WHO keeps the servers up?
on
NetSlaves
·
· Score: 2
Funny; I used to work with one of the authors at a Silicon Alley startup. He was a complete non-techie, a "Webmaster" who produced neither HTML nor English for the site, but instead concentrated on the "overall architecture" of the site. Whatever that meant.
He's probably angry and bitter his experience there didn't make him rich. On the other hand, he didn't stay around long enough to vest any of his options. On the third hand, I never noticed him working particularly long hours.
He feels sorry for the guys who kept the servers going? That would be a couple of 23-year-old college dropouts (and one Columbia grad), who rarely showed up before 11 a.m. but often stayed past midnight, got paged at 3 a.m., and worked a lot of weekends. I'm glad the authors feel sorry for them... but at least one of them doesn't know a thing about that life.
(Since you're probably wondering: I was a lead developer, then "Manager, Content Acquisition Software," for a team of five Unix/Perl programmers and one NT/C++ guy. I didn't work very long hours, unless you counted the productive four hours every day on the train. I stayed around long enough to vest about a third of my ten-cent-a-share options. But when the company laid off most of the technical staff, I saw the writing on the wall. I found a job closer to home, and let my options expire. The company's doing okay, with twenty employees instead of the hundred and twenty it had when I was there. I don't think it'll ever IPO. Bottom line: I'm not rich, either. And, yes, when the sysadmins weren't around, I was one of the other guys who kept the servers up.)
The first public information on Unix (other than printing the troff files from the distribution tapes) was the Unix issue (July-August 1978) of the Bell System Technical Journal. Thompson & Ritchie had an article on the kernel, I think there were articles on the compiler, the shell (Mashey or Bourne?), ed, and nroff/troff.
This issue holds the distinction of being the most-stolen magazine in AT&T/Lucent/Telecordia technical libraries.
The "Bell System" was AT&T, including Western Electric (now Lucent, more or less), the "baby Bells" (now Bell Atlantic et. al.), and Bell Telephone Laboratories (now AT&T Labs, Bell Labs, and Telecordia). It was broken up when AT&T "divested" the local operating companies in 1984.
When I program in C and C++, I always worry a stray pointer will bring my program crashing down. In Perl, I can use an undefined variable, or call a function I expected to be autoloaded... but in practice, I'm far less likely to see these kinds of bugs once I've done a sane amount of testing. (Put another way: An equal amount of testing gives me more confidence in a Perl program than the "equivalent" C or C++ program.)
P.S.: I've yet to overflow a Perl string. Wish I could say that about C char arrays....-(
Previous/. posts suggest Linux is being used widely through the United Kingdom's civil service. How widely? Are some folks running servers (old legacy boxes or new applications) on Solaris, BSD, NT, or some other OS?
Check out the Index of Famous Dog names. "Welcome to Dog Central! The title says it all; this is an Index of Famous Dogs throughout history, both in reality and in fiction. This is not to be considered a complete list, nor can I account for the accuracy of all the information here (just some of it)."
Cats and other critters also available on that site.
(Back in 1982, we needed to move our systems from one building to another. We were told to lose the Star Trek themed names. We picked state names; useful since all of them have two letter abbreviations. "lznv" was the "Nevada" system in the Lincroft building (location code "lz"; dunno what the "z" stood for). It was funny, though, tracing the bus from Hawaii to New Hampshire.-)
I use (*sigh*) IE5 to show me offered cookies (my making custom changes to the Internet security zone). Any image server that wants to give me a persistent cookie goes into my Restricted Sites zone, which (among other things) are prevented from giving me cookies.
Here's my list:
*.accendo.com
*.admex.com
*.admonitor.com
*.ap-adcenter.com
*.avenuea.com
*.bfast.com
*.burstnet.com
cookies.cmpnet.com
*.doubleclick.com
*.enliven.com
*.eu-adcenter.com
*.flycast.com
*.focalink.com
ads.guardianunlimited.com.uk
*.hitbox.com
*.iadnet.com
ads.icq.com
*.imgis.com
*.inet1.com
*.isyndicate.com
*.linkforads.com
*.linkexchange.com (Microsoft!)
*.mediaplex.com
*.netgravity.net
*.ngadcenter.net
privacyproxy.nytimes.com
ngnetwork.pcworld.com
*.preferences.com
*.smartclicks.com
*.surfree.com
*.thecounter.com
*.track-star.com
*.tripod.com
*.uexpress.com
*.valueclick.com
*.webconnect.com
The hosts without the * are the scariest. ngnetwork.pcworld.com is offering me a cookie named NGADsomething. Want to bet that host, though in the pcworld.com domain, is actually the IP address of an ngadcenter host? In other words, you can block the ngadcenter domain and NGAD can still track you. Ouch.
I haven't (yet) set up host lookup to set all those hosts and domains to 127.0.0.1, but I'm thinking about it. --PSRC
"The Unofficial Guide to..." was a series of MS Windows books edited (sometimes written) for Addison Wesley by Woody Leonhard. He's since edited (and sometimes written) the "Annoyances" series of MS Windows books for O'Reilly. (Both series recommended.)
Mr. Katz suggests making (in effect) 'Net-based, short turnaround peer review a standard way of doing business; journalism for the 21st century.
I see two potentially serious problems with this (in addition to the obvious problem with not waiting to release "scoops"):
(1) In no time at all, there would be an unacceptably low ratio of qualified reviewers to material to be reviewed. (For example, Slashdot provides a large and self-moderating community, but could even we review every ZD story for accuracy?)
(2) Unless posted in some secure area, publishing drafts is still publishing, and could raise additional probability of slander lawsuits. (One would presumably publish early drafts, simultaneously with the usual internal review that checks for this kind of liability.)
In the consumer electronics business (at least in the U.S.), there are some well known "sweet spots" in the price spectrum. They're mostly multiples of $100. Hit these price points, and you'll encourage impulse buys by various kinds of consumers in various circumstances. The most well known price points are $100, $300, $500, and $700. (Think what price a VCR, or a fairly big screen TV, or a camcorder had to hit before you considered buying one.)
$700 PCs are a huge hit. WebTV seems stuck at $500, and it's anything but taking the market by storm. 3DO proved conclusively that the $500 price point is too high for game consoles.
Nintendo and Sony have found a good price point at $200 for game consoles. Sega had to hit that point.
Households with home LANS or cable modems, almost by definition, live higher on the price curve. It makes sense for Sega to offer an upgrade kit, or a higher priced (because they can get away with it) version, for that market segment. But Sega shouldn't blow the price point for the majority of their customer base.
(I worked on four different consumer telecommunications services from 1992 to 1998. I saw lots of products miss the price points, even though we knew better.)-:
Seems to me the real problem is the AOL password is stored in the clear on every client's machine. (That's the only way a trojan horse could e-mail it out, right?)
The last place to insist on the "corporate" look is Wall Street. You will show up at 8 p.m. promptly, in a nice suit, work past the point of exhaustion every day, produce lots of crap in a hurry, get paid six figures, and burn out in six months.
All so business guys (being paid seven figures) can move $$$ from point A to point B and take a percentage.
Wall Street does jump on the hottest technology (or technological fad) pretty fast. They had Solaris workstations years ago (on non-techies' desks, because that's what the applications were designed to run on; moving to NT now), used Smalltalk for a while, jumped to Java early. (Their business logic changes at whim, so they need really rapid application development.) It can be an interesting entry on your resume. Not mine, thanks.
(They also suck up 90% of the good technologists, and inflate salaries enormously, in Manhatten. When I was a manager at a Silicon Alley startup, that made recruting tough.)
India was a British colony for many years. Most Indians who come to the U.S. speak excellent English, in addition to an Indian language.
There are approximately a zillion official languages in India. The only one spoken everywhere in the country is English, so that's everyone's second language. Like Rob, I'd hate to see them treated as second-class citizens. The good news is, many get sponsored for citizenship.
The South Asians in the U.S. are affluent enough for plane fare here. Nearly 100% of them have bachelor's degrees; many of them have master's degrees; the majority of them have some industry experience.
The big hype about "Indian" programmers was how the marketing types were going to outsource development to folks living in Pakistan, India, and other places where salaries are much lower. Hasn't much happened yet.
As someone else said, breaking up Microsoft is like breaking up an anthrax spore: you get two of what you didn't want one of.
Regulating Microsoft is like the IRS taking over the Mustang Ranch and still operating it as a brothel: the customers still get screwed, just not as creatively. (I heard the Mustang Ranch is going to be operated as a real ranch for real horses.)
What's left? Bill Gates laughed at a million dollar a day fine. We don't want him to laugh at the proposed remedy; we want him scared enough to modify his behavior.
How about a nice round billion dollar (US) fine?
I'm serious. It's somewhere between a slap on the wrist and chopping a hand off. In and of itself, it doesn't prevent Microsoft from being bad in the future, but if future offenses can lead to future significant fines, it might get their attention. They might change their ways, they might agree to a settlement. I haven't heard any other proposal that, realistically could lead that way.
The C++ and Perl versions were about equally fast; the C version was about five times faster. (I'm not counting the broken Windows deque-based version.) See p. 81 for the numbers.
I hope (in my copious spare time:-) to figure out just why the C++ version is so much slower.
What John Ousterhout (creator of Tcl/Tk) does about his RSI. Summary: Windows PC running Dragon Systems NaturallySpeaking, plus a2x to make the Windows PC act as a keyboard for a Unix box running X.
Here is an article from Byte reviewing good mikes with NaturallySpeaking. (Formatting's terrible, but at least it's all on one Web page.)
Having said that, let me emphasize the "closest thing" isn't very close! In many engineering disciplines, there is a well-understood, well agreed-to body of practices that no competent professional would violate. In software "engineering," there's almost concensus that goto statements should be avoided most of the time, and that's about it.
As a result, darned near no one has a CCP, and most employers don't know what it means, let along care whether or not you have one.
Forget grad school and go out into industry unless you are into distributed systems
...
U Wisconsin Madison
When I got my MSCS from UW/Madison (1979-1981), I wrote tens of thousands of lines of code: compilers, interpreters, a database management system, even mock operating systems and device drivers. I believe it immensely strengthened my programming skills.
Interesting thing about MSCS programs: they have (or had at the time) huge numbers of students with BS (or even BA) degrees in fields other than computer science. (I started with a BS in Physics and an MS in Mathematics.) The non-CS majors weren't coddled; there were a couple of senior level classes we could take for graduate credit to "catch up", but we were expected to learn C and Unix based on a couple of one hour (each) supplemental lectures.
After complaining to the SETI folks that the Windoze client software would run so much faster if they used 3DNow! optimizations (you did see this was a 3DNow! advocacy site, didn't you?), they say:
What should SETI@Home do about it? Limit the submission of units to one machine per email address. Disallow the transfer of units between accounts. Retire the Top 20 teams and accounts of all categories every 2 months, that will make it less appealing to push so hard to the top.
But, above all, let people have the fastest client software possible, that is the least they deserve if you want them to support the project. Put the code into open source and let the best coders of the planet tackle it!
Cynical translation: "Kick off all those darned RISC users with their really fast floating point units and their farms of multiprocessor servers! Let us recode your client so our favorite processor looks better!"
Now I remember why we're looking for intelligence Out There....
Interesting... one of the articles says Myst creator Robyn Miller has started a CGI film production company, "Land of Point". The article has a URL to the company's site... but the server (running IIS 4.0) returns HTTP Error 403 - Forbidden. Afraid of the Slashdot effect?
(Remember a song, "Me and my Arrow", used many years ago in a car commercial? It was from a film called "The Point," which takes place in the Land of Point.
Slashdot Mirror
Funny; I used to work with one of the authors at a Silicon Alley startup. He was a complete non-techie, a "Webmaster" who produced neither HTML nor English for the site, but instead concentrated on the "overall architecture" of the site. Whatever that meant.
He's probably angry and bitter his experience there didn't make him rich. On the other hand, he didn't stay around long enough to vest any of his options. On the third hand, I never noticed him working particularly long hours.
He feels sorry for the guys who kept the servers going? That would be a couple of 23-year-old college dropouts (and one Columbia grad), who rarely showed up before 11 a.m. but often stayed past midnight, got paged at 3 a.m., and worked a lot of weekends. I'm glad the authors feel sorry for them ... but at least one of them doesn't know a thing about that life.
(Since you're probably wondering: I was a lead developer, then "Manager, Content Acquisition Software," for a team of five Unix/Perl programmers and one NT/C++ guy. I didn't work very long hours, unless you counted the productive four hours every day on the train. I stayed around long enough to vest about a third of my ten-cent-a-share options. But when the company laid off most of the technical staff, I saw the writing on the wall. I found a job closer to home, and let my options expire. The company's doing okay, with twenty employees instead of the hundred and twenty it had when I was there. I don't think it'll ever IPO. Bottom line: I'm not rich, either. And, yes, when the sysadmins weren't around, I was one of the other guys who kept the servers up.)
The first public information on Unix (other than printing the troff files from the distribution tapes) was the Unix issue (July-August 1978) of the Bell System Technical Journal. Thompson & Ritchie had an article on the kernel, I think there were articles on the compiler, the shell (Mashey or Bourne?), ed, and nroff/troff.
This issue holds the distinction of being the most-stolen magazine in AT&T/Lucent/Telecordia technical libraries.
The "Bell System" was AT&T, including Western Electric (now Lucent, more or less), the "baby Bells" (now Bell Atlantic et. al.), and Bell Telephone Laboratories (now AT&T Labs, Bell Labs, and Telecordia). It was broken up when AT&T "divested" the local operating companies in 1984.
When I program in C and C++, I always worry a stray pointer will bring my program crashing down. In Perl, I can use an undefined variable, or call a function I expected to be autoloaded ... but in practice, I'm far less likely to see these kinds of bugs once I've done a sane amount of testing. (Put another way: An equal amount of testing gives me more confidence in a Perl program than the "equivalent" C or C++ program.)
P.S.: I've yet to overflow a Perl string. Wish I could say that about C char arrays....-(
Previous /. posts suggest Linux is being used widely through the United Kingdom's civil service. How widely? Are some folks running servers (old legacy boxes or new applications) on Solaris, BSD, NT, or some other OS?
Check out the Index of Famous Dog names. "Welcome to Dog Central! The title says it all; this is an Index of Famous Dogs throughout history, both in reality and in fiction. This is not to be considered a complete list, nor can I account for the accuracy of all the information here (just some of it)."
Cats and other critters also available on that site.
(Back in 1982, we needed to move our systems from one building to another. We were told to lose the Star Trek themed names. We picked state names; useful since all of them have two letter abbreviations. "lznv" was the "Nevada" system in the Lincroft building (location code "lz"; dunno what the "z" stood for). It was funny, though, tracing the bus from Hawaii to New Hampshire.-)
I use (*sigh*) IE5 to show me offered cookies (my making custom changes to the Internet security zone). Any image server that wants to give me a persistent cookie goes into my Restricted Sites zone, which (among other things) are prevented from giving me cookies.
Here's my list:
The hosts without the * are the scariest. ngnetwork.pcworld.com is offering me a cookie named NGADsomething. Want to bet that host, though in the pcworld.com domain, is actually the IP address of an ngadcenter host? In other words, you can block the ngadcenter domain and NGAD can still track you. Ouch.
I haven't (yet) set up host lookup to set all those hosts and domains to 127.0.0.1, but I'm thinking about it. --PSRC
Netscape's bookmarks are just a Web page. Make that your start page!
"The Unofficial Guide to ..." was a series of MS Windows books edited (sometimes written) for Addison Wesley by Woody Leonhard. He's since edited (and sometimes written) the "Annoyances" series of MS Windows books for O'Reilly. (Both series recommended.)
Coincidence?
Mr. Katz suggests making (in effect) 'Net-based, short turnaround peer review a standard way of doing business; journalism for the 21st century.
I see two potentially serious problems with this (in addition to the obvious problem with not waiting to release "scoops"):
(1) In no time at all, there would be an unacceptably low ratio of qualified reviewers to material to be reviewed. (For example, Slashdot provides a large and self-moderating community, but could even we review every ZD story for accuracy?)
(2) Unless posted in some secure area, publishing drafts is still publishing, and could raise additional probability of slander lawsuits. (One would presumably publish early drafts, simultaneously with the usual internal review that checks for this kind of liability.)
In the consumer electronics business (at least in the U.S.), there are some well known "sweet spots" in the price spectrum. They're mostly multiples of $100. Hit these price points, and you'll encourage impulse buys by various kinds of consumers in various circumstances. The most well known price points are $100, $300, $500, and $700. (Think what price a VCR, or a fairly big screen TV, or a camcorder had to hit before you considered buying one.)
$700 PCs are a huge hit. WebTV seems stuck at $500, and it's anything but taking the market by storm. 3DO proved conclusively that the $500 price point is too high for game consoles.
Nintendo and Sony have found a good price point at $200 for game consoles. Sega had to hit that point.
Households with home LANS or cable modems, almost by definition, live higher on the price curve. It makes sense for Sega to offer an upgrade kit, or a higher priced (because they can get away with it) version, for that market segment. But Sega shouldn't blow the price point for the majority of their customer base.
(I worked on four different consumer telecommunications services from 1992 to 1998. I saw lots of products miss the price points, even though we knew better.)-:
Seems to me the real problem is the AOL password is stored in the clear on every client's machine. (That's the only way a trojan horse could e-mail it out, right?)
The last place to insist on the "corporate" look is Wall Street. You will show up at 8 p.m. promptly, in a nice suit, work past the point of exhaustion every day, produce lots of crap in a hurry, get paid six figures, and burn out in six months.
All so business guys (being paid seven figures) can move $$$ from point A to point B and take a percentage.
Wall Street does jump on the hottest technology (or technological fad) pretty fast. They had Solaris workstations years ago (on non-techies' desks, because that's what the applications were designed to run on; moving to NT now), used Smalltalk for a while, jumped to Java early. (Their business logic changes at whim, so they need really rapid application development.) It can be an interesting entry on your resume. Not mine, thanks.
(They also suck up 90% of the good technologists, and inflate salaries enormously, in Manhatten. When I was a manager at a Silicon Alley startup, that made recruting tough.)
India was a British colony for many years. Most Indians who come to the U.S. speak excellent English, in addition to an Indian language.
There are approximately a zillion official languages in India. The only one spoken everywhere in the country is English, so that's everyone's second language. Like Rob, I'd hate to see them treated as second-class citizens. The good news is, many get sponsored for citizenship.
The South Asians in the U.S. are affluent enough for plane fare here. Nearly 100% of them have bachelor's degrees; many of them have master's degrees; the majority of them have some industry experience.
The big hype about "Indian" programmers was how the marketing types were going to outsource development to folks living in Pakistan, India, and other places where salaries are much lower. Hasn't much happened yet.
As someone else said, breaking up Microsoft is like breaking up an anthrax spore: you get two of what you didn't want one of.
Regulating Microsoft is like the IRS taking over the Mustang Ranch and still operating it as a brothel: the customers still get screwed, just not as creatively. (I heard the Mustang Ranch is going to be operated as a real ranch for real horses.)
What's left? Bill Gates laughed at a million dollar a day fine. We don't want him to laugh at the proposed remedy; we want him scared enough to modify his behavior.
How about a nice round billion dollar (US) fine?
I'm serious. It's somewhere between a slap on the wrist and chopping a hand off. In and of itself, it doesn't prevent Microsoft from being bad in the future, but if future offenses can lead to future significant fines, it might get their attention. They might change their ways, they might agree to a settlement. I haven't heard any other proposal that, realistically could lead that way.
I don't know where this idea that C++ generates slow exectuables came from.
Here's one place.
In Kernighan and Pike's (excellent) The Practice of Programming (check out the Slashdot review), a program is written in C, C++, Perl, Java, and Awk.
The C++ and Perl versions were about equally fast; the C version was about five times faster. (I'm not counting the broken Windows deque-based version.) See p. 81 for the numbers.
I hope (in my copious spare time:-) to figure out just why the C++ version is so much slower.
http://www.scriptics.com /people/john.ousterhout/wrist.html
What John Ousterhout (creator of Tcl/Tk) does about his RSI. Summary: Windows PC running Dragon Systems NaturallySpeaking, plus a2x to make the Windows PC act as a keyboard for a Unix box running X.
Here is an article from Byte reviewing good mikes with NaturallySpeaking. (Formatting's terrible, but at least it's all on one Web page.)
Accuracy is reported to be 99%+.
http://www.perform ancecomputing.com/reviews/software/9904c.shtml
Gateway is married to Intel. They always have been. They have never used AMD chips
Wrong.
Check this URL for the Gateway Select(tm) line of Gateway AMD-based systems.
Gateway will discontinue selling AMD-based systems, but they haven't yet.
Forget the MCSE/CNE/A+ stuff. The closest thing programmers have to a PE is a Certified Computing Professional designation from the Institute for Certification of Computing Professionals.
Having said that, let me emphasize the "closest thing" isn't very close! In many engineering disciplines, there is a well-understood, well agreed-to body of practices that no competent professional would violate. In software "engineering," there's almost concensus that goto statements should be avoided most of the time, and that's about it.
As a result, darned near no one has a CCP, and most employers don't know what it means, let along care whether or not you have one.
When I got my MSCS from UW/Madison (1979-1981), I wrote tens of thousands of lines of code: compilers, interpreters, a database management system, even mock operating systems and device drivers. I believe it immensely strengthened my programming skills.
Interesting thing about MSCS programs: they have (or had at the time) huge numbers of students with BS (or even BA) degrees in fields other than computer science. (I started with a BS in Physics and an MS in Mathematics.) The non-CS majors weren't coddled; there were a couple of senior level classes we could take for graduate credit to "catch up", but we were expected to learn C and Unix based on a couple of one hour (each) supplemental lectures.
What should SETI@Home do about it? Limit the submission of units to one machine per email address. Disallow the transfer of units between accounts. Retire the Top 20 teams and accounts of all categories every 2 months, that will make it less appealing to push so hard to the top.
But, above all, let people have the fastest client software possible, that is the least they deserve if you want them to support the project. Put the code into open source and let the best coders of the planet tackle it!
Cynical translation: "Kick off all those darned RISC users with their really fast floating point units and their farms of multiprocessor servers! Let us recode your client so our favorite processor looks better!"
Now I remember why we're looking for intelligence Out There....
Luke Skywalker was 19 in STAR WARS: A New Hope... and I was 19 when I first saw it in 1977 (only 9 times that summer).
(Remember a song, "Me and my Arrow", used many years ago in a car commercial? It was from a film called "The Point," which takes place in the Land of Point.
You're speaking just for yourself when you say that, right?-) --PSRC
... option 4 (which I rather like) is "Debian Linux", not "Debian GNU/Linux". Even the Debian folks don't pronounce "Linux" the way RMS wants???"