When it comes to FDA-approved medical devices, such things as disclaimers don't exist or don't mean anything. Most people don't know that if you are an executive for a company that makes FDA-approved anythings that you are 1) personally liable for criminal acts resulting from your products (standard corporate liability protection go out the window), and 2) you are guilty until proven innocent, possibly at the risk prison time. The given justification is that human life is at risk by their choices, actions and leadership. It also puts quite a damper on innovative thinking - from what I've seen, not surprisingly executive micromanagement is pretty common in such companies.
HP has long had similar terms on products (except for the former Medical Group products - now part of Philips Medical). IBM does also.
Yet Microsoft has no such language that I've seen in their licenses, T's & C's or EULAs. Not that that surprises me. There has always been a vast divide betweeen the ethics and morality of the "old line" computer firms and Microsoft.
Can you have a career in computers without typing, preferrably touch-typing, skills? Sure. Anyone can be a writer without good spelling or grammar also, but you're tying one arm behind your back by doing so. I could still use a computer if I didn't have both arms but it would be harder and put me at a disadvantage. Since most people have a choice...
Thankfully my mother "forced" me to suffer a summer of typing classes between the 8th and 9th grades - all on manual Royal typewriters. It's always made using computers so much easier. It also helped my finger strength when I started piano lessons in my 30s. I believe (w/o evidence) that good typing skills can immunize you from carpal-tunnel.
There's enormous advantage to being able to type. For me programming languages and shell commands and their standard themes pretty much "chunk" like words. This makes Unix-based OSes incredibly efficient compared to mousing everything (like Windows Sys Admin - blech!). Using Unix/Shell well goes hand-in-hand with typing.
Being able to touch type (like I am now) is even better (BTW "touch typing" means typing without looking at the keys - and some go further and define it as not looking at your typed output either but only looking at some original source you may be copying/expositioning from - all the while hitting >30-40 wpm with high accuracy). The delay between thought and action becomes nearly non-existent as typing becomes muscle memory.
And then there's being able to compose programs in a minute or so (e.g. in perl or C) by touch type using just 'cat > myprogram.pl' and having them compile/run the first time. You're truly getting hardcore when you get to that point!:-) That's generally the point when I feel I've truly mastered a language. I'm working on OCaml now.
With essentially BSD under the hood, I've never seen the motivation to put Linux on my Mac. With Fink, Qt, Mono, X11, etc., most things from the Linux world are already available, plus the nicer UI to boot. I don't have enough hours in the day as it is, and the time to admin/RTFM has been my biggest gripe about keeping Linux boxes as production machines at home.
Actually overloading happens stochastically with nearly all audio reproduction depending on the source material (most recorded & mixed poorly) and/or the inappropriate volume level given the capabilities/performance of equipment. Yes, the world would be beautiful and ideal if everything were linear (which is what you are implying is doable and obligatory) but the reality is that electronic devices are generally so profoundly nonlinear it makes more sense to simply define "how badly nonlinear".
If this makes me sound like a Tube-a-phile, let me mention I'm an analog circuit designer and my profession and personal opinion that >90% of most tube-a-philes are ignorant fools. There are a handful of exceptions. Norman Koren, though he is out of the audio hobby, knows what he's talking about. His writings should be required reading and required baseline knowledge for anyone who wants to mindlessly spout off about tubes being better over transistors. He's pro-Tube, BTW. His analysis is some of the only cogent and technically correct writing I've seen on the subject. AFAIHS, most pro-Tube audiophile magazine articles are written by people without actually knowledge of or experience in analog circuit design (building one or two tube amplifiers in your garage doesn't count) so I'm always dubious but I'm open to qualified and valid arguments, either way. This question of Tube-vs.-Transistor is usually irrelevent with bad circuit design: transistor amplifiers can be as good as the best tube amplifiers and tube amplifiers can be as bad as the worst transistor amplifiers. Device technology is not some magic bullet and claiming such only demonstrates one's stupidity and ignorance.
That said, one need only look at the rise of MP3 to see that most of the population can't hear the difference if there ever was one. This is something that the RIAA complete missed. It's also something that SuperCD and AudioDVD format promoters seem to have fatally overlooked (from an MBA sense, the market cap for such formats are far smaller than they claim or seem to believe). Most environments in which we listen to music are noisy (car, office and even home), and further most of us can't hear well enough or have the ear training to discern bad from good even with moderate quality equipment. The available "channel capacity" between our audio sources and our ears is generally far less than the 16-bits dynamic range/44.1 KHz data rate due to this ambient noise floor. Add to that the channel capacity limits between our ears and brain: I had my hearing checked when I was 19 and even then I had no significant perception over 16 KHz (which is statistically "normal" for 19 yo males). I'm in my 40s now and I've noticed my hearing getting worse since that! My iPod and its MP3 are certainly lower quality than the ideal but I get to take my entire audio collection with me anywhere in the world - nothing like sitting on the beach in Nusa Dua, Bali and feeling a particularly obscure recording from your collection would be appropriate for the moment and just playing it! That and hearing fidelity limitations tends to trump the quality argument in most cases.
Golden Ear performance is a requirement for only a tiny and limited market of audiophiles and historical archival use. The claim that overload handling differences is real and potentially relevant. Mr. Koren's analysis shows (from the pro-Tube camp) distortion is often an artifact of bad circuit design rather than necessarily a device technology issue (esp read his article on negative feedback) - bad design pervades both the Tube and Transistor sides of the audio industry. Most people won't be able to tell the difference anyway, which, from an economic-forces-driving-technology-options-and-dev elopment point of view, that's all the matters in the long run. Hence most audio is IC transistor-based, and increasingly, computer/synthesizer-based anyway.
Licensing only works for slowly changing profession fields. Licensing is a government/quasi-government (profession org) administered system which necessarily lags the rate of technological change by years if not decades. Government == slow. That's good for democracy because it tempers the evil temper of the majority but it would be the kiss of death to technological progress. That's sounds like Microsoft PR-crap but Microsoft is starting to get what's coming to them now through economic and public opinion feedback.
This time-constant mismatch is why PE certs for EEs have never taken off - the business and state-of-the-art changes too quickly. If you've ever taken the EIT or the PE you can see it instantly: the test questions cover things that are many years out of date. When I took the EIT in California in the mid-80s there were still questions about biasing vaccuum tubes, about Hoover dam-style generator maintenance and no questions at all covering my field of study: IC design. The test had a 1975 copyright date on it. I'd been doing electronics as a hobby in 1975 and vacuum tubes should not have been such an exam even then!
Power engineering in EE has certs only because power is also goverment regulated so innovation rates conveniently match certs exam update rates (try getting a solar/battery/invertor system signed off some time with the power company - it takes years though technologically it should be a 1 week thing with off-the-shelf products already available on the market).
If such certification were mandatory we'd see the field slow to something like the power industry - namely glacial. I, for one, don't want to live in a world that progresses that slowly. I'll leave the US first, despite being 4th generation, born & raised here in the US.
If certification were an entirely optional thing like the useless but optional MSCE or something, maybe. That could be done now and we'd let the market decide how useful it really is. I'd put money on such a cert being about as useful as MSCEs - essentially uncorrelated with any meaningful measure of delivered employee value.
My direct experience working in the gov't is that our greatest protection against Big Brother has been (until recently) the pervasive Byzantine and internecine nature of the government bureaucracy. In general the gov'ts fundamental inefficiency has been underestimated by the tin-hat crowd. The scary part of much of the WOT is that it seeks to eliminate that protecting inefficiency and self-destructiveness. Efficiency != Democracy. Mussolini made the trains run on time.
I started working professionally in 1980 (while I was still in undergrad), and in all this time since then I have never gotten a job by sending out resumes blindly; every job was due to contacts, accidental and intentional.
There was one summer between semesters when I was trying to get an internship anywhere. I needed money bad for school. I sent out several hundred resumes. I've keep the rejects in a box to remind me how useless that tactic is. On top of all the letters is the reject letter I got from Intel which came a month after I had already started working at Intel. How did I get that job? I talked my way into an interview over the phone, nearly by accident!
Even in the pre-Resumex/Monster.com days most HR depts operated as automated reject mills. I was following up by phone with to each company I'd sent a resume.
When I called Intel's HR department for the 2nd or 3rd time expecting the normal brush off, I said just one thing that suddenly changed it from a "don't call us, we'll call you" conversation to a "let me transfer you to one of our engineers immediately for a phone interview" conversation. The specific thing, which may or may not work today, was to mention "my befuddlement at their obstinence as my professor had sent my resume to them" (Cajones Event #1). I talked to an engineer for 15 minutes and was hired on the spot (this during the 1982 recession). As it turned out unbeknownst to me, Intel had a fast-track college internship program going on and I had just pushed the right buttons (Dumb Luck Event #1). My professor really had sent my resume to a number of companies, but ironically, Intel wasn't one of them! (Dumb Luck Event #2) I found this out when my same professor came to our very department (out of all the hundreds of departments in Intel at the time) (Dumb Luck Event #3) to seek project funding for his lab at school. After his presentation, I talked to him mentioning that the stuff his lab was doing sounded really cool and boldly asked if he ever hired undergrads into his graduate lab (Cajones Event #2). He said he did if the prospect was especially talented (Dumb Luck Event #4); I had gotten in A in his class the semester before but he didn't remember me despite my excitement over having him send out my resume (along with those of the other 30 people in class, apparently). I dropped by his office when I got back to school; he had looked my grade and he promptly offered my an RA (research assistantship) position which included full tuition (at a private university I could barely afford), a livable stipend and only optional teaching duties (which I gladly did anyway). All this while still in my junior year. After grad school I worked at a military think-tank with which my professor had research contracts.
Randomness is a big factor in life, but playing lotto is a losing game; play life by stacking the odds. Remember the 6 degrees of separation we all have to everyone else - that gives you far better and more powerful odds than playing "send off a resume" lottery. Think of life as a river and you as a kayak: you can't control much as the river is infinitely stronger than you, but you can look ahead and prepare for the randomness ahead and maybe position yourself for a really cool ride as long as you don't obsess about the tactical situation itself and learn to recognize how and when to react to what gets thrown at you.
The above true story is how life really works. I could not have planned it if I'd tried! Embrace the chaos of the Universe! There is an order to it as long as you don't look to hard for it.:-)
I certainly agree: you don't absolutely need to be an MBA to write a business plan. You don't even need an engineering degree to do good engineering either. Attitude and "showing up" counts for 80% of success.:-) But 1) the degree does give knowledge advantages and 2) people read/presume ability into the piece of paper, justified or not.
Does a marketing specialist need an MBA? It depends on the company, I guess. It depends on the person. Depends on the market they work in.
I use four computers 18+ hours a day without RAID (just normal backups) and yet in 5-odd years, I've never had a full-on disk failure of any type. All these computers are running 24/7. There must be a reason some folks have these repeated failures yet others don't. Any insights? I'd love to have a realistic and engineering valid rationale for when I really need to go to RAID.
My company has a small business server at a server farm that runs RAID on linux, and it's down far more often than all my home machines combined (Mac OSX, Suse 9 Linux, Windows XP and FreeBSD). The failure is almost always a disk issue.
The only things I can think of that might be different with me vs. the original poster: 1) I don't go out of my way to find the absolute cheapest parts - many/.ers seem to do otherwise, 2) my computer room/office is alway kept below 80 (and more usually below 75) (air con) - most failure mechanisms are heat-activated, and 3) I don't overclock or have a hellacious video card so I'm not generally pushing any component limits in general or generating a a lot of heat in the cases - hey I suck at games and I just want to use my computer to get things done. Also redundancy doesn't necessarily buy you anything if the mechanisms to create it add more error rate individually or in overhead.
Would love comments and insights...
Re:No! I use CapsLock as my "ESC" key
on
Is Caps Lock Dead?
·
· Score: 1
:-) True. Forgot about that (still have several typewriters in my garage). I'm been in the electronic world most of my life (early 70s) so for me the electronic keyboards have always trumped typewriters (even though I learned to touch-type on a manual tw).
If you don't have someone with an MBA on your board and executive team, you absolutely will not get up-to-bat, let alone to first base, with a VC, Angel, etc. It's like starting an engineered product company without having someone with an engineering degree. It can be done but nobody is going to bet their money on it.
Forensic DNA matching does not involving matching all or even a significant amount of your DNA to a challenge. Try 13 short, matching sequences of a dozen or so base pairs out of 3.2B base pairs for the mostly commonly used type of forensic DNA testing. Sure there are 6B people so that doesn't sound too bad, but are they indepenently distributed in geography or DNA space? Nobody knows - a whole company exists to try to find out - eventually). Afterall a single American White Male was the primary source of the DNA for Celera HSG work. What does that say about statistical relevance?
To put the statistics of forensic DNA matching in perspective, imagine that you wanted to "match" a suspect piece of literature to a book missing from your library (assuming that you have every copy of each author's work - no twins/clones) and that you could only match it based on the content of the book's words. Now say you decided to define a match by randomly picking one word that started with the letter 'H' of from each chapter from a book the size of War and Peace, up to 10 chapters only, and then saying that if more than 5 out of 10 words are the same, that defines a "match". Only to a certain condifidence interval you can claim to have a match and even then it's certainly never 100% sure. Basically that's how forensic DNA matching works.
You can probably be pretty sure that a given work of James Joyce is not Edgar Allen Poe, but declaring a match between some unknown challenge against a random library edition is a bit more iffy. The cases you hear about "long-shot hits" on "dead file" cases are both very long shots and (should) never (be) sufficient for conviction on DNA alone, railroading excepted. If used for a narrow range of matching tasks, DNA is superlative. Used for the wrong matching task, you might as well draw suspect names from a hat.
Everyone writing articles are on tech futurism should have read Christianson Otherwise anything of that genre from such an author should be summarily ignored as suspect and likely ignorant. The tone of this article strongly suggests he has not read it.
A key feature of all new disruptive technologies is that initially they always suck compared to what they proport or appear to replace. That's what gives them the opening to take over.
A second feature of all new disruptive technologies is that the established players dismiss them, often calling them "toys". This article itself seems to be another point of evidence though the conclusions drawn are typical "established player" self-deceptions.
Another key feature is that the price and business model is often radically different from the assumptions on price and business model held by the current technology. Usually the new technology is cheaper.
Cringley's article suggests it's WLAN VOIP will be disruptive. It seems to qualify as such.
Consider other disruptive technologies: minicomputers disrupting mainframes, microcomputers disrupting minicomputers, the internet disrupting desktop microcomputers, etc., etc.
Their solution here probably won't qualify as "due diligence" I think.:-) Let the lawyers be unleashed!
Re:No! I use CapsLock as my "ESC" key
on
Is Caps Lock Dead?
·
· Score: 1
> So, again, why not just use your Esc key as your Esc key?
An excellent question. For the youngsters here too young to remember pre-IBM PC keyboards, especially the vernerable DEC VT52 & VT100 and their predecessors on the real "original" PCs: the current PC caps lock location is where previously *everyone* had the control key. That was one of the peeves many of us had about the original IBM PC's keyboard.
Believe me, EMACS was so much easier on the left-hand ergonomics back then - just shift the pinky off the touch-type home key position and use the ring finger for ctrl-A and ctrl-S. Between this key change (with the routine unavailability of simple remapping on *other people's* IBM PC computers) and the more universal availability on Unix of vi, were what triggered my conversion from primarily EMACS (in college) to primarily vi (in the work world). ESC was always a problem with EMACS ergonomically also; statistically less common with vi.
Sorry, No. When I said "school", I meant college, undergrad and graduate. High school is about as distant as kindergarten from my point of view. A few graduate degrees will do that. Perhaps you see the past differently.
My biology-majoring school (I mean, university) friends, both in undergrad and graduate school did, in point of their own words, pick biology because of the minimal math. Those people are now runnng biotech and pharma companies today. Statistically significant? - well, a dozen data points.
Even my own sister, who is a PhD in biology and director of operations at a major biotech firm, has told me the same - she picked biology over engineering or physics, in part because of the milder math requirement. She is quite good at math I should note and probably could have easily done physics or EE.
I am thankful that bioinformatics curricula are becoming so prominent. It's a step in the right direction vis-a-vis the true complexity of the biological system. I think the profession is trending to greater mathematic sophistication but technical knowledge and expectations are always generational. I'm still involved in a number of proteomics and metabolic circuit projects so I see it first-hand. Unfortunately nothing in these will be used by the biotech industry for a good 5-10 years. Academic lab projects != commercial R&D projects.
I have read the original article, thank you. My comment was specifically with regard to my actual conversations with actual PhD's in biology running actual top-ten (earnings, reputation, your pick) biotechnology and pharma companies. This type of origianl journal article's discovery, from what I've seen personally, actually surprises folks who I would have thought/hoped knew better, but clearly don't, despite their PhDs (my sister excepted, of course). Admittedly even PhDs can become PHBs when they become VPs.
The profit motive causes as much corner-cutting in biotech as any other commercial field. Further, the opinions of those without the "gold standard" degree (in biotech that would be a PhD in biology) are discounted as much as they are in other industries, which leads to naive strategies and decisions despite the availability of other voices (sounds like Iraq). Similar things happen elsewhere: try getting promoted at Agilent above group manager without an EE degree - it will never happen! It causes the same blindspot problem.
Irrational enthusiasm expressed by too many biotech execs (I used to be in the business - my sister and brother-in-law are "wheels" in the business) is concerning.
This article is about is genomics knowledge which is one of the best understood areas of biotechnology and molecular biology, yet it's always bugged me that PhDs in biology would simply dismiss what didn't fit into their neat little model as "junk DNA". That "junk DNA" was conserved gave serious doubts about it being junk. That it has to be a "control system" component has pretty obvious.
Until recently though, math and systems theory have not been strengths of biologists in general - when I was in school, biology was what people took to be able to do science without a lot of math. Ask a biologist about Laplace, Linvill or Liapunov and you'll get a blank stare - which is truly scary if they're mucking around with living feedback systems being spread into the broader environment. There's still a generation that probably needs to be purged before the profession can be deemed "systems theory aware".
What's scarier: the whole knowledge-base of proteomics and enzyme/metabolic circuitry is far more primitive that genomics, yet this area represents far more of the biology activity in cells than genomics. Which makes plunging head-long into rolling out things like Monsanto safflower extremely dubious and dangerous.
That said, I'd be the last to advocate ceasing this type of genetic research and technology development - only it is different from most every potentially dangerous technology humanity has developed, so considerable caution and process safe-guards are needed.
I am disappointedly agreeing: he's misquoted or clueless - if anything HW is the only refuge for profit. And I really like Java too.:-(
That Sun doesn't have the executive awareness of this fundamental advantage that hardware gives bodes poorly for the company. The lynchpin of his whole subscription argument is that Sun (magically) holds some type of de facto monopoly that prevents substitution with Open Source or cheap hardware. Flaw: there are no barriers to exit for customers and few or no barriers to entry for competitive solutions. It also assumes that all markets require the same thing. That might be convenient for a COO's point of view - since it's easy to plan production that way. Tragically it fundamentally ignores customer and market reality that no two customers ever need quite the same thing from you and your product. THAT is really the only leverage point to create profit in the first place, i.e. it's a very good thing. In general, suscription business plans tend to come from companies dominated by Spreadsheet Jockeys. In general: Spreadsheet Jockeys == bad business persons. They see a cash flow and immediately have orgasms without considering whether the cash flow exists or is sustainable. Then bet company. $$$!
I would agree that certain types of hardware, especially computer hardware will be/is commodified. This is not all types being commodified, even computer harware or software.
The subtlety that he seems to miss is that profits in high tech come from the d/dt term of innovation, not the constant or linear terms: that is, profit is dynamically defined by market and product innovation momentum. He's talking like someone who only sees the static, linear terms. Again, I think that it's his COO background - that part of a company thrives and strives on conformity and uniformity as a personality trait, so it's creating a massive blindspot because he probably has no sales, marketing or R&D experience worth drawing on, those groups collectively tend to embrace those d/dt terms. This all goes back to the classic technology adoption curve: either you surf the leading edge through innovation/invention or you will be always treading water, risking capsize on the trailing, cost-cutting edge.
The classic example is what we had at HP (now the Agilent side but applies to some HP products also): HP was always pretty open with information about test instruments. You could almost always "roll-your-own" using the same off-the-shelf component instruments, as either as an end-user, reseller or OEM-user. But there was advantage to getting the whole thing at once (knowledge embedded in a turn-key solution was worth something). There are also advantages you can design in when you "own" the hardware that OTS folks won't have access to. You can do that somewhat with SW but it's much, much harder. This is analogous to a mixed-license open-source product strategy. The difference is that you capture profit or not on the software without endangering your Porter position, but you always sell hardware either way, but with a lower aggregate GS&A. This is pretty much what IBM is doing succesfully.
With Sun that advantage was with the SPARC, multiprocessing, etc. Moving hardware is always an advantage over moving software. But it appears, like some many USian companies, just like a spoiled children rant, "It's just tooooo hard. I don't wanna do it anymore". He seems to lack imagination but I think the entire Fortune 1000 does also: most Fortune 1000 companies today will be gone in 20 years!
Slashdot Mirror
Sigh, an Agilent box I'm guessing. :-(
It's sad how they've turned into an embedded Windows house.
When it comes to FDA-approved medical devices, such things as disclaimers don't exist or don't mean anything. Most people don't know that if you are an executive for a company that makes FDA-approved anythings that you are 1) personally liable for criminal acts resulting from your products (standard corporate liability protection go out the window), and 2) you are guilty until proven innocent, possibly at the risk prison time. The given justification is that human life is at risk by their choices, actions and leadership. It also puts quite a damper on innovative thinking - from what I've seen, not surprisingly executive micromanagement is pretty common in such companies.
Yet Microsoft has no such language that I've seen in their licenses, T's & C's or EULAs. Not that that surprises me. There has always been a vast divide betweeen the ethics and morality of the "old line" computer firms and Microsoft.
Thankfully my mother "forced" me to suffer a summer of typing classes between the 8th and 9th grades - all on manual Royal typewriters. It's always made using computers so much easier. It also helped my finger strength when I started piano lessons in my 30s. I believe (w/o evidence) that good typing skills can immunize you from carpal-tunnel.
There's enormous advantage to being able to type. For me programming languages and shell commands and their standard themes pretty much "chunk" like words. This makes Unix-based OSes incredibly efficient compared to mousing everything (like Windows Sys Admin - blech!). Using Unix/Shell well goes hand-in-hand with typing.
Being able to touch type (like I am now) is even better (BTW "touch typing" means typing without looking at the keys - and some go further and define it as not looking at your typed output either but only looking at some original source you may be copying/expositioning from - all the while hitting >30-40 wpm with high accuracy). The delay between thought and action becomes nearly non-existent as typing becomes muscle memory.
And then there's being able to compose programs in a minute or so (e.g. in perl or C) by touch type using just 'cat > myprogram.pl' and having them compile/run the first time. You're truly getting hardcore when you get to that point! :-) That's generally the point when I feel I've truly mastered a language. I'm working on OCaml now.
JG
With essentially BSD under the hood, I've never seen the motivation to put Linux on my Mac. With Fink, Qt, Mono, X11, etc., most things from the Linux world are already available, plus the nicer UI to boot. I don't have enough hours in the day as it is, and the time to admin/RTFM has been my biggest gripe about keeping Linux boxes as production machines at home.
I assume Microsoft is counting installations the way they always have: 1 PC = 1 Windows Install
If this makes me sound like a Tube-a-phile, let me mention I'm an analog circuit designer and my profession and personal opinion that >90% of most tube-a-philes are ignorant fools. There are a handful of exceptions. Norman Koren, though he is out of the audio hobby, knows what he's talking about. His writings should be required reading and required baseline knowledge for anyone who wants to mindlessly spout off about tubes being better over transistors. He's pro-Tube, BTW. His analysis is some of the only cogent and technically correct writing I've seen on the subject. AFAIHS, most pro-Tube audiophile magazine articles are written by people without actually knowledge of or experience in analog circuit design (building one or two tube amplifiers in your garage doesn't count) so I'm always dubious but I'm open to qualified and valid arguments, either way. This question of Tube-vs.-Transistor is usually irrelevent with bad circuit design: transistor amplifiers can be as good as the best tube amplifiers and tube amplifiers can be as bad as the worst transistor amplifiers. Device technology is not some magic bullet and claiming such only demonstrates one's stupidity and ignorance.
That said, one need only look at the rise of MP3 to see that most of the population can't hear the difference if there ever was one. This is something that the RIAA complete missed. It's also something that SuperCD and AudioDVD format promoters seem to have fatally overlooked (from an MBA sense, the market cap for such formats are far smaller than they claim or seem to believe). Most environments in which we listen to music are noisy (car, office and even home), and further most of us can't hear well enough or have the ear training to discern bad from good even with moderate quality equipment. The available "channel capacity" between our audio sources and our ears is generally far less than the 16-bits dynamic range/44.1 KHz data rate due to this ambient noise floor. Add to that the channel capacity limits between our ears and brain: I had my hearing checked when I was 19 and even then I had no significant perception over 16 KHz (which is statistically "normal" for 19 yo males). I'm in my 40s now and I've noticed my hearing getting worse since that! My iPod and its MP3 are certainly lower quality than the ideal but I get to take my entire audio collection with me anywhere in the world - nothing like sitting on the beach in Nusa Dua, Bali and feeling a particularly obscure recording from your collection would be appropriate for the moment and just playing it! That and hearing fidelity limitations tends to trump the quality argument in most cases.
Golden Ear performance is a requirement for only a tiny and limited market of audiophiles and historical archival use. The claim that overload handling differences is real and potentially relevant. Mr. Koren's analysis shows (from the pro-Tube camp) distortion is often an artifact of bad circuit design rather than necessarily a device technology issue (esp read his article on negative feedback) - bad design pervades both the Tube and Transistor sides of the audio industry. Most people won't be able to tell the difference anyway, which, from an economic-forces-driving-technology-options-and-dev elopment point of view, that's all the matters in the long run. Hence most audio is IC transistor-based, and increasingly, computer/synthesizer-based anyway.
JG
History doesn't repeat itself but it does rhyme - Mark Twain
So does this mean BHO really stands for "Butt-Hole Objects"? (Apologies to Mac 7100 owners and the late Carl Sagan)
This time-constant mismatch is why PE certs for EEs have never taken off - the business and state-of-the-art changes too quickly. If you've ever taken the EIT or the PE you can see it instantly: the test questions cover things that are many years out of date. When I took the EIT in California in the mid-80s there were still questions about biasing vaccuum tubes, about Hoover dam-style generator maintenance and no questions at all covering my field of study: IC design. The test had a 1975 copyright date on it. I'd been doing electronics as a hobby in 1975 and vacuum tubes should not have been such an exam even then!
Power engineering in EE has certs only because power is also goverment regulated so innovation rates conveniently match certs exam update rates (try getting a solar/battery/invertor system signed off some time with the power company - it takes years though technologically it should be a 1 week thing with off-the-shelf products already available on the market).
If such certification were mandatory we'd see the field slow to something like the power industry - namely glacial. I, for one, don't want to live in a world that progresses that slowly. I'll leave the US first, despite being 4th generation, born & raised here in the US.
If certification were an entirely optional thing like the useless but optional MSCE or something, maybe. That could be done now and we'd let the market decide how useful it really is. I'd put money on such a cert being about as useful as MSCEs - essentially uncorrelated with any meaningful measure of delivered employee value.
My direct experience working in the gov't is that our greatest protection against Big Brother has been (until recently) the pervasive Byzantine and internecine nature of the government bureaucracy. In general the gov'ts fundamental inefficiency has been underestimated by the tin-hat crowd. The scary part of much of the WOT is that it seeks to eliminate that protecting inefficiency and self-destructiveness. Efficiency != Democracy. Mussolini made the trains run on time.
I started working professionally in 1980 (while I was still in undergrad), and in all this time since then I have never gotten a job by sending out resumes blindly; every job was due to contacts, accidental and intentional.
There was one summer between semesters when I was trying to get an internship anywhere. I needed money bad for school. I sent out several hundred resumes. I've keep the rejects in a box to remind me how useless that tactic is. On top of all the letters is the reject letter I got from Intel which came a month after I had already started working at Intel. How did I get that job? I talked my way into an interview over the phone, nearly by accident! Even in the pre-Resumex/Monster.com days most HR depts operated as automated reject mills. I was following up by phone with to each company I'd sent a resume.
When I called Intel's HR department for the 2nd or 3rd time expecting the normal brush off, I said just one thing that suddenly changed it from a "don't call us, we'll call you" conversation to a "let me transfer you to one of our engineers immediately for a phone interview" conversation. The specific thing, which may or may not work today, was to mention "my befuddlement at their obstinence as my professor had sent my resume to them" (Cajones Event #1). I talked to an engineer for 15 minutes and was hired on the spot (this during the 1982 recession). As it turned out unbeknownst to me, Intel had a fast-track college internship program going on and I had just pushed the right buttons (Dumb Luck Event #1). My professor really had sent my resume to a number of companies, but ironically, Intel wasn't one of them! (Dumb Luck Event #2) I found this out when my same professor came to our very department (out of all the hundreds of departments in Intel at the time) (Dumb Luck Event #3) to seek project funding for his lab at school. After his presentation, I talked to him mentioning that the stuff his lab was doing sounded really cool and boldly asked if he ever hired undergrads into his graduate lab (Cajones Event #2). He said he did if the prospect was especially talented (Dumb Luck Event #4); I had gotten in A in his class the semester before but he didn't remember me despite my excitement over having him send out my resume (along with those of the other 30 people in class, apparently). I dropped by his office when I got back to school; he had looked my grade and he promptly offered my an RA (research assistantship) position which included full tuition (at a private university I could barely afford), a livable stipend and only optional teaching duties (which I gladly did anyway). All this while still in my junior year. After grad school I worked at a military think-tank with which my professor had research contracts.
Randomness is a big factor in life, but playing lotto is a losing game; play life by stacking the odds. Remember the 6 degrees of separation we all have to everyone else - that gives you far better and more powerful odds than playing "send off a resume" lottery. Think of life as a river and you as a kayak: you can't control much as the river is infinitely stronger than you, but you can look ahead and prepare for the randomness ahead and maybe position yourself for a really cool ride as long as you don't obsess about the tactical situation itself and learn to recognize how and when to react to what gets thrown at you.
The above true story is how life really works. I could not have planned it if I'd tried! Embrace the chaos of the Universe! There is an order to it as long as you don't look to hard for it. :-)
Does a marketing specialist need an MBA? It depends on the company, I guess. It depends on the person. Depends on the market they work in.
I use four computers 18+ hours a day without RAID (just normal backups) and yet in 5-odd years, I've never had a full-on disk failure of any type. All these computers are running 24/7. There must be a reason some folks have these repeated failures yet others don't. Any insights? I'd love to have a realistic and engineering valid rationale for when I really need to go to RAID.
My company has a small business server at a server farm that runs RAID on linux, and it's down far more often than all my home machines combined (Mac OSX, Suse 9 Linux, Windows XP and FreeBSD). The failure is almost always a disk issue.
The only things I can think of that might be different with me vs. the original poster: 1) I don't go out of my way to find the absolute cheapest parts - many /.ers seem to do otherwise, 2) my computer room/office is alway kept below 80 (and more usually below 75) (air con) - most failure mechanisms are heat-activated, and 3) I don't overclock or have a hellacious video card so I'm not generally pushing any component limits in general or generating a a lot of heat in the cases - hey I suck at games and I just want to use my computer to get things done. Also redundancy doesn't necessarily buy you anything if the mechanisms to create it add more error rate individually or in overhead.
Would love comments and insights...
:-) True. Forgot about that (still have several typewriters in my garage). I'm been in the electronic world most of my life (early 70s) so for me the electronic keyboards have always trumped typewriters (even though I learned to touch-type on a manual tw).
This sounds about as vague as the dot-com boom. I don't believe it.
If you don't have someone with an MBA on your board and executive team, you absolutely will not get up-to-bat, let alone to first base, with a VC, Angel, etc. It's like starting an engineered product company without having someone with an engineering degree. It can be done but nobody is going to bet their money on it.
Forensic DNA matching does not involving matching all or even a significant amount of your DNA to a challenge. Try 13 short, matching sequences of a dozen or so base pairs out of 3.2B base pairs for the mostly commonly used type of forensic DNA testing. Sure there are 6B people so that doesn't sound too bad, but are they indepenently distributed in geography or DNA space? Nobody knows - a whole company exists to try to find out - eventually). Afterall a single American White Male was the primary source of the DNA for Celera HSG work. What does that say about statistical relevance?
To put the statistics of forensic DNA matching in perspective, imagine that you wanted to "match" a suspect piece of literature to a book missing from your library (assuming that you have every copy of each author's work - no twins/clones) and that you could only match it based on the content of the book's words. Now say you decided to define a match by randomly picking one word that started with the letter 'H' of from each chapter from a book the size of War and Peace, up to 10 chapters only, and then saying that if more than 5 out of 10 words are the same, that defines a "match". Only to a certain condifidence interval you can claim to have a match and even then it's certainly never 100% sure. Basically that's how forensic DNA matching works.
You can probably be pretty sure that a given work of James Joyce is not Edgar Allen Poe, but declaring a match between some unknown challenge against a random library edition is a bit more iffy. The cases you hear about "long-shot hits" on "dead file" cases are both very long shots and (should) never (be) sufficient for conviction on DNA alone, railroading excepted. If used for a narrow range of matching tasks, DNA is superlative. Used for the wrong matching task, you might as well draw suspect names from a hat.
A key feature of all new disruptive technologies is that initially they always suck compared to what they proport or appear to replace. That's what gives them the opening to take over.
A second feature of all new disruptive technologies is that the established players dismiss them, often calling them "toys". This article itself seems to be another point of evidence though the conclusions drawn are typical "established player" self-deceptions.
Another key feature is that the price and business model is often radically different from the assumptions on price and business model held by the current technology. Usually the new technology is cheaper.
Cringley's article suggests it's WLAN VOIP will be disruptive. It seems to qualify as such.
Consider other disruptive technologies: minicomputers disrupting mainframes, microcomputers disrupting minicomputers, the internet disrupting desktop microcomputers, etc., etc.
Their solution here probably won't qualify as "due diligence" I think. :-) Let the lawyers be unleashed!
An excellent question. For the youngsters here too young to remember pre-IBM PC keyboards, especially the vernerable DEC VT52 & VT100 and their predecessors on the real "original" PCs: the current PC caps lock location is where previously *everyone* had the control key. That was one of the peeves many of us had about the original IBM PC's keyboard.
Believe me, EMACS was so much easier on the left-hand ergonomics back then - just shift the pinky off the touch-type home key position and use the ring finger for ctrl-A and ctrl-S. Between this key change (with the routine unavailability of simple remapping on *other people's* IBM PC computers) and the more universal availability on Unix of vi, were what triggered my conversion from primarily EMACS (in college) to primarily vi (in the work world). ESC was always a problem with EMACS ergonomically also; statistically less common with vi.
My biology-majoring school (I mean, university) friends, both in undergrad and graduate school did, in point of their own words, pick biology because of the minimal math. Those people are now runnng biotech and pharma companies today. Statistically significant? - well, a dozen data points. Even my own sister, who is a PhD in biology and director of operations at a major biotech firm, has told me the same - she picked biology over engineering or physics, in part because of the milder math requirement. She is quite good at math I should note and probably could have easily done physics or EE.
I am thankful that bioinformatics curricula are becoming so prominent. It's a step in the right direction vis-a-vis the true complexity of the biological system. I think the profession is trending to greater mathematic sophistication but technical knowledge and expectations are always generational. I'm still involved in a number of proteomics and metabolic circuit projects so I see it first-hand. Unfortunately nothing in these will be used by the biotech industry for a good 5-10 years. Academic lab projects != commercial R&D projects.
I have read the original article, thank you. My comment was specifically with regard to my actual conversations with actual PhD's in biology running actual top-ten (earnings, reputation, your pick) biotechnology and pharma companies. This type of origianl journal article's discovery, from what I've seen personally, actually surprises folks who I would have thought/hoped knew better, but clearly don't, despite their PhDs (my sister excepted, of course). Admittedly even PhDs can become PHBs when they become VPs.
The profit motive causes as much corner-cutting in biotech as any other commercial field. Further, the opinions of those without the "gold standard" degree (in biotech that would be a PhD in biology) are discounted as much as they are in other industries, which leads to naive strategies and decisions despite the availability of other voices (sounds like Iraq). Similar things happen elsewhere: try getting promoted at Agilent above group manager without an EE degree - it will never happen! It causes the same blindspot problem.
This article is about is genomics knowledge which is one of the best understood areas of biotechnology and molecular biology, yet it's always bugged me that PhDs in biology would simply dismiss what didn't fit into their neat little model as "junk DNA". That "junk DNA" was conserved gave serious doubts about it being junk. That it has to be a "control system" component has pretty obvious.
Until recently though, math and systems theory have not been strengths of biologists in general - when I was in school, biology was what people took to be able to do science without a lot of math. Ask a biologist about Laplace, Linvill or Liapunov and you'll get a blank stare - which is truly scary if they're mucking around with living feedback systems being spread into the broader environment. There's still a generation that probably needs to be purged before the profession can be deemed "systems theory aware".
What's scarier: the whole knowledge-base of proteomics and enzyme/metabolic circuitry is far more primitive that genomics, yet this area represents far more of the biology activity in cells than genomics. Which makes plunging head-long into rolling out things like Monsanto safflower extremely dubious and dangerous.
That said, I'd be the last to advocate ceasing this type of genetic research and technology development - only it is different from most every potentially dangerous technology humanity has developed, so considerable caution and process safe-guards are needed.
That Sun doesn't have the executive awareness of this fundamental advantage that hardware gives bodes poorly for the company. The lynchpin of his whole subscription argument is that Sun (magically) holds some type of de facto monopoly that prevents substitution with Open Source or cheap hardware. Flaw: there are no barriers to exit for customers and few or no barriers to entry for competitive solutions. It also assumes that all markets require the same thing. That might be convenient for a COO's point of view - since it's easy to plan production that way. Tragically it fundamentally ignores customer and market reality that no two customers ever need quite the same thing from you and your product. THAT is really the only leverage point to create profit in the first place, i.e. it's a very good thing. In general, suscription business plans tend to come from companies dominated by Spreadsheet Jockeys. In general: Spreadsheet Jockeys == bad business persons. They see a cash flow and immediately have orgasms without considering whether the cash flow exists or is sustainable. Then bet company. $$$!
I would agree that certain types of hardware, especially computer hardware will be/is commodified. This is not all types being commodified, even computer harware or software. The subtlety that he seems to miss is that profits in high tech come from the d/dt term of innovation, not the constant or linear terms: that is, profit is dynamically defined by market and product innovation momentum. He's talking like someone who only sees the static, linear terms. Again, I think that it's his COO background - that part of a company thrives and strives on conformity and uniformity as a personality trait, so it's creating a massive blindspot because he probably has no sales, marketing or R&D experience worth drawing on, those groups collectively tend to embrace those d/dt terms. This all goes back to the classic technology adoption curve: either you surf the leading edge through innovation/invention or you will be always treading water, risking capsize on the trailing, cost-cutting edge.
The classic example is what we had at HP (now the Agilent side but applies to some HP products also): HP was always pretty open with information about test instruments. You could almost always "roll-your-own" using the same off-the-shelf component instruments, as either as an end-user, reseller or OEM-user. But there was advantage to getting the whole thing at once (knowledge embedded in a turn-key solution was worth something). There are also advantages you can design in when you "own" the hardware that OTS folks won't have access to. You can do that somewhat with SW but it's much, much harder. This is analogous to a mixed-license open-source product strategy. The difference is that you capture profit or not on the software without endangering your Porter position, but you always sell hardware either way, but with a lower aggregate GS&A. This is pretty much what IBM is doing succesfully.
With Sun that advantage was with the SPARC, multiprocessing, etc. Moving hardware is always an advantage over moving software. But it appears, like some many USian companies, just like a spoiled children rant, "It's just tooooo hard. I don't wanna do it anymore". He seems to lack imagination but I think the entire Fortune 1000 does also: most Fortune 1000 companies today will be gone in 20 years!