hand is doing. This is in response to not having kits at OSDevCon and the many individuals at FOSDEM who had ordered the kit last month, never to receive them. I told him about the new web page above and that this should improve the situation greatly.

We talked about the FreeBSD documentation, how it is very excellent and the recent re-design of the web site aligned with PHP docs. I feel that we should look to these leaders in information design as we determine if it is something we can do for docs.sun.com. Because I'm with these developers, administrators, communities members so often, I do hear very often about deficiencies of docs.sun.com, and one is crying out for PDFs and collections of them lately. But, again I think we should look at FreeBSD handbook and overall design to learn where we can improve.

Service Management Facility, Detlef Drewanz
-------------------------------------------

Immediate questions folks ask about SMF:
What is manifest loading?
Can I switch it off?
Where are the scripts?

SMF is a common service management framework
System administrators have to live with this system
Complete change in the boot process
We now have service dependencies

In S10 we have same number of services, but different dependencies, so we do things in parallel during the boot process, no longer is it sequential, it is milestone-based. So, we get better, more explicit indication of the fixed definition state of the system for a particular milestone. Now developers can introduce a new milestone. Benefit: You can switch to milestone3, test it, then go back to milestone2, test, go forward, etc. Couldn't do that before.

Also have automated restart of services, and secure delegation of administrative tasks.

SMF gives us common service development and deployment framework using service manifests, and includes handling for upgrade. Formalizes the service development and makes common start stop, etc.

Next questions:
How do I identify a service: FMRI, fault managed resource ID
How do I describe a service: Service Manifest XML file
-name it using FMRI
-define dependencies
-create methods (start/stop/refresh) with properties
%svcadm enable 'service-name'

To create your own service, use this
HowTo

Because the manual is good, but sometimes the manual is not so good.

[note] I asked Detlef for further information about this statement after his talk and he said the following:

Documentation is spread across too many web sites and we need one place to go for information. However, he also did say that it was useful to have the How To articles, so it is a difficult problem to solve. He said that he would not expect to find the information about SMF in the Advanced Administration Guide and felt strongly that SMF needs its own document, its own book, because folks are unlikely to find it in Adv. Admin Guide, but he did find it there, by using search, and was surprised by that location. He felt that the information should be in the developer's collection or on OpenSolaris.org and the information should describe development of manifests.

SysAdmins love SMF largely because the status of the service is maintained across the upgrades.

Questions:
What is different between restart and refresh? Restart is stop and start, refresh is just refresh cache or re-read the config files.

[note] This question received loud comment from many folks, it was a very experienced crowd, so no dumb questions aloud, pretty typical UG audience, IMHO. Ask a newbie question and the group roars back with a response like, 'of course you need both restart and refresh and how could you be so silly to not understand the difference'.

March 1, 2007
Dahlem-Dorf, Berlin Germany
German Unix User Group
----------------------

Detlef Drewanz gave the conference introduction, he leads the GUUG. I got some initial questions while waiting for things to get started, as follows:

0. Are you from Marketing?

1. Is the documentation source on opensolaris the same as the docs.sun.com?

2. I've downloaded it once and did not want to grep through the HTML, need PDF instead in order to search.

[note] I continue to believe that the sources will be important in the future, they are not that interesting now. No, not from Marketing. Yes, the sources are the same as those on docs.sun.com.

[note] I need to be sure to link to Privilege Bracketing HOWTO from bigdoclist on Docs Community:
http://opensolaris.org/os/community/security/library/howto/privbracket

Telephony summit and other conferences are hosted by GUUG in Germany for many years, at least 14 years ago, 1994 first Linux conference/congress summit, no one commercially using it at that time, built up a community, put together lot of interesting projects through the congress. Good contacts with professionals in solaris/opensolaris, so they concluded that there must be an OpenSolaris conference to achieve the same things as they did with the Linux congress.

In parallel to this conference there is GUUG conference ongoing in same building, UNIX administration topics will be covered. Now on to Simon's keynote. Casper Dik was not able to join us this week. Next time we do this conference will be in 2008. The entire line of OpenSolaris distribution maintainers are here with us.

Simon Phipp's Keynote:
[Announcement: get your free OpenSolaris starter kit!]

Simon introduced Java to IBM in 1995, didn't seem important at the time. Similar situation today in this conference, note this date because it is the beginning of something big.

Back in 1991, traveling on business talking to people about video conferencing. Different times, telephones, cash and traveler's checks, slow post, etc. Today, credit cards, e-tickets, mobile phone, email.

Suggestion that we've reached the end of the consumer age, it is history. Topology has changed, from governments at the center and consumers at the spokes, to a meshed topology, where everyone is in the middle, anyone may post a question, vote on the issues raised and government responds. Heading back to pre-industrial-revolution society, world is a village. This new topology changes the way you handle security, technology, government.

Not just boundaries anymore, it must work with digital identity. Software, not monolithic systems, but small loosely joined pieces, service oriented. Pricing, vendor was at the hub in the past, now must bring focus to the value. Marketing, once produced polished messages, but consumers are now participants in that conversation. Development was a closed room, now statistically all smart people are outside the closed room, now open the room, open source.

What is open source? Fundamentally, the work of a community of developers. Sharing code commons and creating wealth from the commons. What makes you feel rich? Open Source puts the means of production in the hands of the people, with perfect visibility.

Ecosystem:
Artisan: licenses software through free commons

Co-developer: makes changes that he feels create richness, then contributes back in order to minimize the cost of maintenance of that richness:

-Licenses either allow you to create richness or not
-Governance makes room to contribute back, commit rights, open SCM
-Motivational models, numbers of models indicate health
o do all participants of the community work for same company, speak same language, have no arguments? If so, these might be a sign of poor health.
-Patent protection comes from good governance.

Deploy-developers: change in the way you get your software is most important, acquisition.

End-user: change in the value inside the software bundle, making choices about how you spend money, by identifying some value and paying for it.

Sun returning to our roots as a software company working with
community to create richness.

Question: why is Sun giving money to FSF?

Sun saw FSF group helping us to license Java, working on OpenJdk, so we thought it important to support them because they share our values. Does this splinter?? [not sure of the question] There is Sun part of world and GNU part of world? Simon: no there isn't, we can come together. There is much more that unites than divides.

Question: Debian: GNU and solaris kernel, have there been continued discussions of Debian working with opensolaris as supported kernel? Simon: it is a long process, will continue to be so, but there is progress in that a sun employee is on way to becoming a Debian committer, less fighting between the groups today, but long way to go.

Announcement S10 University challenge contest 2006:
November 2005-June 2006, showcase, 330 entries from 15 countries.
Dr. Sergej Alekseev and Gunter Stiege Oldenburg University for JDLabAgent and debugging agent for automatic Generation and Analysis of Tracing. Like a flight recorder for Solaris system, uses DTrace, but extends with graph theoretical algorithms, finds a subset of monitoring points and reconstructs events from this. next, java debugging lab architecture. JDLabAgent is shared object loaded into JVM then uses front-end for analysis of kernel flow.

http://jdlabagent.sourceforge.org LGPL
http://jdlabstudio.sourceforge.org GPL

Reward $5,000 student workstation and 100K HW donation to University from Sun.

Tuesday Feb 27, 2007

On the second day of the FOSDEM conference we distributed the remaining 200 OpenSolaris Starter Kits, for a total of 500, and word-of-mouth was very good. We had many students coming to the both specifically for the kits and we ran out before the day's end. We might have been able to distribute yet another 100 kits based on the high interest.

I was pleased that, at FOSDEM, most visitors to our booth knew about OpenSolaris and had questions that indicated some previous knowledge of the project. The most frequent question was not 'what is the difference between Solaris and OpenSolaris?', as was the case at LinuxWorld and SIGCSE last year. At FOSDEM, the question was 'why should I use OpenSolaris instead of Linux?' For me, this is a much easier question to address, frankly. It also demonstrated to me that almost all of the visitors to our booth have working knowledge of Linux and have used it, in its many flavors, extensively for some time.

So, why should one use OpenSolaris instead?

-Observability tools;
DTrace provides extensive debugging of the system and applications. It is available on Mac OSX and a project is underway to port it to PowerPC. If there was an interest, a demonstration of simple DTrace outputs was helpful to most developers.

-Virtualization;
Zones, containers, and resource management components in OpenSolaris provide virtualization infrastructure to enable a huge number of development environments on a single machine. This technology likewise provides isolation of processes and applications for testing. Branded zones enable one to run applications optimized on other platforms, on the OpenSolaris platform, to take advantage of unique OpenSolaris tools.

-File System;
ZFS is a revolutionary file system that is easy to administer and eliminates the need for volume management. Most students knew about ZFS and have heard incredible stories from peers about the functionality.

-Security;
Role-based access controls in OpenSolaris enable secure separation of rights for fine grained management of privileges. New user interfaces for privilege setup and configuration make this concept easy to understand and implement on-the-fly. Trusted extensions enable process-level controls for each running instance per session, so an individual process is restricted by its trusted meta-data assignment, not only by the privileges assigned to a particular user.

-Java;
OpenSolaris includes complete integration of the latest Java Runtime Environment.

-Documentation;
OpenSolaris provides documented support for all active features of the operating system, comprising thousands of man pages and hundreds of technical developer and administrator manuals. The Documentation community provides source files for the most active of these documents so you are able to customize the information to your needs, feedback on documentation usability and accuracy, or request to participate in new information development. Extensive training and certification programs exist for Solaris administration and development in addition to excellent free online training, articles, and lively forums that give you access to the experts and supportive communities of all types of users and developers of emerging technologies.

With many thanks to Erik and Steve for their thoughtful comments, I've updated this draft curriculum. More comments and thoughts are encouraged. C++! It's just too funny that I had C++ in there, I must have been brain dead from travel or seen too many of those hilarious x-mas Apple commercials...

[Read More]

This blog copyright 2009 by MissMichelle