Dirty Earnie

Pretty smug lookin' bitch/Eichmann, ain't she?

The article is a good peak at how the mainstream little Eichmanns think.

Especially the recorded stuff. Inhuman volume & precision. Overwhelming, stunting virtuosity. There ought to be special unplugged colonies that ban all such electronic shit. The kids would grow up to be much more extroverted and healthy.

Maybe my problem at Conley/EMC was that I wasn't prepared to deal with the kinds of people and personalities that you find operating situated up at a fairly high level in the cash nexus. Sharks, weasels, weaselly sharks, sharky weasels.

It's weird, Gary Gagnon, who I worked with at MITRE, once talked to me about communication protocols and mentioned these two, which I had never heard of at that time. Then a couple years later I was working on a project in South Carolina that include working with RS-232 and then a couple years after that I was living in SCSI city, so to speak, in Maryland, and that continued right up until 2000 when I was fired from EMC. I didn't really work much down in the guts of the driver at the SCSI interface level while at Conley/EMC, but that was what was used to communicate between the UNIX workstations and Symmetrix RAID boxes. Last I heard, Gary had moved, along with his family, down to DC to be part of the MITRE Corp. offices located there. He was, on balance, a miserable, self-important, officious, jumped-up little French-Canadian hard-on and I don't give a rat's ass about him.

You heard me.

But what friend will come to my aid, who will protect me, from the nightmare horror of James Taylor songs?

This song perfectly epitomizes Hannah Arendt's famous "banal face of evil". This is the song that's gently wafting through the head of the SS officer as he drops the Zyklon-B canisters into the gas chambers at Auschwitz.

At one point in this song Taylor gently whispers to the listener the advice to not let 'them' take your soul. Well I got news for you and it ain't good. If you're listening to and enjoying this song, you ain't got no fuckin' soul. Capiche?

Boston seems to be largely populated by little Eichmanns. Either accomplished ones slotted neatly into their comfy little niches or tons of little Eichmanns in training, i.e., the ubiquitous students.

Happy Little Eichmanns would be a good name for a blog about living in Boston.

EMC lost a lot of very good kernel programming talent because of this project. This Oliveira guy was a real ball-busting tyrant. Last I heard they'd shelved the project. Must gall EMC to have to use another company's (Veritas) Volume Manage for managing the volumes on their RAID Boxes. Fred was my boss when I got fired. And he was a real miserable bastard, as well (of course). I didn't really have that much to do with him. As soon as we moved into the 11 Cambridge Center quarters he told me that he and Helen Raizen and Mark Kaufman were all fairly close, having worked together at another company. I think it was Kaufman (who was the big boss) who brought Oliveira and Raizen on board at Conley before the EMC buyout. Anyway, the implication was clear. I had fucked with one of his allies, i.e., Helen, and I could expect a united front of opposition and antagonism from all 3 of them. So why bother trying to work in such an environment? Just working for the dour and saturnine Kaufman himself was enough to make me lose the will to live. Fuck it.

Anyway, all of the Volume Manager churn and drama happened after I'd left. My bud Testardi told me a little about it. He was one of the ace programmers that Oliveira chewed up and spat out, figuratively speaking. Despicable guy, Oliveira, in my opinion. Real low life.

I used to work with Dave at Mitre. He was a few years younger than me. He managed to get promoted to vice president of technical development of QStar, a permanenty struggling little start-up. He it was that got me the job.

QStar had a special room that it used for demos, the walls and ceiling were black. The carpet was a light grey, there was a long conference table in the middle surrounded by leather-like upholstered swivel chairs. All around and against the walls of the room were various makes and models of UNIX workstations. They called it he Bat Cave.

At one point we needed to install Novell Netware on one of the workstations in that room and the installation process was so lengthy and hairy that you needed to have attended an official course given by Novell to get through it. I was willing to wing it, in fact I think I had but I'd gotten screwed up after only a few steps into it. One of the QA guys, Andy Feldman had taken the course and agreed to stay a little late and install it. We got some racks of beers and Dave and I, and I forget who else, proceeded to get drunk. Dave had a drinking problem, by the way. He also had an opiate addiction problem. But anyway, it was fun shooting the shit and drinking and keeping Andy company while he did the job. Andy seemed to be enjoying it all, even though he had to keep his head clear and couldn't drink. I remember at one point telling that old joke about "2 wrongs don't make a right but 3 lefts do" and "668, neighbor of the beast".

Dave ended up getting fired not too long after. But just about everybody ended up getting fired from that place. It was positively medieval, in my view.

The Bat Cave was in the building when QStar was located in Rockville. It was a 5 story mirror-glass building and QStar was located in the top floor. The owner, a crazy young Cherokee named Brian Swafford, had a thing for shiny things. That's one of the reasons he got into optical disk storage, because they're such bright and shiny things.

Anyway, when finances got even tighter Swafford and his family decided to move the company to cheaper quarters down in Ft. Walton Beach, FL. Andy was one of the ones who followed the company down there, as did I. He got sick of it after a few months, though, and took a job back in Maryland with a tape storage company almost completely owned and staffed by jews, believe it or not. He was a good kid. I hope he's doin' alright.

Summers was a real jerk. Not much of a friend at all, which I'm sure he'd be more than happy to acknowledge. He's probably died of emphysema by now. He was quite a heavy smoker.

Of course I have no idea what I'm talking about.

My compiler course teacher, Bob Russell, used to harp on about how all this modularity and data type stuff had already been done by the Swedes back in '67 with Simula. He was a pretty good prof. Very straightforward, decent guy.

Crappy Cocksuckin' Cunts & Intelligence. I hear they added another C to the series now. Don't give a damn. Wish t'fuck I'd never heard of the miserable place.

He was my Masters project advisor, Jim Weiner's, PhD advisor. They had a falling out a few years before he died. Maybe they ultimately patched it up and had a tearful, death-bed reconciliation, like in a Hollywood movie. Who cares. Goguen's big claim to fame, as far as I knew, was his research on axiomatic characterizations of abstract data types. But I guess he mined that topic past it's relevance. I read a short informal paper by him that was posted on-line, something about throwing lotus blossoms down the divide, about how his approach had fallen out of favor. He also mentioned the documentary, Fast, Cheap and Out of Control. I watched it some years later and I don't think he got the message of the film, it was all about man's arrogance with respect to nature, either using it in some twisted way or trying to outdo it with bizarro robotics.

Weiner was kind of a miserable little prick too, on balance. I'll have to blog some more about him some other time. One thing I recall was him talking loudly to one of his UNH faculty colleagues, Gene Freuder, about Emma Goldman. I had no idea who she was at the time and only really learned about her in the past 3 years. I'm not really a fan. I tried reading her 2 volume autobiography but gave up. Too many trivial details. Kind of like this blog. Ha ha.

Anyway, here's Weiner's UNH stuff. You look, I can't bear to look at the little asshole.

I rather fancy it does. There might be some lingering odors of shit, considering how much of it I've had to eat in my life -- shit, crow, humble pie. And the fact that it's standard fare means nothing to me. The fact the I've been strapped to a chair alongside millions of others at a huge table set for an enormous banquet of shit doesn't make it taste any sweeter.

I fought to be allowed to own Powermt, the PowerPath admin tool. Then I royally screwed it up. It had been written in C++, but the kernel programmers looked down their nose at such fancy pretensions. The original programmer, Khang Can, had choosen C++ because he wanted to learn it. It was plenty good enough. He defined a data type that would keep the latest info read from the PowerPath driver about the state of all the paths for a given disc until it was really necessary to do another read from the driver. Oh, and another reason to get rid of C++ was that Ric Calvillo was a tricky little cunt and he wanted to needle Khang by throwing his design out like it was total trash. Ha ha, pretty funny, huh? Especially as it was Khang's whole idea to do PowerPath (called SafePath, originally) to begin with and that this ultimately resulted in Calvillo making a cool $15 million on the sale of his company (Conley) to EMC. Nice way for Calvillo to show his gratitude, eh? But that was his way to overcome feelings of over dependence on any one person, I guess, and that resulted in him doing his best to discount or discredit that person. Whatever floats your boat. Ric had a lot of peculiar ideas.

Anyway, I threw out all the C++ code, got rid of the persistent data state (global variables are a no-no in functional/logic programming, i.e., Lisp/Prolog) and rewrote the code in a functional programming style. This mostly meant that I had one function that applied another given function to a collection of objects ('apply' in Lisp), usually either to a collection of discs or disc paths. It also meant that I read the info from the driver a huge number times more than was strictly necessary, which when there were a lot of discs and a fair number of paths could bog down the status display of the disc paths. Things got worse from there as we moved from the Sidney St. location to the 11 Cambridge Center building.

In my defense I did realize that the design I choose was inefficient and that I might be called on to speed it up, but I only waited for that to happen instead of anticipating and fixing it as soon as I was aware of it. I had one smart-ass Pakistani QA guy, Amir Shariff, who was needling me that the status display wasn't sensitive enough to path status changes, and this zillion-of-reads approach was my response. Kind of childish. One of the new kernel programmers noticed it and put in his own time to come up with a persistent-status-maintaining strategy, much the same as Khang's, but in plain C, and then sprung it on the assembled company full blown. Dan Murphy I think his name was. A real prick. He was assigned to some weird German OS that nobody had any familiarity with, maybe Siemens. He had worked at OFS in the same building that the PowerPath people had moved into, 11 Cambridge Center. A real, tight ass, dried up, pinched faced, miserable little prick. He had been responsible for managing mutual time agreement between OFS hosts on a LAN, I believe. Who gives a shit. Turd.

At least I don't have to work with that shit, or those shitheads, anymore. I don't give a fuck how much it pays. I'd rather panhandle and sleep under bush down by the railroad tracks that live in that kind of an environment.

Oh, and my throwing out the C++ code had the added effect of pissing off Khang, who has to have been one of the nicest guys I've ever worked with in the high-tech business, i.e., he was pissed at ME. He wasn't one to hold a grudge, but still... So it was a big loser deal for me all the way around. I should have bailed out of that place when they sold it to EMC. But I was beyond, or below, clueless. Utterly hopeless and beaten down.

Also, another odious creep I had to work with, called ... Oh, who cares. Life's too short to work with jerks like him.

Rest secure that enormous commercial, financial, governmental transacting and recording entities will continue to operate smoothly and without a hitch in the event that some sub-sub-sys-admin gofer doofus trips and accidentally pulls out an adapter cable.

This was EMC's best-selling software product at the time I was working there. Relatively small-ticket item, but a big seller.

Consisted of a driver that lived below the disc driver in UNIX workstations that were attached to Symmetrix RAID boxes. In the applications layer there was a command-line admin tool that translated commands and arguments into IOCTLs to send to the driver. The admin tool also managed a small text file that contained the saved settings. And in some systems there was an additional daemon used to ensure that the driver page was never swapped out, or something like that. I never really understood.

There was an element of dishonesty to the design, I believe. The capability to chose different schemes for sending the disc data packets down the available redundant paths was sold as a possible performance enhancement. Perhaps to forestall any possible complaints that an additional driver handling the data packets might involve a significant slowdown. But since there was no real potential for true parallelism, any such performance gain was, a priori, an impossibility. In other words, regardless of the available paths the data packets had to be sent down to the disc in exactly the same order as they were received.

It's funny now that I think about it, the functionality was always spoken of in terms of writes and never reads. Surely one would get similar behavior in either case if a path became broken. Maybe most of these systems were envisioned as primarily data sinks, I don't know.

Intense guy, early forties, tall, thin and missing the upper part of his esophagus and trachea due to accidentally drinking Drano at an early age. Worked as a high-priced consultant to IBM in Austin, TX. Was discovered by Rich Krueger, the Conley sales guy. The technical lead at the time was Kevin Rodgers who had been recruited from C.L.A.M. associates in Cambridge for his IBM AIX expertise, and general technical excellence.

Apparently C.L.A.M. associates is now called Availant. And they, in turn, may have been swallowed by Visions Solutions. Who give a crap.

Anyway, talking to a guy with an incomplete trachea is painful and Kevin just wasn't up to interfacing with Perry much. Maybe there was some AIX expertise rivalry, I don't know. Perry was a pretty bullheaded guy. I was given the job of interfacing with Perry, something I didn't do that well, I didn't challenge Perry's approaches enough, I just passively accepted them. Anywho, what I remember most was eating with Perry one evening at Legal Seafood in Kendall Square. I seem to recall eating breakfast with him that same day in Central Square at the gone, but not forgotten, Golden Donut diner, as well. The deal is that in order to eat he would have to chew his food and them spit it out into a funnel in the end of a tube that went in up under his sweater and somehow into his stomach. And then he had to take a rubber bulb and place it in the tube and squeeze to push the food through. It was a little unusual, to say the least. He also had some problems with fluid accumulating in his lungs, now that I think about it. But overall his disability didn't seem to slow him down much. That guy was a real force of nature: really, really intense.

He clashed a few times with Rodgers and talked to me on the phone about it and asked how the big boss, Ric 'the prick' Calvillo felt. I was delighted to tell him that Calvillo felt the same way as Rodgers, which was a total poisonous lie on my part. You gotta get your licks in where you can, kids. Anyway, my lie had the desired effect, so that was the end of that business relationship.

I exchanged a couple e-mails with Perry after I was fired. He told me about his new house that he and his wife had built up on a hill over Austin. He and his wife had put their heart and soul into designing it according to their lifestyle and vision and Perry told me that the first time he came up to visit the finished house he told me, without the slightest trace of irony, that it was so magnificent and perfect that make him believe in the existence of god. In fact he said it with more than a trace of exultation in his voice.

All for now kids.

Oh, and apparently Krueger's bailed on Calvillo's latest start-up, Crap-cipient. For what that's worth. I have a little more I can say about Krueger that might be good enough to blog, I don't know.

Brad Butz was a field engineer located in Ireland. He did a bad no-no by contacting me, a developer, directly to help a customer at Seimens in Germany. I pounced on the opportunity to violate company protocol. I also had fun cutting my grotesque, sub-human monster (with no moral compass) of a boss, Helen Raizen, out of the loop. She got her revenge a few months later by giving me a maximally bad performance review. That didn't bother me since I no longer considered her as a valid, living entity. What bothered me was that I was stuck at a shithole of a company that would tolerate such a person, even for a day. Did I mention that this was the great and glorious EMC?

Brad Butz visited Cambridge a few months later and I had dinner with him at Legal Seafood in Kendall Square. He was American, but he had married an Irish girl and started to speak with an Irish accent. He had also been busy in the offspring department and had about 4 of them. He told me he liked to buy a lot of shoes when he visited the states because kids run through a lot of shoes and they're much cheaper in the states. A happy little techie nerd with a nice little Irish mama. I envy him.

This was code layer to provide maximum platform independence. I don't remember any details, I was pretty much lost to the world while that part of the generic PowerPath driver was being developed. And I was never a kernel programmer anyway. Just worked on the application layer admin tool and the HPUX make file and software installation package generation. And did QA and other miscellaneous support for the kernel programmers and other QA guys. Dull as dishwater, you may believe.

One in a million.

I was assigned to work with Mike because he was difficult to work with and because I too was difficult.

Getting back to his wife, she was, and might still be, a corrections officer of some sort. Says quite a bit about a person to hold such a job. Perhaps she works at the Florence supermax, it's in the same state. In my mind's eye I imagine her stopping by every day and poking Ted Kaczynski in his cell with a long, sharp stick, just for fun.

Thanks a tip of the Hatlo hat to Metallica.

To be followed by a sort of strangled screaming with your hand clutching your throat and saying in tones of ever increasing intensity and desperation: "MUST BE.... MUST BE.... CONTROOOOOOLED!!!!!!" Got that from MST3K, I believe. Was fun to do in the corridors of the various high tech places I worked at when taking breaks from writing and debugging code.

Which also reminds me of a job interview I had at Intermetrics in Cambridge back in 1983. During one part of the interview they took me over to a building that was kind of dark and there were strange jungle noises coming from the programmers in the back of the building. They were working on and 8-bit compiler. Not that that explains anything. One of the more interesting of my job experiences. They declined to hire me because I didn't seem interested enough in what they were doing. Story of my life. Got the interview via my acquaintance with a strange little, troll-like Israeli mathematician that I had worked with back at UNH called Dov Harel. He's quasi-famous now for discovering a code optimization algorithm. Knowing Dov, it was probably pretty elegant. Unfortunately he neglected to consider one of the possible cases that could arise. This was filled in, probably somewhat inelegantly, but about 4 other mathematician/computer scientists. The upshot of it all being that the optimization took too long to carry out relative to the speed gains achieved in the resulting code.

Dov was working for Microsoft and living in Haifa last I knew. I found his e-mail address. He wasn't real thrilled to hear from me. He had been let go (tenure declined) from UNH and he thought I was still there, although I'd left it only a year after he did. Easy to get stuck in a rut, I guess.

Busy shepherding us all into a glorious, ever more high-tech future. Yes. Shepherding. Future. One of the masters of progress.

I purposely skipped the meeting in which this was described and never made the effort to read the brief write-up. Instead of using the jump table, as proscribed, in the HPUX code I was responsible for I just included the linkage statically at compile time. Some months later the interface guys over at Hopkinton had reverse engineered our code, somewhat, (not sure exactly why) and discovered my faux pas. One of the leads in Cambridge, a black guy, Kevin Rodgers, told me about receiving the stuff from Hopkinton and then he proceeded to mutter "unbelievable" over and over while shaking his head. I'm a bad boy. And not much of a team player, to say the least. I mention that Kevin was black because he was very black. This was EMC, one of my brief brushes with glory. That and being a groupie for John Zerzan.

Tried to work with him as his two-man team while at QStar. Very antagonistic situation. One of QStar's salesman had tricked or coerced someone at Merrill Lynch into agreeing to buy a huge 12 inch optical platter jukebox and the associated hierarchical storage management software. The hitch was that QStar had to provide a programming interface so that the Merrill-Lynch programmers could send their financial record transactions to the optical platters. By the time the project landed in my lap the Merrill Lynch guys were anxious to pull out. I can't blame them, QStar was one slipshod outfit. And I was not much of an exception, sadly. The last straw came when there was an unexplainable discrepancy between the timing of writes to the optical drives, that is, on my testbed the writes took significantly less time to perform than on the Merrill-Lynch testbed. Turns out that they had the 'write&verify' function turned on on their optical drive and I didn't. I should have been more thoroughly versed in the hardware functionality.

One other interesting aspect of the story was the software analysis package that the ultra-anal guys at Merrill Lynch wanted us to buy and put our software through before delivering it to them. It was quite a nifty little package, but I forget the name of it. It found a lot of little flaws in code, but the most important of them included finding memory leaks and null pointers. I guess I never had the necessary dedication to be a really good professional programmer.

I believe the code to be analyzed was instrumented and then run in a GUI debugging environment type thing. Pretty nifty, really. Reminded me of a lot of ACM and IEEE articles I looked at back when I was in grad school and working at MITRE.

I include these old reminiscences because I have a nephew who thinks that writing code is a little like Futurama. I personally found it to be mind-numbingly boring drudgery. Talking about it at a high level can be exciting, I suppose, but that goes for a lot of things. In fact, I believe that's a big part of the formula for a lot of pulp fiction. Everybody wants to rise above their daily grind and soar like a bird and take it all in in a glance.