Monday Sep 01, 2008


Despite having half a dozen computing devices within reach at any time (Eric Schmidt was right in 1995 when he said "In the future your body will have 5 IP addresses; where you put them is your business"), I'm still a stickler for sticky notes, notebooks, scribbled thoughts on the back of tear-off calendar pages, and literally back of the envelope calculations. At one point I had two "work notebooks" (one new, one old), a collection of stickies on my home iMac, another set on my ancient TravelMac, and an assortment of things that seemed to accrete in the reporter's notebook I keep in my car's center console.

I crave simple. I still use vi for editing text files, and I send email without capital letters (this is because of a long-standing belief that "chording" on the keyboard contributes to carpal tunnel syndrome, so I attribute 28 years of healthy paws to my devolved typographical style. Thanks, pep). At the same time, I am constantly tearing items out of print materials or bookmarketing them for later blog ideas (not at all obvious from the lack of recent writing) and need to be able to sort out work ideas, the "honey do list", and a pun waiting for a blog to congeal around it.

Enter Evernote. It combines the text editor like functions that us habitual list makers crave, with a web clipping function, tagging for easier search and organization, and it has a variety of mobile interfaces. Voila. I can start an idea at home, finish it in the airport, edit it or review it on my (somewhat despised) iPhone, and even print off an old-fashioned hardcopy for the close-paren function of the habitual list maker: ticking things off when they're done.

In ten days it's become a top-five application along with Firefox, Thunderbird and OpenOffice.

Wednesday Mar 26, 2008

Getting the "H" Out Of Yahoo

The hullabaloo, hoopla, hearsay and hand-wringing over the Microsoft-Yahoo deal has me more than mildly amused, and not just because it would fill the H-stanza in King Crimson's (entirely appropriate) Elephant Talk. Somehow, Google thinks that the deal will damage the internet in some way that having opaque code that colors our search glasses doesn't. Having one overwhelming path to an answer, whether it's search, email, personal productivity apps (there's a misnomer), or advertising, is bad for the internet. I'm regularly reminded of Larry Wall's comments about the ornate and duplicative nature of perl: it's is about having many ways to accomplish the same thing, it's about having paper and crayons and not just an Etch-a-Sketch with which to express your crudely drawn ideas. That's what makes the internet good, frightening, disruptive, harsh and useful, all at the same time: there are many paths, some of which seem wrong but all equally transparent.

Personally, I think the Microsoft-Yahoo mash-up is a Truly Bad Idea, but not because of competitive or market forces. It's something that "can only end in tears" to quote one of my favorite staff peers. I'm neither a financial analyst nor do I play one in my blog, but on the surface, the deal seems to have interesting, material impacts like accounting for the premium paid, and the minor bump in net income to cover other Microsoft experiments in media. And finally, I don't see how owning an advertising platform makes Microsoft into a media company, if that's where their joint strategic vector is aimed. It's difficult to discern sources of other operational leverage, unless Microsoft wants to re-engineer Yahoo's platforms (pure cost always has a bad return on investment unless the future cost savings are measured in orders of magnitude, not percentages). The New York Times summarized the situation quite neatly by comparing the deal to building a spaceship out of spare parts.

So what bugs me about the whole deal, and makes me smile to see Google grating? It's a long prolegomena to any future meat-physics, built on three premises: (1) The difference between advertising platforms and media companies; (2) Yahoo's creed of "connecting people" and (3) the decreased utility of thinking in hierarchies.

Yahoo is an advertising platform, not a media company. True media companies make their money from end users paying for content, and in a world of long-tail economics, they grow those markets through other types of content -- blogs, linked references, word of mouth, email, tagged content, or just about anything else that expresses opinion or organizes ontologically related thoughts. The "head of the tail" will still be advertising driven, but the long tail doesn't rely on ad displays, click throughs, or content analysis to drive volume. Part of the itchiness here is that Microsoft only makes those "head of the tail" products, and monetizes essentially nothing in the long tail. So buying an advertising platform aimed at other heads seems reasonable -- and is a reflection of the media company == advertising platform structure that brought us the writer's strike.

Yahoo's creed is to connect people to their passions. I love ice hockey, eBay, baseball, the New Jersey Devils, Princeton, and my family, and aside from some very infrequently used Yahoo group-based mailing lists, there's zero net interaction with Yahoo there. I used flickr but switched to SmugMug; anything I want to tag I put on Facebook. I'm not sure where the connections are supposed to come from, and I certainly don't find clumps of targeted ads useful in these narrow, single-topic facets of my life. So Yahoo hasn't helped me find any other parts of my interest taxonomy, nor have its ads stimulated me to buy things based on intersections of those interests. Some might claim that I'm not taking full advantage of the Yahoo experience, but that's because I find the experience too much like using the Yellow Pages (the alphabetical, printed kind, not the maiden name of NIS). I don't want more content, I want more organization. Which leads me to...

The "H" in Yahoo means "hierarchy". Perhaps the root cause of Yahoo's consumption stems from holding on to the "H" in the likely ex post facto Yang and Filo acronym for too long. Hierarchies imply that you start at the general and work your way to the more specific; in the advertising platform of the future (or better yet, the media company growth platform of the future), you'll go from specific to specific. When I do my annual Locus magazine sci-fi shopping run, I'm not starting with "Books" or even "Space Opera" but rather looking for additional input from writers and editors whose output I trust, and in three years of picking my beach reading this way I've discovered half a dozen authors and consumed all of their works. No click-throughs, no banner ads, and no in-store display required.

That, in a nutshell, is the issue: connecting people to concepts (things, communities, ideas, markets) is a graph problem, and if you want to focus on the graph problem, you need to think about the edges of the graph, not the nodes in the graph. Let the nodes (the content sites, the media companies) flourish, and build a better, richer, more wonderful graph traversal experience, and you have succeeded. It's why I read BoingBoing and why getting slashdotted is so powerful. It's why FaceBook has promise, once it figures out how to get around the walls of its content garden.

But unless Microsoft dramatically changes its designs on media company status, I believe the focus will be on the nodes in the graph, and advertisements for those nodes, rather than the edges connecting the nodes. And at some point, the "Yet Another" part of the acronym expansion dominates the end result, while the "Officious" and "Oracle" potentials - to create node-to-node links that nobody suspected were interesting until they discovered passion in their targets - are forever encapsulated in the cells of an M&A spreadsheet. And that's something to cry over.

Saturday Feb 23, 2008

Help Me, I'm Melting

Like a lot of folks, I use my Sun blog for work- and technology-related things, and keep my personal rants, raves, faves and professional sports bordering on mythology thinking over on a personal blog. In addition to having a never-ending notepad for writing ideas, it's a playground for tweaking PHP code, WordPress themes, and mySQL databases. That's the good news -- the bad news is that I've been using iPowerWeb to host it, and they are suffering scalability issues.

I believe there's a common misperception that deploying open source software is somehow easier or less difficult than putting commercial databases and web servers into production. The acquisition cost is less, but the deployment engineering is the same. This is hard work -- network engineering, scalability design, reliability engineering, recovery design, user self-help and ticketing systems, and instrumentation so you detect problems before your customers come calling. Or try to.

I spent 47 minutes on hold tonight with the iPowerWeb support line, only to get an "expediter" rather than a technical specialist. And the only reason I called is that the trouble ticket that I opened online hadn't received a response in over 72 hours. And what prompted the whole "I need support" blog existential crisis is that since iPowerWeb moved my snowman over to their new hosting architecture, performance has cratered. I use "crater" with such derision that the Moon and/or parts of Arizona might take offense; upwards of 30 seconds to load the WordPress index page and 90 seconds to insert a new entry in a single category, with only 2-3 tags and a modified timestamp seem excessive to me. It's nota WordPress, mySQL or underlying OS problem, because two other sites hosted by the same company on their down-rev infrastructure are still snappy and happy.

This is, I believe, a case of virtualization gone south, in a big way. Whatever they did in separating out the mySQL servers in an attempt to build a more efficient, multi-tenant database engine has resulted in something that's bad for everyone. It can only end in an ABEND (in particular an S522 for those of you who spent undergraduate Saturday nights waiting for jobs to complete).

Friday Feb 15, 2008

Squawk: Big Noises From Small Things

I'm fond of saying that computing is getting more transparent. Your mental image of the "computer center", a big brick building with the mainframe and some high speed line printers, is a Currier & Ives view of how people interact with computing devices. We're all socialized to the idea of small, portable, personal devices with rich computing platforms, but we haven't really tested the limits of rich, interactive computing on a small scale. I won't count the dozens of microprocessors in my car as a "rich" environment, because I can't (and really don't want to) code to them.

Enter Squawk, the micro embedded Java virtual machine and development environment from Sun Labs. In the latest Innovating@Sun podcast, principal investigator Eric Arsenau and I get really small and talk about embedded programming from a technical, financial and cultural perspective. SunSPOTS, the sensors based on the Squawk technology, have a vibrant user community already: my favorite application so far is the UhlerBot because it is a viable robotics platform built out of hobby store parts.

The cross-disciplinary possibilities for small devices encompass social sciences, environmental sciences and the arts; by open sourcing the platform and the hardware specifications we hope to see creative new applications driven by the transparency of the technology and its consumption.

Wednesday Jan 23, 2008

MySQL Dolphins in a Strategic Blue Ocean

To quote Peter Wolf, this one has a little introduction, starting just across the river from the WBCN studios where the Woofer Goofer got his start. If you have heard of "blue ocean strategy", skip to dolphin puns below.

Otherwise, my take on the Sun-mySQL deal starts with an article in the Harvard Business Review by W. Chan Kim and Renee Mauborgne. Their proposed blue ocean strategic model looks for previously non-consumptive or non-existent markets. The corresponding premise is that "red ocean" markets are those already in existence, with each player growing as a function of overall market growth and through taking market share from others. The blue ocean strategy focuses on finding new, non-consumptive markets based on the relative value of product or service features demanded by consumers in those markets. The disruptions come from combining the most valued features in non-intuitive ways to create a new market. Bagged lettuce combines the produce aisle and the convenience of the prepared foods aisle in the supermarket; nobody knew it would be a billion dollar business. One of my favorites (indicated by waistline) is Pret-A-Manger, the UK based sandwich shop that combines the speed of ordering lunch in a fast food outlet with the fresh ingredients and healthier eating choices of a local deli. The name itself is a play on pret-a-porter, the notion of high-end clothing (or food) ready to consume without the time intervention of a tailor or the guy slicing turkey one sandwich at a time.

Blue ocean strategies overlap in two ways with the digital world. First, open source software is a key entry point to non-consumptive markets. The best way to get someone who has never used your software to try it, evaluate it, or take an interest is to remove all barriers to entry. The analysts (and occasional) customers who ask me "How will Sun make money by giving things away?" miss the fact that "giving things away" is a blue ocean strategy that expands markets, while "making money" is a red ocean tactic to compete and take share in those newly entered fields of play. The second dip of the network endpoint in the blue ocean is the use of blue ocean strategic thinking to define new, small markets and identify the attributes that drive consumers to value them. It's Chris Anderson's Long Tail as seen by an MBA, not a web site developer.

Now I'll put those mySQL dolphins in the blue ocean.

I'll argue that mySQL -- and open source databases in general -- form a blue ocean of opportunity. Put another way, customers who want Oracle, Sybase or DB2 are looking for vertically scaled, commercial software that tyipcally has a suite of packaged commercial applications on top of it. Very red ocean, coming from a guy who has seen his share of Jersey shore red tides. Anyone who would buy Oracle (or others) probably wasn't looking at mySQL in the first place. Conversely, mySQL customers probably didn't consider Oracle or DB2; they wanted a free database with a different (not better, not easier, not fewer, but different) set of prioritized features. The purest definition of a blue ocean. They are willing to do more work on availability at the application or monitoring level and use shards to scale horizontally. Different features with different priorities: blue ocean there.

[Sidebar: if you look at the database market in terms of red and blue oceans, it makes Dvorak's diatribe on the situation amusing in both its simplicity and disconnection from actual business strategy. Then again, Dvorak claimed that Motorola was poised for cell phone dominance just 9 months ago, and today's news kind of throws water on that theory too.]

I've read a number of other theories and comments on this deal that seem to ignore the actual physics of the open source, middleware and database software markets:

It's anti-(company) or meant to disrupt (company). Actually, it's a new market for us. Doesn't change, dismiss, disrupt, or interfere with any of our other database partners. Most of them have application and web servers, too, while we have the open source Glassfish project, and the world hasn't ended. On the other hand, Sun has had a database group for at least 15 years, and when you put some really creative database people on these types of problems, well, maybe Solaris and our systems and storage products get better for all databases.

mySQL has a small revenue stream and its customers are small businesses. True, my personal sports blog is written in WordPress, and therefore I am a mySQL customer (indirectly, through the company that hosts my blog), but Google, Facebook and a host of other brand names use mySQL too. See previous comment about companies that want free database software and wil invest in the upper layers of the stack. See also "non consumptive markets".

Sun doesn't do acquisitions well. The X64 products based on the Kealia intellectual property, our market-leading identity management suite build on top of Waveset, the corresponding professional services practice that grew out of Neogent, and the ever-expanding community say "bzzt, but thanks for playing the home version of our game."

Sun will make mySQL proprietary. Theoretical computer scientists love trapdoor functions, that is, a function that only goes one way and can't be reverse engineered even if you know the full range of inputs and associated outputs. Open sourcing a project is a true trapdoor function: once the code is released, attempts to make it "proprietary" amount to trying to reverse engineer the trapdoor. The bits don't go in that direction. And I'll add that optimizing code or adding features don't count as making something proprietary. Again, bad behavior generates vitriol and and reversion; the code gets edited out by the community while the comments get piled on. This week's example involving Apple and Dtrace has a comment thread that runs from the DMCA to correctness of performance data.

Bottom line: the mySQL deal is pretty exciting. Normally, I reserve the redline of excitement for Really Big Things like the NFL Giants in the Super Bowl (check), Princeton ice hockey riding the ECAC (check), or my youth hockey team in the playoffs (check). This deal has the same sense of huge opportunity for all involved.

Friday Jan 18, 2008

Abstraction and Ingredients

I stopped reading package ingredients a few years ago, after getting regularly depressed that what I considered a "snack size" actually contained an entire day's worth of calories, fat, sugar and not nearly enough vitamin content. Let's face it: when you pop open one of those airport snack packs, you aren't thinking about how you'll divvy it up into three servings; you're looking for a sugar fix and you're less concerned about what other things it drags along. Not exactly heart-healthy and aligned with the much-rumored new year's resolutions for 2008, but immensely practical.

Packaging and abstraction drive use. If something is easy to consume, you'll consume more of it, and if it presents an abstraction that hides the ugly {technical, political, nutritional} details, you'll find it easier to use. That, in a (healthy) nutshell, is the motivation for Project Caroline, a Sun Labs effort focused on simplifying service deployment and delivery. (For anyone questioning my failure to craft a pun around Sweet Caroline, it's an anti-New York Rangers sentiment bubbling through the otherwise sugary stuff here).

VP of Advanced Development Rich Zippel and I sat down with our USB microphones to record another Innovating@Sun podcast, extolling the virtues of Caroline in all of its nuts and bolts, from the motivations for a simple service platform to the derivation of the name (again, no Neil Diamond, thankfully).

This is a much bigger deal than a research project and an attmept to rationalize the array of interfaces presented to deployers, not just service developers. The biggest challenge for data center designers today is not choosing a virtualization platform or a networking switch vendor or even a cooling technology. Those are implementation details (large details, to be sure, but details). The challenge is balancing the needs of the CTO or VP of Application Development to "go fast", creating more value in IT for the business, with the needs of the CIO, who wants to "go slow," controlling the rate of change, mananging risk, and squeezing as much utilization (and efficiency) out of the computing assets as possible. The path to achieving this balance doesn't involve Xen (sorry) as much as thinking about abstraction across the entire array of applications: networking, computing, storage, sessions, data caching and persistence, and language.

Moving up to a higher level of abstraction for the data center means that you're less concerned about how it's built and more concerned that it "just works" -- an artifact of what Sun CTO Greg Papadopoulos calls "Network Scale Computing." That's the intent of Project Caroline, and it's a message that has resonated with every CIO with whom I've met in the last two months.

Thursday Oct 25, 2007

The DHCP Server Ate My Homework

For two decades we've heard various theories about how computers were going to change education: make us learn faster, learn more, or expose us to new educational techniques. The bottom line is that computers have served to reinforce the social side of learning -- finding people with similar questions and problems, locating specific answers to specific questions, and making it far easier to detect potential plagiarism using search techniques. While we still learn through most of the approaches that have worked since Biblical times -- small groups, discussions, subject matter experts sharing their knowledge -- technology offers us significant opportunities to improve the context for learning.

Moodle is an open source course management system -- a blend of content management and educational logistics. "Moodle" has become vernacular around our house, as both of my kids use it daily for homework, after school class discussions, and to retrieve notes that would have been on infamous purple ditto sheets in my day. The audience for course management is tough -- they're used to iPods, video on demand, wireless service, and game consoles, so anything that is slow to respond, has a funky user interface or isn't reliable is going to generate a homework excuse. "The DHCP server ate my homework" hasn't popped up yet, but it's a matter of time.

Earlier this week Stuart Sim, CTO of Moodlerooms, joined me for an Innovating@Sun podcast to talk about how they're building out Moodle instances for the most demanding consumers of all (teenagers), at maximum scale, and making money from this open source project all at the same time. Our after-school program covers the spectrum from Moodleroom's use of Niagara-based servers to why term papers might become an historical artifact once and for all.

Tuesday Oct 16, 2007

DRM in Hollywood, with a shout out to Dr. Demento

Greetings from Hollywood, California, home of the silver screen, recording studios, Doctor Demento, and earthquakes. I'm speaking on a DRM panel hosted by the media business law firm Foley and Lardner. I'm one of three technology folks in the room, and I'm going to talk about avoiding legal and technical decisions that limit our future rights, opportunities and markets. I don't think I'm going to be popular; perhaps I should not have worn my Diesel Sweeties pirate shirt; "pirate" foments a violent reaction among this group of legal media ninjas.

And if you surf over to the good Doctor's web site, you'll see that he's switched from a nationally syndicated, advertising supported model to a subscription service. Dr. Demento was a Sunday night staple, fitting into the time slice of my teen years between the Wonderful World of Disney and Sunday night NFL games. The shift away from national syndication doesn't mean that there's no longer an audience for his particular brand of wackiness, but rather that audience isn't sitting by the radio on Sunday night, aggregated in one time slot. Now you can listen to what you want when and where you want, without the FCC's censorship. But it's not "free" in the sense of advertising supported radio being free. If you're willing to spend $2 for a sophomoric laugh when you need it most, it's a good deal.

Tuesday Oct 09, 2007

Madonna, Ichiro and Andy: Big Name Movers

If you've spent any time with Sun co-founder Andy Bechtolsheim, you know that he's very, very high bandwidth. Once he gets going, he's a stream of pure information. This makes conversational style content pretty difficult, because you don't want to slow him down and lose bits. So I opened up our Innovating@Sun podcast with Andy by comparing him to Madonna and Ichiro -- two players who don't need surnames to move mass markets.

In this case, the market being reshaped is that of "commodity computing." This isn't anything new to Andy, since the very first Stanford University Network workstation was a composition of a Motorola processor, a commodity Ethernet chip (at the time) and the commodity operating system for the then-popular DEC VAX systems, BSD Unix. The emphasis isn't on the commodity components; they are defined by intersecting supply, cost and demand curves. It is, as it has been, about building a computing environment that scales in terms of power, space, cooling, reliability and sustainability. Andy goes through everything from common chassis and physical design to the convergence of servers and storage controller architectures to why it's good to have developers for your routers. He speaks from authority, having done just about everything in the list personally, at, by and for high bandwidth.

Monday Oct 01, 2007

Long Tale of Non-Commercial College Radio

Commercial college radio provided me with my first experiences in sales. As an advertising sales "rep" for WPRB-FM, Princeton's student-run radio station, I had to pitch ideas, produce ads, write copy, frequently voice the ads myself, manage our cash stream and do demand generation. It was a great way to finance my growing record collection. There are only a handful of commercial college stations; most are either financed by the affiliated institution or through listeners, much like the Public Broadcasting System ("supported by viewers like you").

But WPRB-FM has a long and technologically illustrative Tiger tale. As the first FM college station, it received a broadcast frequency of 103.5 FM, later swapped for cash and the equally useful 103.3 FM. Most student-run stations are banished to the lower end of the frequency spectrum, where they're less likely to be found by "dial twiddling", something that worked in favor of listeners in the Trenton-Princeton-New Brunswick Route 1 corridor who let go of the tuning knob when something on the radio make them tune in a bit more closely. A broadcast format that included classical and jazz provided an advertising platform for local businesses: jewelers, travel companies, the University Store, and some higher end eateries. The station was, and is, self-supporting without any assistance from Princeton University aside from the use of space in a dormitory basement, and (at the time) a spot for the broadcast antenna on top of the same building.

Afternoons and evenings featured WPRB's "progressive" format, today probably known as alternative, eclectic, or just fun. It included everything from truly obscure local bands (like Regressive Aid and their conjoined one-show band with The Groceries, the aptly named Lunchmeat 2000) to bands that nobody had heard of (at the time) like REM. Our listeners were as varied as our on-air programming; I'm sure that most of my classmates who frequently voiced "WPRB plays crap" as we spun U2's Boy now own copies of Bono vox and How To Dismantle An Atomic Bomb. We had shows that featured reggae, art rock (think Yes, Genesis and King Crimson), metal and alternative, and during one weekend when we were moving a wall, all 20-minute plus tracks so that we could swing hammers and not just bass lines. Our varied programming after lunchtime let us pick up "punk clubs" (King Tut City Gardens) and various other counter-cultural havens as advertisers who had no other routes to listeners.

College radio, powered by a desire to disrupt convention, and expose listeners to something new, is the epitome of a long tail. Only through college radio could my friend P hear Alaskan punk band (don't ask) "The Anemic Boyfriends" on WFMU-FM and declare them high art. The essence of Chris Anderson's long tail economics is to drive more overall volume by first expanding the "tail" of a distribution curve, and then moving demand from the head of the curve (smash hits) to the tail (future micro-smashes). The problem in the 80s was distribution -- as desperate as P was to find the 45 RPM single of The Anemic Boyfriends, it took nearly a year and a trip to the ferro-ciously good used record store in Ithaca, New York, to find the vinyl.

No matter how long the tail, at the end of the distribution curve, there's another distribution curve of even more refined, more obscure listeners. It's what led Dave, one of my managers who knows I have a penchant for being a Yes-man, to point me at Porcupine Tree as the Yes of this generation. He's right, and I spent $45 on content I would have never found through any other channel. And at the transitive closure of those distribution and demand curves, you'll find Indie Rock Pete from Diesel Sweeties, afraid to like anything with a positive listener count.

In the mid-80s, WPRB went through a financial change brought about by the confluence of an expanding New York radio market making our 103.3 FM frequency more of an impediment to our neighbors on the dial who wanted to go bigger. Through several years of negotiations, FCC filings and long meetings, WPRB was able to expand its broadcast area and received a cash infusion at the same time. Numerous exhaustive and exhausting discussions ensured about the financial models that would govern how and where the money would be spent in future years. We debated the risks of investment strategies, regulatory issues and continuity of student leadership, and yet it was forces exogenous to the broadcast industry that reshaped WPRB 20 years later.

When you mix podcasting, blogging, and social networks as ways for students to share their musical likes and dislikes, the on-air pulpit is less appealing. Running a commercial station with a unique value proposition is much harder when that same value-through-unique programming can be obtained with a combination of Google and iTunes. WPRB has switched from a commercial format to a listener supported financial model, still maintaining the quasi-independence of Princeton's direct sponsorship and the fiercely independent creativity in its programming. College radio is far from dead. It does, however, rely on direct support of listeners who delight in hearing something new, finding a reference to it on a web site, and monetizing that interest almost immediately, without intervening ads for unrelated products.

Hall of Fame 2.0

Hall of Fame 2.0 The National Baseball Hall of Fame is a physical and emotional experience. In addition to honoring the great players and builders of the game, it's a repository for the artifacts of great accomplishments, records, and the culture of baseball. As the only team sport measured in defensive success, not on a clock, it encourages us to think of time as malleable, our thoughts drifting between this year's excitement and the youthful memories that first made us fans.

Unfortunately, the Hall feels like a museum, which it is, rather than a glimpse into the collective memory and celebration of the National Pastime. Less than 10% of the balls, bats, bases and beauty of the collection is on display at any time, and the organization of the displays makes it hard to formulate a story out of what's there. Baseball tradition, like religious tradition, is passed on through storytelling and personal action; it's parents telling their kids about famous players, great plays, or playing the sport. It's my father telling my son about a mutual friend who played for Honus Wagner in Pittsburgh, or me telling my kids why and how Willie Stargell inspired me to choose #8 when possible (even when going through toll booths), and at some point in the future, my daughter telling my grandchildren about the night we went to the Giants game to hope that Barry hit 756 into our section of the bleachers (we were a night early, but the memories will remain sans asterisque). It's the equal mix of seriousness and silliness that led me to hand out free ice pops to any Little League baseball or softball player wearing jersey #8, provided they let me say "for Willie Stargell".

Walking through the three floors of the Hall, I found Willie Stargell's plaque, a solo shot in the 1988 inductee class. Around the corner is one of his baseball cards as part of an exhibit geared toward younger kids, and upstairs in the legends of the game alleys, you'll find a Willie Stargell jersey paired with an exhibit about Roberto Clemente. So far, so good, and again, what anyone would expect from a first-rate museum. But on the third floor, the subtleties and opportunities for telling stories emerge. A 1970s World Series program has a page showing the buttons worn by Pirates fans, including "Chicken on the Hill," Willie Stargell's off-season employment and passion. Around the corner, there's the bat Stargell used in the 1979 World Series, where he was voted MVP, next to one of Kent Tekulve's Pirates hats. Sewn around the bucket-shaped lid, and across its top, are "Stargell Stars", player recognition given out by Stargell for particularly good play.

What was most disturbing was the lack of informed help at the Hall. Twice I asked logo-wearing employees where I might find Ron Bloomberg's bat (used in the first at-bat by a designated hitter), and was told it probably wasn't on display (it is, on the third floor in the records room, behind that of 1969 Met Art Shamsky, another item we sought). Conversely, every visitor had his or her own reasons to search something out, to tell its story of relative importance, and to take pictures that will highlight the next time those memories are revisited. While the inductees are selected by the Baseball Writers, it is the individual scribes of the game that truly relate the history of the game.

What's missing from the exhibit room is a place for all of us to share our own experiences with these tools of the baseball trade. To borrow a phrase from Tim O'Reilly, the Hall of Fame lacks an architecture for participation, a Hall of Fame 2.0, where user-generated content including pictures, stories, and our own interpretations can embellish the tools of the trade on display. Here's my ideal Hall of Fame experience: Knowing that you want to revel in Willie Stargell legend and lore, you can find all of the references to "Pops" and plan your own exhibit guide. Posted on the Hall's website would be an email from Stargell's niece explaining how Stargell stars were the one item his family asked for, more than autographs or baseballs, a reward to be given out. It makes Tekulve's hat more impressive, and more personal, perched next to Stargell's bat. I'd have a link to the Chicken on the Hill dining experience at PNC Park where a bit of Willie lives on, so that the World Series program makes sense in contexts both current and three decades old. I'd point to my Facebook picture album of number 8s from around the world, for the same reason NASCAR fans put driver numbers on their rides? Over time, as Willie Stargell said, the number comes to represent you in real life and not just on the roster. And finally, we'd have real merchandising, a place to locate the stores along Cooperstown's Main Street that sell licensed Stargell t-shirts, something to make the 4-hour trip home more comfortable.

The Hall of Fame board of directors is full of baseball management and talent, but no fans. No participation. Not even a hint of technology, from a sport that has always raced to utilize technology for the good of the game. Isn't it time that the fans share their knowledge and emotion, sometimes with religious fervor, in the shrine dedicated to the game's long-term history?

Wednesday Sep 05, 2007

Dark Star Modulo Lyrical Interpretations

I lived with a Dead Head for a year, and lived in his musical light cone for another year, so it's hard to say when exactly I first heard "Dark Star," one of the Dead's platforms for improvisation and fun (if you want a thorough treatment of the philosophical, metaphysical and less musical underpinnings of Dark Star, check out Steven Skaggs' essay on those very things).

Improvisation in code or in music needs a framework to carry it. It has to be accessible; it has to be easy to digest (at least for those not familiar with the rest of the author's work); and done well it both builds on the ideas of others and contributes new phraseology back. That's the end of the literal comparison between the Sun Labs gaming platform and its namesake Dead exploration.

In our latest Innovating@Sun podcast, I traded fours with Jim Waldo on how Darkstar hides some of the real-time and occasionally messy elements of building a game, how we can make it grow to "interesting" economies of scale, and why Java is well-suited for the highly charged (no pun intended) world of real-time development. Jim even runs with my obscure Grateful Dead references, which is fortunate since I was one step away from invoking the confluence of technology and John Perry Barlow by referring to Jim as the estimated prophet of Java.

Wednesday Aug 01, 2007

JavaFX from end to end

We've just posted the second part of the Innovating@Sun end to end JavaFX podcasts. In part one Sun Software co-CTO Bob Brewin talks about the motivations for JavaFX, including the difficulty of making "write once, run anywhere" a reality when "anywhere" encompasses a variety of client devices. It's one of the edgier episodes we've done because Bob isn't afraid to take on any of the established engineering norms for building, deploying and evolving applications. While Bob discusses JavaFX from the perspective of the developer, product line manager Jacob Lehrbaum joins me in part two to discuss what really happens -- and will happen -- on the mobile handset. From the foundation of the SavaJ handset environment to the inclusion of telco stack elements, JavaFX Mobile delivers the other end in the end to end future of the rich network application.

Monday Jun 18, 2007

The Mod(ular) Squad: Elliptic Curve Cryptography

A few years ago, after a long day (for a fifth grader) of studying long division, my daughter exclaimed that she saw no practical use for remainders. It reminded me of a similar day, sitting in a computer science class on computational complexity, of feeling that there was no practical use for knapsack problems. Both, it turns out, are the basis for many of the cryptography systems in vogue for online security and identity based systems. The exponential complexity that makes a problem intractable also makes it stronger in the face of brute-force attack, and the use of remainders (particularly the Chinese remainder theorem) makes it practically computable. Realizing that Professor Steiglitz was most egregiously correct (back in 1983) when he warned us that large prime numbers were in our futures, remainders, NP-complete problems and computational complexity all go "click" when I'm indulging my eBay habit.

Fast forward a few years: large-scale compute grids enable brute-force attacks against weaker (shorter key length) crypto systems, and increasing the key length to stay one or two hops ahead of the bad guys means additional drains on power, performance and time. Particularly bad things if you're worried about securing a data path to your mobile device, where power and time equal battery life. What's needed is a crypto system that uses shorter key lengths to produce a stronger system, and the click-fitting math this time are elliptic curves, providing a more efficient way to tackle the factoring problems underlying crypto systems. The result - elliptic curve cryptography - is a promising step in making systems more efficient and secure at the same time.

Aside from reading Simon Singh's Fermat's Enigma, which neatly tied together modular forms, elliptic curves, Fermat's Last Theorem, and Princeton University, I am, in the words of Napoleon Dynamite, in the need of some skills. For higher math, bigger invention and practical applications of all of the above, I had Sun Labs Distinguished Engineer Vipul Gupta join me for our Innovating@Sun podcast on Cryptography Breakthroughs. It's the current equivalent of being told that large prime numbers are in your future.

Wednesday May 23, 2007

15 Minutes of Real Time Fame

I have time on the brain today. Rob Kolstad of USENIX once wrote that systems administration is one of the few disciplines in which you deal with time scales ranging from the microsecond to the year (megasecond), depending upon what you are dependent upon (memory, disk, grep output or purchase order). As human beings, we are real-time by nature, and living in the metropolitan New York area, those real time bounds are short and regularly tested (it's called impatience and marked by car horns).

Real time, historically, was the domain of embedded devices such as nuclear reactors, industrial controls, and medical equipment. If missing a time deadline meant melting down or flaming out, you had to worry about application architecture down to the level of scheduling sub-tasks. How long your application would run without stopping, how long it might sleep without handling data, and squeezing latency out of every routine made real-time programming slightly less oppressive and tedious than raking a lawn full of autumn leaves.

Real time is enjoying a renaissance today, thanks to (a) an increased emphasis on managing latency out of financial applications (b) the abundance of media and media manipulating applications available on the network and (c) the ease of developing applications that operate within hard time constraints. Greg Bollela, Distinguished Engineer in the Java software group, sat down with me for an Innovating@Sun podcast about real-time Java. Greg discusses how these Java libraries manage the time management for the developer, abstracting away the all-too-real bean counting at the CPU cycle level.

Applications are only as much fun as the data we have to feed them. What good is real time without a boundless supply of media types to synchronize, deliver and edit? With our ability to post videos of everyday life to YouTube, and better than hobby quality video editing capabilities on the desktop, we are enjoying a world equidistant between Andy Warhol's supposed universal fifteen minutes of fame and Spinal Tap keyboardist Viv Savage's motto to "have a good time, all the time."

We did just that -- Dave Cavena, systems engineer for many of the production and entertainment studio companies, and Bob Sokol, Chief Media Architect in our field organization, joined me to talk about what happens on the other side of the microphone, camera lens or sound board. We run the table from digital rights management, content transcoding, why tape (of all flavors) is still good, and what customer designs Sun has to offer in the digital content management space. It's fifteen minutes that preface delivering a customer or employee's online fame.


Hal Stern's thoughts on software, services, cloud computing, security, privacy, and data management


« July 2016