Friday, November 09, 2007

hype versus reality

Like every engineer, I have to admit, upfront, that I have a limited tolerance to spin-meistering and marketing terminology. I find it boring, repetitive, tiring and annoying. It's also totally predictable - after you identify the marketing speakwords, aka buzzwords, its really annoying to see them over-used again and again and again and again - you get the picture. Most engineers can relate to this. The flip-side is that a certain amount of buzzwords replication actually works! OK - while I don't pretend to understand this phenomenon, I'll buy into it, based on anecdotal evidence. But... there is a point where hype and buzzword usage crosses my personal line-in-the-sand. I guess every technocrat has there own line-in-the-sand.

That line, in my case, is where hype and buzzwords go from being hype, that really can't be verified and validated, to where it is totally bogus and unbelievable. In fact, I'll go further and state that, in some extreme cases, it can be plain dumb and/or possibly dishonest. Such is the case with Marc Hamiltons blog entry which states that the indiana project preview had seen 100,000 downloads in less than 72 hours. Why did I think that this number was just plain wrong? Because, I had an email exchange with Jesse Silver of Sun and he asked me if I could provide download numbers for the Project Indiana Developer Preview (filename in-preview.iso) that we were mirroring on and he told me to expect "big numbers". So I asked him, "what do you mean by big numbers?" and he said that they had seen over 100,000 downloads from I was curious - because, my initial reaction, was that I did'nt (personally) feel that there was this level of interest in Indiana - particularly since the marketing team had not released it under the widely expected name of Project Indiana, but instead, had chosen to rename/re-brand it, to the OpenSolaris Developer Preview - which no-one, including me, really expected. Well, after taking a look at's numbers: we had shipped 690 copies at that time (Mon Nov 5 11:52:54 PST 2007), I just did'nt see that level of interest. After a quick back-of-the-napkin calculation, I knew those numbers were just plain wrong - and I advised Jesse that I felt that those numbers were flawed. Why? Well, for the answer, take a look at the following email I sent to Marc Hamilton (3 days) later in the week after I noticed his blog entry (referred to in a post to one of the OpenSolaris mailing lists which prompted me to read it):

--------- begin Marc Hamilton email -----------
Date: Thu, 8 Nov 2007 10:22:09 -0600 (CST)
From: Al Hopper
To: Marc Hamilton
Subject: 100k download - hard to believe

Hi Marc,

I saw your recent blog[1] and the number you quoted for Project Indiana downloads (100,000) does not look reasonable to me. A "back-of-the-napkin" calculation reveals that for 100,000 downloads of 660226048 bytes per iso image, delivered over 72 hours, you'd be pushing over 2Gbits/Sec to the 'net. From a quick test of, it looks like you've got a 5Mbit/Sec cap on your connection (my best techguess).

I had an earlier email "conversation" with Jesse Silver - where he asked me for our stats[2] and quoted his download numbers. I expressed scepticism that his numbers were accurate - suggesting that they may have counted the number of download transactions from the http access logs, rather than accumulating a count of the bytes transferred per in-preview.iso transacation and dividing the result by 660226048 (the size of the iso image). As of about 1 hour ago, we've shipped 808 copies of in-preview.iso.

I would suggest that you update your blog ASAP.
Comments welcome.

[2] we are providing downloads of the in-preview.iso file

--------- end Marc Hamilton email -----------

I did'nt receive any feedback from Marc 28+ hours later - hence this blog.
Should inaccurate hype be allowed to go unchallenged? What do you think?

PS: screen capture of Marc Hamilton blog as of Fri Nov 9th 19:28 Pacific