I think highly focused, technical questions are unfair and a waste of time. Syntax, symantics, and API questions are for lam3rs who have every language and libary reference manual memorized.
I prefer to have a simple "chat". I want to see how an engineer thinks when they are comfortable, not when they are sweating the details of my stupid C++ multiple inheritance question.
If I DO ask technical questions, I explicitly state that pseudo-code is acceptable.
The problem is, companies have forgotten The Mythical Man Month. My company can hire 4 engineers overseas for 1 U.S. engineer.
However, in most cases, it isn't worth it. You get 4 junior programmers with a resulting lower quality and usability. You also get poor performance and hard-to-maintain code. I don't see time savings either. Most projects with these overseas teams are just as far behind schedule.
In one case, a project was dying because the overseas team could not overcome performance problems. A single U.S. engineer (me) was able to improve the performance 1000x. The result was a useful product vs. a useless product.
Personally, I'm not afraid of losing my job to cheap, foreign labor. White collar jobs cannot be so easily transferred to cheap alternatives. So far, the only threat I see is management continuing to repeat the mistakes made famous by The Mythical Man Month.
I have found PROVANTAGE to be the best overall experience. I found them through http://ibuyernet.com/. They don't have the absolute lowest prices, but they are always very close. They're shipping costs are fair. Delivery is very fast.
MOST IMPORTANT: Customer service is top notch. Order tracking and returns are very good.
That was my first thought. We have to redefine the term supercomputer. It used to be that these massively parallel systems were connected through high-bandwidth memory.
According to the detailed article, it is made up of 700 nodes connected to a SAN. Hence, the term supercomputer seems to cover a network of computers.
Given that broader definition, one could argue that metaprocessors like United Devices, SETI@Home, distributed.net are supercomputers. Although their connectivity is poor, they have a lot more than 1400 processors.
I had the fortunate experience of attending a supercomputing conference in 1995 while working for HP. This was back when Thinking Machines, etc. were in their heyday. At the end of the conference, we talked about future supercomputers. Everyone conceded that the internet was the supercomputer of the future, where people would lease out compute power.
I think what we are seeing now is the nascent stages of this technology. Idealistic research projects may give way to capitalistic projects. Imagine getting paid to share your computer. Take the $24.5M and pay people $50 a system. You could get 490,000 systems at your disposal.:-)
Actually, Intel is a sponser. The provider of the service is United Devices.
I always thought SETI@Home was cool, from a technical point of view, but I couldn't get very excited about analyzing random noise from space.
THINK ABOUT THIS: With the proliferation of encryption and compression technology, would we even be able to detect our own data? Ok, sure, it's may be on a carrier signal, blah blah blah, but once we gathered the data, what would we make of it? I think we are boldly assuming that alien transmissions are cleartext.:-)
Anyhow, back to United Devices distributed cancer research...
In 1999, my 3 year old daughter died of a brain tumor. In the process of her treatment I was exposed to the world of cancer research and discovered how little we actually know. After a biospy, they couldn't even classify her tumor. It didn't fit the classification system, which means many of our assumptions are fundamentally flawed.
Most of the cancer "treatments" are just blind guesses as to what might work. Clinical research shows that some treatments have a statistically better effect and are thus put into use.
What appeals to me about the distributed computer research is that it accelerates the whole process of guessing and helps target specific treatments that have a higher chance of working. Thus, the clinical research can spend more time testing better-than-average treatments.
Please consider participating in a truly useful project. Here is the site for my daughter with a link to the UD site: In Memory of Abby and Kevin
DVD+R uses a similar dye-based, write-once medium, like DVD-R, giving it an equal chance of compatibility. In fact, basd on my research, the DVD+R might even be more compatible. It shares the same characteristics of DVD+RW in terms of how the writer identifies a disc and aligns itself for writing.
When you look at DVD-R, DVD-RW, DVD+R, and DVD+RW, they can be categorized along two axes:
1) R vs RW
The key difference here is in the physical recording medium. Write-once media uses dyes, which have relatively high reflectivity and thus play well with DVD-ROM/DVD-Players. Rewriteable media uses organic phase-change compounds that have low reflectivity, and hence confuse many DVD-ROM/DVD-Players.
2) + vs -
The DVD+R(W) vs DVD-R(W) discs differ in how they "communicate" to the writer. Essentially, writable discs have extra info not seen on stamped discs and ignored by readers. This info tells the writer what type of disc it is, and how to align itself for writing. DVD-R(W) uses groove recording with address info in the land areas. DVD+R(W) uses high-frequency wobbled grooves, which the writer uses to align to the disc.
Of course, Jim Taylor explains it better than I can in his DVD FAQ. I highly recommend his book "DVD Demystified Second Edition". Although it doesn't cover DVD+R, the updated FAQ does. This book really should be called "Home Audio and Video Demystified". The explanation of aspect ratios for 4:3 and 16:9 TVs is worth the cost of the book!
Cutesy names are important for humans. Formulaic names are for robots and computers to keep track of.
I use this rule of thumb: If I am going to be talking about a server to my co-workers, then test it out. Pretend (or actually) to talk to someone about the server and think about how painful it is to utter the name.
I absolutely loathe patterned names. Think of it this way - it can actually cost the company money. I actually calculated the time wasted by our stupid corporate (robotic) naming scheme. I was so frustrated, I wanted to prove to management that the names we were using were costing them money (if you believe time == money).
Example:
Me: I'm having trouble with mmxgg1.
Sysadmin: What's the problem?
Me: When I try to NFS mount mmxgg1 from isdmv2, isdmv2 hangs.
Sysadmin: Which mount point?
Me: I'm mounting mmxgg1:/home/me
Sysadmin: Do you want me to reboot isdmv2?
Me: Yes. And change those F'in names in DNS while you're at it!
As many others have pointed out, there are methods for tracking location and function. You should always have a database or spreadsheet to keep track of that.
I can almost guarantee you that once you assign a functional name to a server, it's purpose will be reassigned.
We had a bunch of servers named web01, db01, app01, etc. (names slightly changed to protect the guilty). Before we knew it, databases were being installed on web01, app servers on db01.
Wow! You've really brought out an interesting point. To sum up, corporations cannot vote and hence should not be part of the political process.
The American people have sold their souls to corporations.
What realy bothers me about the whole SSSCA thing is that our government is furiously protecting an industry that has little tangible value, in terms of contributing to society, yet makes the most money. Go figure. If only that much money went to cancer research or world health/hunger problems.
As much as I like capitalism, this is capitalism run amuck.
If you speak in generalities, you could easily defend this guy. Yes, he was a very nice guy - maybe too nice (a whole other topic of discussion).
Your argument is weak. Probably 99.99% of the US population doesn't know what it means to implement 64-bit support in RISC hardware/Un*x OS+tools.
However, this guy is paid because he is that 0.01% of the world that knows this stuff. Remember I said he used to be an engineer. He doesn't have the luxury of being naive about architectural details that are the cornerstones of our jobs.
I don't buy your argument one bit. In a highly specialized development organization, you need to know your business.
We're not talking about cheesy applications software powering the latest dot-bomb marketing site. When you work at tha architecture level of computing, you better get it right, or the whole deck of cards comes crashing down. Now your stupid web app doesn't work because the compiler generated bad code. Try to debug that!
This is like putting the cart before the horse. Your boss is defining the solution instead of the problem.
I am a strong advocate of simplicity, which is why I think Java is a great choice. Multiple inheritance and operator overloading are not sorely missed. Other threads on this topic explain why multiple inheritance is not a problem in Java.
More importantly, IMHO, Java eliminates the #1 bug producer in languages like C++ - pointers and the subsequent memory errors. I surely do not miss managing my own memory in C++, etc.
Java is not perfect. For example, EJBs have been WAY overhyped and overused. Most of the time their usage in the real world is akin to using an excavator to dig a hole for a mailbox post.
Tim certainly deserves a lot more credit than the popular press gives him. Why are we getting all worked about about a stupid article from a stupid magazine?:-) We just can't help it, I guess.
Entertainment should have John Carmack on the list. He helped create a whole genre of gaming? Instead, they have Sonic the Hedgehog on the list. Gimme a break! I would even accept Richard Garriot over the nobody's listed there. Also, the only reason Napster succeeded is because it got so much press and legal suits. The very beast that made it rise to the top also brought it down. I'm quite happy with Gnutella and LimeWire thank you.
You forget... They invented the Open Test paradigm. Pump out some poorly designed (if at all), buggy code, call it a release, and let the whole world test it. Free QA.
I think the one thing we can all agree on is that Linus did not create the Open Source movement.
As others have pointed out, it is hard to identify a single person. My gut reaction was to say, "Hey, Richard Stallman invented the open source movement." After reading the comments here, I realized I was wrong.
However, without the FSF efforts for gcc etc., open source would have a hard time getting to where it is today. I'm pretty sure that I'm not thankful for EMACS though.
They have a great web-based management interface.
Features that I like for my personal domains:
- DNS managed by them they have entries for roostme.com (or whatever) and www.roostme.com, which is all I need.
- Framed or forwarded website hosting, if you need that sort of thing
- Email forwarding. This was the best feature! By using this it bought me some time until I had my sendmail config ironed out.
- DNS MX record for email. Once I had my sendmail working, I canceled forwarding. Then I had their DNS register the MX record for my server.
Note: The interface was not 100% perfect. It took a bit of thinking on my end to be sure I switched off the email forwarding and MX'ed my system properly.
Imagine a company hiring hackers to break into competitors sites to put important keywords in the unthesaurus.
For example, what if you hacked 3com's site to put the words 'ethernet' and 'network' in their unthesaurus. It's unlikely that a professional company like Linksys or others would do this, but it is entirely possible.
You could argue that meta keywords should take precedence, but I'm sure the hacker would remove those words from the meta keyword list.
Remember the band 'The The' from the '80s. It would seem to be damn near impossible to find them via normal search techniques.:-)
I did a quick test, here are the results:
Yahoo: A (listed the band site via their web site listings; official site was 4th in list)
Google: F (quoting didn't help)
Northern Light: C (found relevant matches, but the official site was nowhere to be found on the first 2 pages)
altavista: A+ (official band site was #1 in list)
Nowadays, you need to think about "searchability" when picking the name for just about anything. That is, assuming you want to be easily found on the web.
I guess that's where dopey marketing names like 'Itanium' actually make sense. Very unambiguous search criteria.
Ok. I'll say some more. For most searches, google's algorithm does a tremendous job of bringing the relevant sites to the top of the list.
In fact, when I look for product info and don't get the manufacturer's site first in the list, I consider that a strike against them - i.e. their web presence is put into question.
Slashdot Mirror
'nuff said. :-)
Absolutely!
I think highly focused, technical questions are unfair and a waste of time. Syntax, symantics, and API questions are for lam3rs who have every language and libary reference manual memorized.
I prefer to have a simple "chat". I want to see how an engineer thinks when they are comfortable, not when they are sweating the details of my stupid C++ multiple inheritance question.
If I DO ask technical questions, I explicitly state that pseudo-code is acceptable.
So THAT'S where Quake got it's name!
:-b
I know, it has the opposite effect and induces sleep.
However, if you drink enough of it, your aim can improve. Just aim between the "2 sarges".
The problem is, companies have forgotten The Mythical Man Month. My company can hire 4 engineers overseas for 1 U.S. engineer.
However, in most cases, it isn't worth it. You get 4 junior programmers with a resulting lower quality and usability. You also get poor performance and hard-to-maintain code. I don't see time savings either. Most projects with these overseas teams are just as far behind schedule.
In one case, a project was dying because the overseas team could not overcome performance problems. A single U.S. engineer (me) was able to improve the performance 1000x. The result was a useful product vs. a useless product.
Personally, I'm not afraid of losing my job to cheap, foreign labor. White collar jobs cannot be so easily transferred to cheap alternatives. So far, the only threat I see is management continuing to repeat the mistakes made famous by The Mythical Man Month.
I have found PROVANTAGE to be the best overall experience. I found them through http://ibuyernet.com/. They don't have the absolute lowest prices, but they are always very close. They're shipping costs are fair. Delivery is very fast.
MOST IMPORTANT: Customer service is top notch. Order tracking and returns are very good.
That was my first thought. We have to redefine the term supercomputer. It used to be that these massively parallel systems were connected through high-bandwidth memory.
:-)
According to the detailed article, it is made up of 700 nodes connected to a SAN. Hence, the term supercomputer seems to cover a network of computers.
Given that broader definition, one could argue that metaprocessors like United Devices, SETI@Home, distributed.net are supercomputers. Although their connectivity is poor, they have a lot more than 1400 processors.
I had the fortunate experience of attending a supercomputing conference in 1995 while working for HP. This was back when Thinking Machines, etc. were in their heyday. At the end of the conference, we talked about future supercomputers. Everyone conceded that the internet was the supercomputer of the future, where people would lease out compute power.
I think what we are seeing now is the nascent stages of this technology. Idealistic research projects may give way to capitalistic projects. Imagine getting paid to share your computer. Take the $24.5M and pay people $50 a system. You could get 490,000 systems at your disposal.
Cancer.
:-)
Actually, Intel is a sponser. The provider of the service is United Devices.
I always thought SETI@Home was cool, from a technical point of view, but I couldn't get very excited about analyzing random noise from space.
THINK ABOUT THIS: With the proliferation of encryption and compression technology, would we even be able to detect our own data? Ok, sure, it's may be on a carrier signal, blah blah blah, but once we gathered the data, what would we make of it? I think we are boldly assuming that alien transmissions are cleartext.
Anyhow, back to United Devices distributed cancer research...
In 1999, my 3 year old daughter died of a brain tumor. In the process of her treatment I was exposed to the world of cancer research and discovered how little we actually know. After a biospy, they couldn't even classify her tumor. It didn't fit the classification system, which means many of our assumptions are fundamentally flawed.
Most of the cancer "treatments" are just blind guesses as to what might work. Clinical research shows that some treatments have a statistically better effect and are thus put into use.
What appeals to me about the distributed computer research is that it accelerates the whole process of guessing and helps target specific treatments that have a higher chance of working. Thus, the clinical research can spend more time testing better-than-average treatments.
Please consider participating in a truly useful project. Here is the site for my daughter with a link to the UD site:
In Memory of Abby and Kevin
Sorry - I just had to say it.
DVD+R uses a similar dye-based, write-once medium, like DVD-R, giving it an equal chance of compatibility. In fact, basd on my research, the DVD+R might even be more compatible. It shares the same characteristics of DVD+RW in terms of how the writer identifies a disc and aligns itself for writing.
When you look at DVD-R, DVD-RW, DVD+R, and DVD+RW, they can be categorized along two axes:
1) R vs RW
The key difference here is in the physical recording medium. Write-once media uses dyes, which have relatively high reflectivity and thus play well with DVD-ROM/DVD-Players. Rewriteable media uses organic phase-change compounds that have low reflectivity, and hence confuse many DVD-ROM/DVD-Players.
2) + vs -
The DVD+R(W) vs DVD-R(W) discs differ in how they "communicate" to the writer. Essentially, writable discs have extra info not seen on stamped discs and ignored by readers. This info tells the writer what type of disc it is, and how to align itself for writing. DVD-R(W) uses groove recording with address info in the land areas. DVD+R(W) uses high-frequency wobbled grooves, which the writer uses to align to the disc.
Of course, Jim Taylor explains it better than I can in his DVD FAQ. I highly recommend his book "DVD Demystified Second Edition". Although it doesn't cover DVD+R, the updated FAQ does. This book really should be called "Home Audio and Video Demystified". The explanation of aspect ratios for 4:3 and 16:9 TVs is worth the cost of the book!
Cutesy names are important for humans. Formulaic names are for robots and computers to keep track of.
I use this rule of thumb: If I am going to be talking about a server to my co-workers, then test it out. Pretend (or actually) to talk to someone about the server and think about how painful it is to utter the name.
I absolutely loathe patterned names. Think of it this way - it can actually cost the company money. I actually calculated the time wasted by our stupid corporate (robotic) naming scheme. I was so frustrated, I wanted to prove to management that the names we were using were costing them money (if you believe time == money).
Example:
Me: I'm having trouble with mmxgg1.
Sysadmin: What's the problem?
Me: When I try to NFS mount mmxgg1 from isdmv2, isdmv2 hangs.
Sysadmin: Which mount point?
Me: I'm mounting mmxgg1:/home/me
Sysadmin: Do you want me to reboot isdmv2?
Me: Yes. And change those F'in names in DNS while you're at it!
As many others have pointed out, there are methods for tracking location and function. You should always have a database or spreadsheet to keep track of that.
Absolutely!
I can almost guarantee you that once you assign a functional name to a server, it's purpose will be reassigned.
We had a bunch of servers named web01, db01, app01, etc. (names slightly changed to protect the guilty). Before we knew it, databases were being installed on web01, app servers on db01.
What a stupid mess!
Wow! You've really brought out an interesting point. To sum up, corporations cannot vote and hence should not be part of the political process.
The American people have sold their souls to corporations.
What realy bothers me about the whole SSSCA thing is that our government is furiously protecting an industry that has little tangible value, in terms of contributing to society, yet makes the most money. Go figure. If only that much money went to cancer research or world health/hunger problems.
As much as I like capitalism, this is capitalism run amuck.
If you speak in generalities, you could easily defend this guy. Yes, he was a very nice guy - maybe too nice (a whole other topic of discussion).
Your argument is weak. Probably 99.99% of the US population doesn't know what it means to implement 64-bit support in RISC hardware/Un*x OS+tools.
However, this guy is paid because he is that 0.01% of the world that knows this stuff. Remember I said he used to be an engineer. He doesn't have the luxury of being naive about architectural details that are the cornerstones of our jobs.
I don't buy your argument one bit. In a highly specialized development organization, you need to know your business.
We're not talking about cheesy applications software powering the latest dot-bomb marketing site. When you work at tha architecture level of computing, you better get it right, or the whole deck of cards comes crashing down. Now your stupid web app doesn't work because the compiler generated bad code. Try to debug that!
Sorry to say, he was just THAT stupid. The scary thing is, he used to be an engineer.
FUNNY STORY: I was working for a major Un*x vendor implementing 64 bit support in the dev tools.
On a dare, one of our engineers messed with the manager:
Engineer: "We tried hard, but could only get 63 bits to work."
Manager: "That's ok. We can get that last bit in a patch."
Indeed!
This is like putting the cart before the horse. Your boss is defining the solution instead of the problem.
I am a strong advocate of simplicity, which is why I think Java is a great choice. Multiple inheritance and operator overloading are not sorely missed. Other threads on this topic explain why multiple inheritance is not a problem in Java.
More importantly, IMHO, Java eliminates the #1 bug producer in languages like C++ - pointers and the subsequent memory errors. I surely do not miss managing my own memory in C++, etc.
Java is not perfect. For example, EJBs have been WAY overhyped and overused. Most of the time their usage in the real world is akin to using an excavator to dig a hole for a mailbox post.
Tim certainly deserves a lot more credit than the popular press gives him. Why are we getting all worked about about a stupid article from a stupid magazine? :-) We just can't help it, I guess.
Entertainment should have John Carmack on the list. He helped create a whole genre of gaming? Instead, they have Sonic the Hedgehog on the list. Gimme a break! I would even accept Richard Garriot over the nobody's listed there. Also, the only reason Napster succeeded is because it got so much press and legal suits. The very beast that made it rise to the top also brought it down. I'm quite happy with Gnutella and LimeWire thank you.
You forget... They invented the Open Test paradigm. Pump out some poorly designed (if at all), buggy code, call it a release, and let the whole world test it. Free QA.
I think the one thing we can all agree on is that Linus did not create the Open Source movement. As others have pointed out, it is hard to identify a single person. My gut reaction was to say, "Hey, Richard Stallman invented the open source movement." After reading the comments here, I realized I was wrong. However, without the FSF efforts for gcc etc., open source would have a hard time getting to where it is today. I'm pretty sure that I'm not thankful for EMACS though.
$16/yr - 2yr mininum
They have a great web-based management interface.
Features that I like for my personal domains:
- DNS managed by them they have entries for roostme.com (or whatever) and www.roostme.com, which is all I need.
- Framed or forwarded website hosting, if you need that sort of thing
- Email forwarding. This was the best feature! By using this it bought me some time until I had my sendmail config ironed out.
- DNS MX record for email. Once I had my sendmail working, I canceled forwarding. Then I had their DNS register the MX record for my server.
Note: The interface was not 100% perfect. It took a bit of thinking on my end to be sure I switched off the email forwarding and MX'ed my system properly.
Imagine a company hiring hackers to break into competitors sites to put important keywords in the unthesaurus.
For example, what if you hacked 3com's site to put the words 'ethernet' and 'network' in their unthesaurus. It's unlikely that a professional company like Linksys or others would do this, but it is entirely possible.
You could argue that meta keywords should take precedence, but I'm sure the hacker would remove those words from the meta keyword list.
Remember the band 'The The' from the '80s. It would seem to be damn near impossible to find them via normal search techniques. :-)
I did a quick test, here are the results:
Yahoo: A (listed the band site via their web site listings; official site was 4th in list)
Google: F (quoting didn't help)
Northern Light: C (found relevant matches, but the official site was nowhere to be found on the first 2 pages)
altavista: A+ (official band site was #1 in list)
Nowadays, you need to think about "searchability" when picking the name for just about anything. That is, assuming you want to be easily found on the web.
I guess that's where dopey marketing names like 'Itanium' actually make sense. Very unambiguous search criteria.
'nuf said
Ok. I'll say some more. For most searches, google's algorithm does a tremendous job of bringing the relevant sites to the top of the list.
In fact, when I look for product info and don't get the manufacturer's site first in the list, I consider that a strike against them - i.e. their web presence is put into question.