[A picture of private offices at Fog Creek Software] Alert! This ancient trifle retrieved from the Joel on Software archive is well-past its expiration date. Proceed with care.

Joel on Software

Explaining Steve Gillmor

by Joel Spolsky
Saturday, December 23, 2006

Nobody I know can understand a thing Steve Gillmor is talking about, mainly because he makes so many obscure references without explaining them. I thought as a public service I would provide a detailed exposition of his latest blog post, Bad Sinatra. My goal here was to explain just about everything, and I hope it’s not too tedious. It works like the Talmud. First I’ve got a paragraph quoted directly from Steve. I’ve taken the liberty of adding hyperlinks, which are not in the original article, and correcting an accidental edit which renders one of the paragraphs nonsensical.

Then I went ahead and bolded all the terms that needed definition, including names that were relevant. Those are defined in the “Rashi” section below each paragraph.

Finally, I added my own commentary in the “Tosefot” section. The idea is that the Rashi stuff helps you understood what Gillmor literally meant, while the Tosefot section contains more in-depth commentary addressing the actual points he’s trying to make.

It took me over three hours to research and explain all this, and, as you’ll see if you follow closely, Gillmor’s entire argument fell apart under scrutiny, so I don’t think I’m going to be doing in-depth explanation like this again.

So without further ado, here’s Bad Sinatra, with Rashi and Tosefot.

Bad Sinatra, by Steve Gillmor.

Jonathan Schwartz has a problem. Me. I read his blog today, starting with the most recent post and eventually landing on one a few days ago about the resurgence of the thick client. Let me weigh in thusly : what a load of shit this is. When Sun leadership starts moving away from the Google sweet spot and toward god knows what users-love-client-code idiocy, it’s big trouble for the Sun boys. Dave Winer is swimming in muddy waters too with his anti-Lucovsky Google-is-deprecating-SOAP of all things in favor of the Ajax RESTian Web-only API that Mark is evangelizing on Scoble’s show.”


Jonathan Schwartz: CEO and President of Sun Microsystems and blogger. Thick client: A software architecture in which custom applications are run on the desktop talking to a server. Compare to thin client, in which only a web browser is run on the desktop. Sun leadership: Schwartz. The Google sweet spot: Ajax, that is, web-based applications like Gmail which are neither too thin (plain HTML) nor too thick (heavyweight Java applets that need to be downloaded or custom applications) and with just the right compromise between functionality and simplicity. Users-love-client-code idiocy: Theory that desktop applications provide a better user experience. Dave Winer: Blogger and pundit. Lucovsky: (Mark), Google developer formerly of Microsoft. SOAP: Complex protocol for remote procedure calls, implemented by sending XML over HTTP (usually). RESTian: Using REST, a much simpler alternative to SOAP, implemented by sending a URL to a web server and getting back XML (usually). Mark: Lucovsky. Evangelizing: promoting. Scoble: (Robert), blogger and video interviewer.

“First, Jonathan. I went to a Sun press party tonight that was remarkable in its failure to deliver any promised executives. Dan Farber found a few lurking in the rear of the room, but no Jonathan, no Papadopoulos, no Fowler. I found myself longing for the good old McNealy days, when Scott’s Microsoft jabs and hockey jive kept the room moving. A Google party a few weeks ago was in full swing before Sergey and Larry showed up, and their presence almost came as an afterthought. Marc Benioff threw a luncheon to announce the latest iterations of the Salesforce build-out, and instead of playing to the middle of the pack, excelled in a detail-rich deep dive into his company’s mining of its customer base as the evolution of Microsoft’s developer strategy. That’s entertainment, folks.”


Deliver: be attended by. Dan Farber: editor of ZDNet and CNET. Papadopoulos: (Greg), CTO and VP of R&D, Sun Microsystems. Fowler: (John), EVP Systems, Sun Microsystems. McNealy: (Scott), ex-CEO Sun, preceded Jonathan Schwartz, characterized by almost perpetual denigration of Microsoft in public. Hockey: McNealy liked ice-hockey. Sergey: Brin, co-founder of Google. Larry: Page, co-founder of Google. Marc Benioff: CEO, Salesforce.com. Salesforce: Web sales-force management software. Playing to the middle of the pack: Avoiding technical detail so as to remain comprehensible to mainstream journalists.


Mining Gillmor is president of and was a co-founder of AttentionTrust, a non-profit organization tightly connected to the for-profit startup RootMarkets, so it’s safe to assume that he agrees with their belief that users on the Internet should own their attention data. For example, if a user goes to Amazon and visits certain product pages, Amazon obtains data about what that user is paying attention to. Under current practice, Amazon “owns” that data. Briefly, the AttentionTrust/RootMarkets idea is that the user should have a right to control that data, delete it from Amazon’s servers if they choose, sell it to the highest bidder, etc. The “sell it to the highest bidder” part is where they hope to make their money, creating an exchange, Root Exchange where user’s “intentions” can be traded. In practice this means trading leads: contact information for customers who have expressed a desire to purchase some commodity. For example a user might go to a website and say they’re buying a house and want a mortgage; this lead would be traded on an exchange just like a stock or bond and lenders who wanted to write mortgage loans would bid on it. Participants in this exchange would be (presumably) required to abide by strict privacy rules and the principle that users own their attention data (i.e. they have the right to control the information that they are interested in a mortgage right now), which makes this an attempt to create a legitimate and ethical version of what has traditionally been a shady, boiler-room operation characterized by illicit markets for names and phone numbers of the elderly and the feeble-minded who are likely to fall for scams (c.f. Glengarry Glen Ross). Whether it works or not remains to be seen. In any case it is fair to assume that Gillmor strongly objects to Benioff’s mining his customer database, which, according to the AttentionTrust philosophy, amounts to theft.

Microsoft’s developer strategy Benioff and Salesforce.com have a history of overstating the importance of their own work. Even though their product is primarily a sales force automation product, Salesforce.com advertising historically focused only on the way it was delivered, as a hosted web application, as if this was a more important aspect than the functionality of the web service itself. I was at a Salesforce.com presentation that started out with a quick five minute summary of the evolution of programming (Eniac, mainframes, minicomputers, microcomputers, and web-based applications) and then presented Salesforce’s latest feature as if this was the next big step in the evolution of programming, on a par, say, with the shift from PCs to the web. The latest feature? Custom Fields. The highly savvy audience literally laughed at the pretense. Now, Microsoft’s “developer strategy” probably refers to their massively successful strategy of creating platforms, especially Windows, and doing everything imaginable to make it easy and profitable for developers to create software for that platform, thus strengthening the value of the platform and creating massive network effects. Presumably what happened at the Salesforce.com luncheon was that Benioff described his company’s efforts to mine its customer data as if this were an accomplishment on a par with Microsoft’s greatly successful strategy of courting third-party developers. Gillmor mocks the hubris of this comparison. On the other hand, it seems more likely that Gillmor might just have been conflating two things Benioff was talking about: their own customer data mining and Salesforce.com’s new platform strategy which encourages developers to write their own enterprise applications to run on Salesforce.com.

“I thought I would miss the end of the Gang more than I have. Mike Arrington’s flareup with Sethi and his revolving news desk door would have made for a lively session or two, but Mike’s week-later wrap of the Sethi situation blew away the bullshit and out-ValleyWagged Denton to boot. I’ve been pitching in with Jason to tighten up his Cast for the Kids, letting me troll the blogosphere for news bites without having to work too hard or miss the roundtable so much. I spent a great three days or so with Doc, Phil Windley, Kaliya, Dick Hardt, and Kim Cameron among others in the Identity Workshop, talked with Mike Vizard and Dana Gardner intermittently, and watched Robert Anderson and Cori Schlegel cross the line and merge GestureBank with the AttentionTrust only a few weeks behind schedule. And Gabe Rivera chimed in a few days ago in email asking what it would look like if I resurfaced. An eight-foot invisible rabbit.”


Gang: The Gillmor Gang, Gillmor’s group podcast, which recently ceased production. Mike Arrington: Editor and creator of TechCrunch and a sometime-member of The Gillmor Gang. Sethi: (Sam), former editor of the UK version of TechCrunch, fired by Arrington for reasons which this margin is too narrow to contain. Valleywag: Snarky, for-profit Silicon Valley gossip site. Denton: (Nick), creator of Valleywag. Jason: Calacanis, occasional member of The Gillmor Gang. Cast: Calacanis’s new podcast. Troll the blogosphere: Read weblogs. Roundtable: The Gillmor Gang. Line: Finish line (he doesn’t mean crossing an ethical line). GestureBank: Gillmor had left AttentionTrust for a period of time to set up GestureBank, which, with his return to AttentionTrust in September, was merged in.


Lively session or two: The Gillmor Gang often devoted a substantial amount of time to discussing fights and rows; most of the final two episodes were devoted to Jason Calacanis’s departure from AOL.

Rabbit: Most of this paragraph appears to be here primarily to bring us up to date on Gillmor’s latest doings, it is not related to the entry’s main points about Jonathan Schwartz’s comments on the thin client or Dave Winer’s comments on web protocols.

“The thick client: why is Jonathan floating this lead balloon? To get some blogoversy? Probably. His slick sales pitches don’t register, and one of the best guys on his feet in the business seems hamstrung by his day job. He should take a look at what Jon Udell is doing. The new show business is actionable information on demand, spiced with a healthy disrespect for marketing bullshit and strategic kindergarten. Take Edelman’s alliance with Newsgator to provide pre-gamed conversations. Nice: avoid the messy run-up to PayPerPost stench, and go right for the stupid notion that people won’t immediately look not at the approved conversation but at the telltale odor of the missing links. Attention: the “customers” are listening now. Think of the stream as meditation.”


Blogoversy: Weblog controversy. Slick sales pitches: Most of Schwartz’s weblog posts are earnest and unsubtle sales pitches for Sun’s products. Don’t register: As a result Schwartz’s weblog has very little influence and is rarely linked to by other bloggers. It has only 2342 incoming links as reported by Technorati compared to, say, Scoble, who has 16,090. Hamstrung: Unable to write what he really thinks on his weblog. John Udell: Former InfoWorld columnist who recently joined Microsoft as a professional blogger. Actionable information: News You Can Use. Strategic kindergarden: Childish strategies attempting to get publicity through weblogs. Edelman: Large, old, established public relations firm. PR firms attempt to gain publicity for their clients by creating news stories, as opposed to advertising firms which simply purchase paid ads. Newsgator: RSS software company. PayPerPost: Loathsome, despicable company which proposes to pay bloggers to mention products, pissing in the well and destroying what little trust we had earned. Stream: list of URLs which the web user visits in the order they visited them.


The thick client. After a brief sojourn in his social calendar, Gillmor threatens to return to the topic at hand, the critique of the blog post by Schwartz.

New show business. Jonathan Schwartz’s slick marketing weblog is ineffective because readers are on the lookout for marketing BS and attempts to create bogus product placements on weblogs.

Stupid notion. PR agencies have traditionally done very well by building up relationships with journalists and news outlets so that when one of their clients has a product to push, they can get news stories about that product into the media. This is far more effective than advertising, because readers consider editorial content to be more credible than the advertising, which has almost no credibility at all. As more and more attention shifts from mainstream media (MSM) to the web and especially weblogs, PR firms are naturally concerned about how to do PR in this new environment. Their initial attempts consist almost entirely of treating popular bloggers as if they were traditional media outlets, so, for example, I get at least half a dozen “press releases” every day from PR firms, written in the style of a newspaper article, as if Joel on Software were some kind of newfangled newspaper. These releases are sent by completely clueless PR professionals who have absolutely no idea what’s going on; not only are they clueless, but they actually think they’re clever by adding a few bloggers to their spam directories. What they’re really doing is training bloggers’ bayesian spam filters to ignore the company they’re supposedly promoting, and generating ill-will towards their clients by spamming on their behalf. Steve Rubel of Edelman has been blogging for quite a while and certainly knows more than most PR professionals how blogging works, however, the Edelman/Newsgator plan, besides being almost impossible to understand, won’t work, because readers will immediately see it for what it is: paid advertising. Gillmor argues that users will recognize it as paid advertising because it lacks hyperlinks which are characteristic of genuine blog posts.

Microsoft’s market standards into a commodity. Spare me the garbage that we need real code running on the client. If you want to know what a vendor is afraid of, figure out what competitor they’re more afraid of. Is it Adobe with Apollo, eroding Java’s device penetration lead? It’s certainly not Microsoft? If you want to see why I’m so relaxed, go look at CrittendenIV’s post about this five tags bullshit. He mentioned me at the end so it bubbled up in my vanity feed, but there’s no way I can match or even come close to this guy. A star is born. By contrast, Jonathan’s bizarre thesis that the browser is a Winerian locked trunk is a) probably true, and b) so the fuck what. Mozilla forever put the lie to that theory when Firefox jujitsued Microsoft’s market standards into a commodity. Papadopoulos levels about Sun’s difficulty in driving sales in the Web 2 value chain no matter how right they are in their technology bets. Jonathan changes the subject to clients. Why?”


Apollo: Project by Adobe to allow Flash applications to run on the desktop on any operating system. CrittendenIV: Oregon blogger. Five tags: Blog game in which a blogger writes five things that their readers did not know about them, and then tags five other bloggers who are expected to then do the same thing. Vanity feed: Tracking mentions of oneself on the Internet. Winerian locked trunk: Metaphor by Dave Winer in which a large company creates a platform that allows small companies to participate in a market, but only as second class participants and at the whim of the large company. Papadopoulos: (Greg), CTO of Sun.


Microsoft’s market There appears to have been an inadvertent edit between the time the text was written and the way it appears on the web. I have redacted the text to the way that I believe Gillmor originally wrote it.

Firefox jujitsued The success of the web browser and the fact that it is a commodity, available for free on every platform, and the fact that developers now develop most new code to run over the web in the browser, has removed a lot of Microsoft’s market standards power. When all your applications run on a browser, Windows is just another commodity platform for running web browsers, no better than Macintoshes or Linux machines.

Levels Not sure where this “leveling” is taking place. Lately Papadopoulos has been talking about how the world only needs 5 computers, by which he means megacomputer clusters like Google, live.com, Yahoo!, etc. None of those clusters is running very much Sun hardware, so somewhere Gillmor might have read Papadopoulos admitting this.

Why? It appears that Gillmor is interpreting Schwartz’s claim in favor of “rich clients” as meaning that people will write Java applications to run on the desktop, as opposed to Ajax web applications. This is why he brings up Apollo, which, if successful, would directly compete with Java as a platform for building desktop clients. Gillmor makes several errors here. First, “Spare me the garbage that we need real code running on the client…” Gillmor himself considers Ajax applications in web browsers the “sweet spot”, and those involve an awful lot of real code running on the client. Second, in evaluating competitors, it is almost absurd to consider Apollo, which is still in beta and has virtually no penetration, as a serious competitor. It might become one, but that would be years away. The CEO of Sun doesn’t spend his days worrying about every project under development when he has real, current competitors to think about. (One of the frequent failings of pundits is a strong tendency to extrapolate wildly about trends which haven’t even begun yet, and assume that they will take over the world. There was a maddening dialog on one of the last episodes of The Gillmor Gang about whether Microsoft Office would move to the web, like Google Documents, in which Steve Gillmor and Jason Calacanis repeatedly talked about “flipping the switch,” as if porting hundreds of megabytes of Windows code to the web was a simple matter of “flipping the switch.” There was no recognition that web-based word processors and spreadsheets, while amusing, have merely a sliver of the capabilities of Windows-based apps.) The third mistake is even more drastic; Schwartz’s argument in favor of rich clients, if you read it, is actually an argument in favor of Ajax clients. Schwartz simply redefines those as “rich” clients. He writes: “In my book, it’s inaccurate to say Google or YouTube are ‘thin clients’ - they’re services that leverage someone else’s thick client. A browser…. Innovation on clients is back - we see it in a flurry of Web 2.0 companies investing in creative desktop interaction and the resurgence of JavaScript…” So actually, Gillmor’s entire “argument” with Schwartz is one based on a misinterpretation of what Schwartz is saying, and in fact they both seem to agree that the “Google sweet spot” is where the world is moving.

“Similarly, Dave is changing the subject from god knows what to JSON and Ajax API and for what reason? Dave is one of the most efficient if not the supreme political pragmatist, so why is he bringing up these subjects? Who am I supposed to be scared of? Google? Nope, if the Ajax API and the terms of service around including unaltered adsense are so counter to user interest, that will precipitate a decline in usage and therefore less adoption of Google properties. Seems self-correcting to me: user votes, user wins. Why do we need saving here?”


JSON: JavaScript Object Notation, the most efficient and concise way to transmit data from a web server to JavaScript code running on a browser. Unaltered adsense: Google’s requirement that if you use their API to perform search queries, you must show the sponsored links that come back with those results, unaltered.


Why is he bringing up these subjects? Dave recently misinterpreted the goal of JSON, thinking that it was an interop technology — reinventing the wheel that he himself invented in the form of XML/RPC and its successor SOAP. His readers corrected him, and he accepted the correction and moved on gracefully.

“Who, then, is Dave’s competitor? Is he caught, like Microsoft, competing against his own success? RSS won, and Dave did it. I’ll wait while nobody argues with this. Good. Now we live in an RSS world. What to do next? I say it’s gestures, but I don’t care what you think about that. Let’s say I’m right, that the world will move from inference to direct testimony, from links to gesture feeds, from push to accept, from pressure to permission. In that world, do we need protection? Or does Dave need to reinvent himself in this brave new world he launched?”


Nobody argues with this: Winer’s frequent insistence on getting credit for RSS leads many people to tease him by denying or downplaying his role, which makes him angry. Gestures: An esoteric astronaut architecture.

“It’s tough to teach an old dog new tricks, the saying goes. But Dave is not your average bear. He’s tough, cunning, honest, and vulnerable. There’s an opening to ignore what I’m saying as personal, but honestly all politics are local, and deeply personal. I’m not counting Dave, or Jonathan, out. But they need to face the music. They’ve outgrown the jobs they invented for themselves, and it’s time to grow again.”


Bear: Affectionate term for a husky, large man with a lot of body hair.


Outgrown the jobs: There are, it seems, two main points in Gillmor’s post. First, Jonathan Schwartz is wrong about rich clients. It appears that Schwartz was just being clever and using the term rich client to mean Web Browser + JavaScript plus a few other cutesy toys, nothing to be alarmed about, and Gillmor and Schwartz don’t seem to disagree at all. So that point evaporates. The other point is about why Dave Winer is changing the subject (what, are we only allowed to have one subject on the entire internet?), and it turns out that was just a simple misunderstanding, which is common on blogs: bloggers will frequently post something before fully researching a subject. Gillmor’s entire thesis about old dogs and new tricks thus completely collapses in a puff of smoke. And it only took me 3 hours to figure it out. Luckily there’s one more unrelated paragraph:

“Oh, and Microsoft, you guys better step up right now and cut this RSS patent cancer out before we do it without anesthesia.”


Before we do it: Microsoft recently applied for a patent on RSS, a disgraceful attempt to patent things that other people invented long before Microsoft was paying any attention. Gillmor’s saber rattling is endearing but completely impotent.

Have you been wondering about Distributed Version Control? It has been a huge productivity boon for us, so I wrote Hg Init, a Mercurial tutorial—check it out!

Want to know more?

You’re reading Joel on Software, stuffed with years and years of completely raving mad articles about software development, managing software teams, designing user interfaces, running successful software companies, and rubber duckies.

About the author.

I’m Joel Spolsky, co-founder of Fog Creek Software, a New York company that proves that you can treat programmers well and still be highly profitable. Programmers get private offices, free lunch, and work 40 hours a week. Customers only pay for software if they’re delighted. We make Trello, insanely simple project management, FogBugz, an enlightened bug tracker designed to help great teams develop brilliant software, and Kiln, which simplifies source control. I’m also the co-founder and CEO of Stack Exchange. More about me.

© 2000-2014 Joel Spolsky