Confessions of an Internet "Shock Jock"

An anonymous reader followed up on the Windows memory-leak fraud scandal, which is worth reading before you read the perpetrator's justification. "Randall C. Kennedy comes clean about his past, his relationship to Craig Barth and how it all came tumbling down. Includes an inside look at the politics of IDG and why you can never trust an IT publication that's as obsessed with page views as InfoWorld."

No Choice at This Point

[eldavojohn]

It's a lengthy piece, you can probably skip the parts about Intel and Wall Street (although some of you may be convinced that becoming such a spinster/shill/whore/liar requires years of training). But What I found most interesting:

And implode it did. After publishing a particularly alarming set of findings – which I still stand behind while continuing to evaluate new data – the internet became engulfed in controversy. As the furor grew, and as more and more media outlets questioned just who this Craig Barth fellow really was and what made DMS tick, the house of cards came crumbling down. The persona of Craig Barth was exposed as one Randall C. Kennedy, and the entire web of half-truths and misdirection was exposed as the ruse that it was.

(Emphasis mine.) It seems like he has a reasonably technical background. What has he found that cannot be explained by SuperFetch (high memory usage) and Native Command Queuing (backlogged disk I/O queue)? Those were the two big percentage differences and apparently explainable if not desirable for the average user.

So, what next? For starters, neither the exo.performance.network or Devil Mountain Software, Inc., are going anywhere anytime soon.

Surely he must realize that open sourcing everything about exo.performance.network is the only thing he can do at this point. I mean, no one's going to trust him again if he has any way to manipulate the data/results without subject to complete inspection. The only option I see is to open source the software client and post the raw data alongside his own analysis. Without that I'm not stupid enough to trust an adoption rate quoted from this guy let alone average disk I/O queue on Windows 7. Without this kind of auditing, I'm sure those numbers will turn up to be just enough to make my eyes widen and my finger click his link. I am saddened that people will probably continue to run his client without knowing this whole story of how they were manipulated by a particularly crafty scam artist.

Meanwhile...

[bmo]

Robert Enderle still gets playtime on NPR.

Maybe it's better to just be an asshole than to be an asshole and try to hide behind a nom de plume.

--
BMO

Re:Uh...

[Anonymusing]

So why pretend that Slashdot too doesn't use shock articles sometimes to try and increase hits?

InfoWorld writes and generates news. Slashdot merely links to it and provides a discussion forum. Infoworld asks you to assume that it has credibility; Slashdot asks you to assume nothing except "this link might be interesting to technically-minded people."

You're right that Slashdot linked to the original article [slashdot.org] in this sorry mess. Infoworld claimed its conclusions were correct. Slashdot did not; it merely said, "Hey, look what Infoworld says" -- and then enabled a lengthy discussion of the merits and problems of the Infoworld article. Much of that discussion questioned Infoworld's results. Frankly, that's exactly what Slashdot is for. It actually is innocent in this.

What a piece of work

[FrankPoole]

This guy drags journalism through the mud, celebrates it like a pig rooting in his own feces, and then has the nerve to blame the media for blowing everything out of proportion and now is trying to claim his 15 minutes of fame like he's a GD Survivor villain. What a jerk. Oh, and by the way, XPNet's Windows 7 data is flat-out wrong and anyone who knows anything about Windows and memory will tell you the same thing.

Delusions of a Dickhead

[steve-o-yeah]

This guy drives me nuts, I can only presume that this post was some last-ditch effort to salvage some credibility, but in his quest to restore said trust, he continues to bloviate. He refers to himself several times as an "Internet 'shock jock'" and (my favourite) "industry’s most notorious internet “shock jock”.

Just like George Costanza couldn't pick his own nickname ("T-Bone"), YOU cannot decide who the most "notorious shock jock" is. Until I heard about your lying bullshit, I had never heard of you before.

Cram it up your ass you self-important douche.

Re:Here I'll help

[Blakey Rat]

Oh and if you haven't been following, the main cause of problems was (partially) that their tool was comparing committed bytes against physical bytes. The problem is that memory is committed against the pagefile, not physical memory... therefore it's quite possible for my computer to have:

4 GB total physical RAM
4 GB committed
3 GB available physical RAM

Via his tool, my computer would show up as memory 100% full, paging like mad. In reality, it's not paging at all. The only reasonable conclusion you can draw from that data is that my pagefile is at least 7 GB large.

Their tool was also measuring Page Ins as a stat, without realizing that memory-mapped files will trigger Page Ins even if they're already in memory. As happens with, for example, every .exe file you run, since Windows memory-maps those first thing.

The guy claims to love Windows NT, but he sure loves to slander it... oh well.