Blog by Sumana Harihareswara, Changeset founder

31 Oct 2007, 10:30 a.m.

Had Me, Lost Me

Hi, reader. I wrote this in 2007 and it's now more than five years old. So it may be very out of date; the world, and I, have changed a lot since I wrote it! I'm keeping this up for historical archive purposes, but the me of today may 100% disagree with what I said then. I rarely edit posts after publishing them, but if I do, I usually leave a note in italics to mark the edit and the reason. If this post is particularly offensive or breaches someone's privacy, please contact me.

At a VMWare recruiting session on Monday, the Ryan North lookalike Eli Collins mentioned dtrace. I thought, "I should learn more about dtrace!" Today I happened across Bryan Cantrill's Google Tech Talk on dtrace and decided to watch it.

Within the first five minutes, Cantrill mistakenly calls Dreaming in Code a bad book and regales an audience at frickin' Google with the "is software information or a machine? both!" conundrum as though it's new. OK, fine, I'll just go read a tutorial somewhere rather than listen to this guy for an hour.

Comments

Bryan Cantrill
http://blogs.sun.com/bmc
31 Oct 2007, 16:18 p.m.

Well, first of all, I very much stand by my assertion that "Dreaming in Code" is essentially a bad book -- and if you would like to debate that, kindly name your time and place. (Perhaps we could get Google to sponsor it as a tech talk?) Secondly, your bravado around the putative "obviousness" of the software paradox is completely unjustified; indeed, I don't think any of us yet truly understands its myriad and subtle ramifications. (Talk to me in a couple of hundred years on that front.) Finally, as for not being able to stomach my presentation for an hour, perhaps you can point me to a talk of yours and I can similarly return the favor of a review based on its first ninety seconds?

Bryan Cantrill
http://blogs.sun.com/bmc
31 Oct 2007, 17:18 p.m.

Sorry, me again. Curious as to what kind of person would lay down that kind of invective based on such precious little data, I learned a little bit more about you. Imagine my surprise that you are completely non-technical! I have to say, I'm impressed: most of you hangers-on gadfly types get silent and uncomfortable when the conversation drifts deeply technical -- but not you! Those subtle cues that, in the rest of us, tell us when we should probably lay low and try to learn something for a second are refreshingly absent in you, and I for one welcome it! But now I'm even more depressed that you didn't watch the video. What would have happened, I ask myself, when I drifted into the cross-call manifestations of virtual address cache conflicts starting at 40:50? Would you have met your match? I bet not, and I wish I could see you dismiss it! Best of luck to you -- and may I one day have the privilege of presenting to you in person.

Nick Moffitt
http://groups.google.com/groups?selm=5543dq%24qjd%40engnews2.Eng.Sun.COM
31 Oct 2007, 18:10 p.m.

Yes, but have you ever kissed a girl?

Bryan Cantrill
http://blogs.sun.com/bmc
31 Oct 2007, 20:22 p.m.

Touche!<br/>

Sumana Harihareswara
http://brainwane.net
01 Nov 2007, 9:29 a.m.

Bryan Cantrill:<br/>

<br/>Thank you for your work on DTrace! I now know that it's a terrific innovation and I look forward to enjoying its fruits, and perhaps even someday using it.<br/>

<br/>A common theme, I find, in answering your objections is that something is or is not appropriate for a certain audience or situation or intent.<br/>

<br/>"Dreaming in Code" is a good and useful book for an aspiring geek or smart layperson to read, to give her a sense of big problems in software engineering. It's not a good book for an adamantium-hard-core geek like Bryan Cantrill to read, because you already know the theories and the history, and you'll only notice where analogies break down instead of where they help. I live in New York; let me know when you're in town and we'll discuss it over a beverage.<br/>

<br/>I've watched several Google tech talks to get a sense of interesting topics or tools. One hour's video won't ever make me an expert, of course, and I don't know enough to grok, say, a discussion of cross-call manifestations of virtual address cache conflicts. But I was going to watch your talk to increase my breadth of knowledge, not for depth. I am not as technical as you, but I'm technical enough to want to know about DTrace.<br/>

<br/>Your talk on DTrace was not appropriate for me, because I found your style unpleasant. I'm sure it was very informative, but I disliked that you started off your talk by slamming a good book, and then doing the rhetorical question:<br/>

<br/>"Here's the paradox. Is software information? Or is software a machine?"<br/>

<br/>Pause as the audience realizes that you're waiting for them to provide the answer.<br/>

<br/>"Both."<br/>

<br/>"Exactly! The answer is, it's both."<br/>

<br/>You said I called it obvious. No, I said it's not new. Thus your audience participation moment wasn't energizing, just an recitation of a known fact to make you get on with it.<br/>

<br/>I'm not going to spend an hour watching a presentation that promises to be both abrasive and boring, even if it's chock full of useful information. I can get information other ways. I've now read some basic FAQs and tutorials; this Jarod Jenson interview was also helpful.<br/>

<br/>I don't think any of my teaching at Berkeley got recorded, otherwise I'd offer it to you. Here's some of my stand-up comedy. I do hope you have the chance to enjoy a talk of mine someday, and that I can someday enjoy a talk of yours.