I misread the subject line on this article, thought it read Fisting for Phishers. Now that is a punishment that would work pretty good, once word got out!
No, you are pretty much straight on - it's funny because nobody can ever become an expert at C++. Even Kernighan once admitted 'Fuck if I know, I making this shit up as I go.'
It's called the 400sc, and you can pick one up for $313 shipped right now from their small business division (they are blowing them out to clear the channel.) See my sig for details, although my link may be out of date.
Honestly I'm not an AMD fanboy (even though I should be, all things considered)... but I would say that the article glossed over the part I found most interesting : with an entry price of about a hundred bucks the Celeron D will overclock to 3.5GHz with relative ease.
I wouldn't say he meant the same computer failing every few weeks - lets say clone boxes only have a single problem (PSU/hard drive/fan/whatever) once a year (pretty fair assumption if you get cheap enough hardware, ya?).
If you have a 1/year failure rate per machine and 20 machines, that's a failure rate of about one every 20 days, distributed evenly if you are lucky (and several in one week if you are not lucky.)
That said, almost ALL of the system failures I have seen in the past 15 years point to 'insufficient cooling' as the root cause. Power supply fans, CPU fans, and case fans are easy to skimp on when building cheap machines and almost always the first thing to fail - and when your $3 CPU fan fails it kills the entire $500 machine. Hard drive lifespan drops due to heat - and the heat comes from a weak exhaust fan (doesn't have to fail, just be 'weak'.) I'm guessing most PSU failures are from overheating also, or from 'running hot' for a long period of time (once again due to weak airflow, not necessarily a fan that stopped.)
Last time I had anything to do with hardware, MTBF for hard drives had an unwritten disclaimer : Mean Time Between Failures assuming you replace the hard drives with new (identical) hard drives every year (or at least on a regular schedule.)
Granted it has been a while, and nobody actually does that, and I can't find any supporting documentation on the matter... but I believe that MTBF carries that stipulation. The theory being that if you replace the drives before they hit the other side of the bathtub curve (this was a little longer than a year when I was doing it, damn MFM drives) then the MTBF went way up. If unexpected failure is really, really bad - then it is a way to keep unexpected failures to a minimum (and a good way to fill your home network with cheap 1 year old hard drives.)
This is a good place for someone to slam me for being wrong, if so inclined.
Go to Dell, Small Busines side, click servers. Dell is blowing out their entry level server to make room for the next generation (which isn't nearly as nice, IMHO.) Left hand side has a link to their entry level 400sc for $323 shipped to your door.
P4 2.8GHz w/ HT 128M ECC RAM 48x CD / keyboard / mouse 40G IDE, upgrade to 80G for $20, or to 160G for $60. Intel Gigabit NIC PCI 8M ATI video card, system has an AGP slot if needed. One year onsite warranty.
Systems run cool and silent as delivered (I have two in my home office; I use them as desktops.) Have onboard 2x SATA and I believe will do hardware RAID 1 (will not do RAID 0 using the onboard) - maybe not, I don't remember. Onboard sound, if anybody cares, and 6x USB 2.0
It will use regular DDR pc3200 in dual channel if you get memory designed for the Dimension 4600 - I recommend Crucial or Kingston. You are going to want more memory.
Maybe you can build a fast, reliable, silent box for less than $325 per, including gigabit networking, ECC memory, server chipsets and onsite warranty - but honestly how much cheaper? Plus think a year down the road when something bad does happen... when you need to defend the stability of your network and server uptime, which is going to hold more water : 'hey I built those machines myself out of the cheapest parts available" or 'hey I bought Dell business class server hardware.'
In the long term you will appreciate having a single system to support, esp if you are the only admin. I am abou to retire two other perfectly good systems at home and add two more of these just so I can have a single Gold Disk for reinstalls / recovery - being able to put a single spare machine in the closet and know that it guarantees your service level agreements as a complete cold swap machine (or spare parts box) is priceless.
Actually the Linux folks see it as a fun hack but don't actually use it because it is so slow. Windows users will flock to it and probably won't notice the difference.
That said, answers: 1. You are going up against people that will work for free, and people that have 25+ years experience (generally not the same people, however.) You have enough experience to do intermediate level Python, VB, VBA, and SAS. I wouldn't mention SAS because nobody uses it except college students doing statistics. * 2. If you put your credentials on the web, make sure the domain you use is your own domain, and that you use it only for professional purposes. If the company can poke around and find pictures or anything not 'work related' it will work against you. Professional business cards are a good idea, as is a professional business email address. 3. How about 'Free Trial Thinking' - then come up with a story about how you came up with the name. 4. Get some professional experience, even if you have to do it for free. The United Way has a hundred different non-profit agencies in your area that could use your help and would acknowledge your contributions.
Footnote * : companies have proved that they will hire warm dead bodies from a morgue if the bill rate is low enough. If you don't mind living like you lived in college (ie. poor, with roomates) you can present yourself as a qualified developer working for $12.50 an hour and go to work tomorrow. Want to hold out for $60 an hour... expect to put some serious work into your job search. One nice thing about contracting is that you have no loyalty - get a better gig, bail with little or no notice (if a company wanted loyalty they should have hired you full time at a reasonable salary.)
Oh man, clients that screwed you over once are the BEST clients - if they come to you begging and crying for your help, knowing that you are upset about the last time they boned you... they are ready to pay top dollar, and pay it up front.
I had a company screw me over some completely weak amount, less than $500 - and they came back crying about me being the only one that knew how their network was set up and could fix their problems. I still haven't made as much money per hour, paid up front, as I did on that gig.
I agree with taking care of your good clients though.
Want to know how to get signed on retainer? You do work on a regular basis for a company for quite some time until they decide that you are the best person for the job. Tell them that you can continue to do that job for half the cost (or less)each month but will be working on an as-needed basis, fixing whatever needs fixed when it needs fixing (and generally remotely.) Negotiate something that is benefitial to them fiscally and at the same time open up your schedule and time frame. Bingo - retainer, and both parties benefit from the agreement.
Go to RealRates and look through there. Get one or more of Janet's books, and go to the message boards / BBS run on that site for more insight and feedback - that place is to contractors what this place is for nerds.
Enjoy. And come up with an answer to the following (follow the sig-link) :
Where do you plan on working? If you plan on working in India, it is probably a good plan.
After the last three completely fuxored projects outsourced from my department alone - if someone came to an interview proudly announcing that he was trained in India I would show him to the door with a polite 'don't call us / we'll call you.'
No offense intended, but the corporate focus is slowly shifting back from 'absolutely cheapest' to 'doesn't suck'. Thank God.
I thought that the Naval Nuke power program was your golden ticket, let you walk into any 3MileIsland power plant (I mean that in a good way - that's the only one I can name off the top of my head) and get started doing whatever Nuke power plant guys do.
Dang, I would like to think it is better than you represent just because I almost went that route.
So what the hell do we do with it? Granted it is a little bigger, faster than the 10GHZ to 30GHz cluster I'm asking about, but after you work out the details it is just a matter of scale.
This seems as good a place as any to ask that question (see my sig for details.)
Someone else has your password and is logging in, reading you email. The contents of Hotmail email can be read without flagging it as 'read' if you know what you are doing. That's why my ex-gf's hotmail account lasted more than a year after we broke up, even when she didn't connect to it for 31+ days in a row. Rhonda, is that you? If so, I still love you. Call me.
No, I would say the #1 problem is that Hotmail deletes your account if you are inactive for a month and there is no way to get your email history back.
I have seen it with my own eyes (in 1993) - it looks amazingly like our own Space Shuttle with the only striking difference I noted being the configuration of the jet engine cones / exhausts in the back. I don't remember off the top of my head what the differences were, only that I remember them being slightly different.
Granted I wasn't all up close and touchy / feely with it, and I had to be slightly discrete with my interest - I will have to go look over the picture(s) and see what else I can see.
I misread the subject line on this article, thought it read Fisting for Phishers.
Now that is a punishment that would work pretty good, once word got out!
$599 for the 2001fp right now at Dell (through tomorrow.)
Enjoy, I think I'm gonna get one too.
No, you are pretty much straight on - it's funny because nobody can ever become an expert at C++.
Even Kernighan once admitted 'Fuck if I know, I making this shit up as I go.'
Debian - Nerd
Gentoo - Nerd
Knoppix - Nerd
Linspire - Nerd that shops at Walmart
Mandrake - Nerd
MEPIS - Nerd
Red Hat - Nerd
Slackware - Nerd
SUSE - Nerd
It's called the 400sc, and you can pick one up for $313 shipped right now from their small business division (they are blowing them out to clear the channel.)
See my sig for details, although my link may be out of date.
Honestly I'm not an AMD fanboy (even though I should be, all things considered) ... but I would say that the article glossed over the part I found most interesting : with an entry price of about a hundred bucks the Celeron D will overclock to 3.5GHz with relative ease.
3.5GHz CPU for $100.
Damn.
I have a cheap cluster at home.
Now the question is WTF do I do with it? (Details in sig)
I wouldn't say he meant the same computer failing every few weeks - lets say clone boxes only have a single problem (PSU/hard drive/fan/whatever) once a year (pretty fair assumption if you get cheap enough hardware, ya?).
If you have a 1/year failure rate per machine and 20 machines, that's a failure rate of about one every 20 days, distributed evenly if you are lucky (and several in one week if you are not lucky.)
That said, almost ALL of the system failures I have seen in the past 15 years point to 'insufficient cooling' as the root cause. Power supply fans, CPU fans, and case fans are easy to skimp on when building cheap machines and almost always the first thing to fail - and when your $3 CPU fan fails it kills the entire $500 machine. Hard drive lifespan drops due to heat - and the heat comes from a weak exhaust fan (doesn't have to fail, just be 'weak'.) I'm guessing most PSU failures are from overheating also, or from 'running hot' for a long period of time (once again due to weak airflow, not necessarily a fan that stopped.)
Last time I had anything to do with hardware, MTBF for hard drives had an unwritten disclaimer : Mean Time Between Failures assuming you replace the hard drives with new (identical) hard drives every year (or at least on a regular schedule.)
... but I believe that MTBF carries that stipulation. The theory being that if you replace the drives before they hit the other side of the bathtub curve (this was a little longer than a year when I was doing it, damn MFM drives) then the MTBF went way up. If unexpected failure is really, really bad - then it is a way to keep unexpected failures to a minimum (and a good way to fill your home network with cheap 1 year old hard drives.)
Granted it has been a while, and nobody actually does that, and I can't find any supporting documentation on the matter
This is a good place for someone to slam me for being wrong, if so inclined.
Go to Dell, Small Busines side, click servers.
... when you need to defend the stability of your network and server uptime, which is going to hold more water : 'hey I built those machines myself out of the cheapest parts available" or 'hey I bought Dell business class server hardware.'
Dell is blowing out their entry level server to make room for the next generation (which isn't nearly as nice, IMHO.)
Left hand side has a link to their entry level 400sc for $323 shipped to your door.
P4 2.8GHz w/ HT
128M ECC RAM
48x CD / keyboard / mouse
40G IDE, upgrade to 80G for $20, or to 160G for $60.
Intel Gigabit NIC
PCI 8M ATI video card, system has an AGP slot if needed.
One year onsite warranty.
Systems run cool and silent as delivered (I have two in my home office; I use them as desktops.)
Have onboard 2x SATA and I believe will do hardware RAID 1 (will not do RAID 0 using the onboard) - maybe not, I don't remember.
Onboard sound, if anybody cares, and 6x USB 2.0
It will use regular DDR pc3200 in dual channel if you get memory designed for the Dimension 4600 - I recommend Crucial or Kingston. You are going to want more memory.
Maybe you can build a fast, reliable, silent box for less than $325 per, including gigabit networking, ECC memory, server chipsets and onsite warranty - but honestly how much cheaper? Plus think a year down the road when something bad does happen
In the long term you will appreciate having a single system to support, esp if you are the only admin. I am abou to retire two other perfectly good systems at home and add two more of these just so I can have a single Gold Disk for reinstalls / recovery - being able to put a single spare machine in the closet and know that it guarantees your service level agreements as a complete cold swap machine (or spare parts box) is priceless.
Sig follows - go answer the question for me :
Guess which ones pick up more chicks?
If they are bragging about their computers, I'm guessing neither.
Actually the Linux folks see it as a fun hack but don't actually use it because it is so slow.
Windows users will flock to it and probably won't notice the difference.
This isn't 1999, and the .COM boom is over.
:
... expect to put some serious work into your job search. One nice thing about contracting is that you have no loyalty - get a better gig, bail with little or no notice (if a company wanted loyalty they should have hired you full time at a reasonable salary.)
That said, answers
1. You are going up against people that will work for free, and people that have 25+ years experience (generally not the same people, however.)
You have enough experience to do intermediate level Python, VB, VBA, and SAS. I wouldn't mention SAS because nobody uses it except college students doing statistics. *
2. If you put your credentials on the web, make sure the domain you use is your own domain, and that you use it only for professional purposes. If the company can poke around and find pictures or anything not 'work related' it will work against you. Professional business cards are a good idea, as is a professional business email address.
3. How about 'Free Trial Thinking' - then come up with a story about how you came up with the name.
4. Get some professional experience, even if you have to do it for free. The United Way has a hundred different non-profit agencies in your area that could use your help and would acknowledge your contributions.
Footnote * : companies have proved that they will hire warm dead bodies from a morgue if the bill rate is low enough. If you don't mind living like you lived in college (ie. poor, with roomates) you can present yourself as a qualified developer working for $12.50 an hour and go to work tomorrow. Want to hold out for $60 an hour
Oh man, clients that screwed you over once are the BEST clients - if they come to you begging and crying for your help, knowing that you are upset about the last time they boned you ... they are ready to pay top dollar, and pay it up front.
I had a company screw me over some completely weak amount, less than $500 - and they came back crying about me being the only one that knew how their network was set up and could fix their problems. I still haven't made as much money per hour, paid up front, as I did on that gig.
I agree with taking care of your good clients though.
Want to know how to get signed on retainer?
You do work on a regular basis for a company for quite some time until they decide that you are the best person for the job. Tell them that you can continue to do that job for half the cost (or less)each month but will be working on an as-needed basis, fixing whatever needs fixed when it needs fixing (and generally remotely.) Negotiate something that is benefitial to them fiscally and at the same time open up your schedule and time frame. Bingo - retainer, and both parties benefit from the agreement.
Go to RealRates and look through there. Get one or more of Janet's books, and go to the message boards / BBS run on that site for more insight and feedback - that place is to contractors what this place is for nerds.
Enjoy.
And come up with an answer to the following (follow the sig-link) :
Where do you plan on working?
If you plan on working in India, it is probably a good plan.
After the last three completely fuxored projects outsourced from my department alone - if someone came to an interview proudly announcing that he was trained in India I would show him to the door with a polite 'don't call us / we'll call you.'
No offense intended, but the corporate focus is slowly shifting back from 'absolutely cheapest' to 'doesn't suck'. Thank God.
I thought that the Naval Nuke power program was your golden ticket, let you walk into any 3MileIsland power plant (I mean that in a good way - that's the only one I can name off the top of my head) and get started doing whatever Nuke power plant guys do.
Dang, I would like to think it is better than you represent just because I almost went that route.
Damn, how's that for a sign from God?
(Yes, I'm probably going to hell for that one.)
So what the hell do we do with it? Granted it is a little bigger, faster than the 10GHZ to 30GHz cluster I'm asking about, but after you work out the details it is just a matter of scale.
This seems as good a place as any to ask that question (see my sig for details.)
And this is different than the dot COM boom ... how?
-
Don't let anybody say I wasn't supportive of this mission.
Here's to India!
May your best initial efforts be equally as successful as the first Apollo mission!
Someone else has your password and is logging in, reading you email. The contents of Hotmail email can be read without flagging it as 'read' if you know what you are doing.
That's why my ex-gf's hotmail account lasted more than a year after we broke up, even when she didn't connect to it for 31+ days in a row.
Rhonda, is that you? If so, I still love you. Call me.
No, I would say the #1 problem is that Hotmail deletes your account if you are inactive for a month and there is no way to get your email history back.
I have seen it with my own eyes (in 1993) - it looks amazingly like our own Space Shuttle with the only striking difference I noted being the configuration of the jet engine cones / exhausts in the back. I don't remember off the top of my head what the differences were, only that I remember them being slightly different.
Granted I wasn't all up close and touchy / feely with it, and I had to be slightly discrete with my interest - I will have to go look over the picture(s) and see what else I can see.