This would not be such a problem if they would not be clueless about what my job is since it was my responsibility for making the deadlines; the programmer is the last in the development chain.
Another thing is that in such a small team (1 analyst, 1 data architect, 1 programmer + 1 clueless PHB) I become a quality control for them since I'm their customer.
If the cost would be only labour vs. hardware, you'd be right. But it also comes to the:
1) Lost productivity on the user's part.
You can have a Beowulf of these;) or Starfire on the back end, but it won't make your badly designed Oracle queries faster; it will just raise the amount of them that can be processed on one server.
And don't remind me about all of these front-end application crashes. Long wait and crashes make users frustrated and therefore less efficient, less motivated and the job suffers.
2) Lost business opportunities.
Related to 1). If my calculation of the build plan works 10 minutes instead of 4 hours, users can play with it and adjust the plan to find the best one instead of just guessing what is the best because the computer system can't process more than one of them a day.
3) Maintainability. Your deadlines might be made, but if the program is a nightmare from that point of view, you're going to pay later (it goes from different accounting bucket, so they don't think about it as much as they should).
4) Dilution of the profession. Sometimes you MUST have the best of the breed (NASA, FAA, embedded systems), but you can't find them easy. Even harder is distinguishing between them and VB lamers.
And this is in the country, where you can't practice most of the professions unless you're a member of the guild or union!
5 years as a programmer (mostly consultant) in the corporate world gave me the following insight:
Everyone's slogan is "Acceptable is enough."
This is why programming ceased to be an art and is now a simple combination of "requirements gathering, design, coding, testing". In one company they even put a wall, making different people do requirements gathering, database design and programming, often excluding actual developer (me) from the making the decisions on the early stages of the project, and holding me responsible for implementing their stupid decisions, based on the lack of understanding of what programmers do!
My rant has the following relation to the quote: in 60-th the cost of a computer was enormous comparing to a programmer's salary. Computers were not entirely pathetic, but any significant system should have been designed and implemented not just well but perfectly; otherwise, computers would not be able to handle it.
In the modern world it is cheaper to buy another server than to hire another couple of programmers; they just throw more iron into a problem to compensate for the lack of quality and efficiency.
So, the problem lies with the clueless managers and executives.
Plus, try thinking about the image of a scientist in the popular culture.
Scientist is Einstein who cannot make a step in the real life without help and forgets everything unrelated to the science (and does not know that Pepsi is better;-) if not evil maniac Fu Manchu.
With such an attitude it is not surprising that a lot of Math and Physics graduate students have Chinese, Indian or Russian surnames. The same is true about programmers too.
It is an irk, but an understandable and forgivable one.
If you look at the Intel platform, they change the packaging too, so that you don't have an upgrade path.
US-I and US-II were in the same packaging since 1995; IIRC, Sun changes the packaging in every odd-numbered version of their chip. This time they needed it in order to increase the bandwidth.
There is a company, called HAL. They belong to Fujitsu. They make UltraSPARC-compatible processors and servers. I've seen somewhere on their site that their next generation of chips will be PIN-compatible with the US-II, so they might become a source of upgrade.
According to the SPEC numbers,
900Mhz US-III is about the same as 750Mhz Alpha-21264A or 1Ghz Pentium-III with Rambus in OREGON-840 motherboard on Integer point.
On floating point, 900Mhz US-III is about 90% of the 750Mhz Alpha-21264A and 130% of the 1Ghz Pentium-III with Rambus in OREGON-840 motherboard.
WARNING: I used Base numbers; if using the Peak ones, Intel bites the dust even more.
Palm is Apple, i.e. successful proprietary product that became a segment-buster. The company uses superior software to sell more expensive hardware.
As this, it has attracted the attention of the big boys. They have decided to enter the segment. Every Apple has its moment of truth, and this is the moment when they still have a choice of becoming Microsoft. It means becoming mostly software company at the expense of the ability to move overpriced hardware;-)
As for the Palm, they were given more breathing room because WinCE devices were too expensive; Microsoft could not use its main marketing method that is dumping.
Don't forget about the tuition. Foreign students pay full price, and they are not eligible to education loans that are taken as granted by the Americans.
They are the richest people in the world because of the Wall Street; their money are the paper ones. If Micro$oft to have a bad quarter, Billy's worth will be decimated.
... they will probably put these inexpensive UltraSPARCs into the next generation Cubes. No one cares what processor is in, especially in the Linux world where you can compile anything for the target architecture.
As for the performance, it will be pretty good. These boxes are not processor-bound anyway.
I also think they will keep Linux as the OS: can anyone spell UltraPenguin ?;-) Another thing is that they will probably release a closed-source C/C++ compiler for Linux on SPARC for free, the way DEC did.
Re:Not Economics, Accounting...
on
Me-Commerce
·
· Score: 1
Lucky Bunny (sic), having 3-4 months off;-)
Don't you have problems explaining the hiring managers why you have these gaps ???
As for making permanent employment worthwhile, they've invented stock options.
Contracting IS a career choice; you make a decision to get more money, but not to become rich when the stock jumps over the roof. Another good thing for contractors is being paid for the overtime;-)))
Re:Not Economics, Accounting...
on
Me-Commerce
·
· Score: 1
Also, 60K/year will bring you an entry-level permanent specialist whether 240K/year contractor will be a top-notch one. His productivity might be (and will be) much higher.
OTOH, you won't be able to squeeze more than 40 hrs/week from the contractor without paying overtime. Even though, the overtimes might not be necessary with a good specialist.
Slashdot Mirror
This would not be such a problem if they would not be clueless about what my job is since it was my responsibility for making the deadlines; the programmer is the last in the development chain.
Another thing is that in such a small team (1 analyst, 1 data architect, 1 programmer + 1 clueless PHB) I become a quality control for them since I'm their customer.
... are your comments posted at -1 ???
You forget something.
;) or Starfire on the back end, but it won't make your badly designed Oracle queries faster; it will just raise the amount of them that can be processed on one server.
If the cost would be only labour vs. hardware, you'd be right. But it also comes to the:
1) Lost productivity on the user's part.
You can have a Beowulf of these
And don't remind me about all of these front-end application crashes. Long wait and crashes make users frustrated and therefore less efficient, less motivated and the job suffers.
2) Lost business opportunities.
Related to 1). If my calculation of the build plan works 10 minutes instead of 4 hours, users can play with it and adjust the plan to find the best one instead of just guessing what is the best because the computer system can't process more than one of them a day.
3) Maintainability. Your deadlines might be made, but if the program is a nightmare from that point of view, you're going to pay later (it goes from different accounting bucket, so they don't think about it as much as they should).
4) Dilution of the profession. Sometimes you MUST have the best of the breed (NASA, FAA, embedded systems), but you can't find them easy. Even harder is distinguishing between them and VB lamers.
And this is in the country, where you can't practice most of the professions unless you're a member of the guild or union!
I've misunderstood you then.
;-)
Uspehov!
5 years as a programmer (mostly consultant) in the corporate world gave me the following insight:
Everyone's slogan is "Acceptable is enough."
This is why programming ceased to be an art and is now a simple combination of "requirements gathering, design, coding, testing". In one company they even put a wall, making different people do requirements gathering, database design and programming, often excluding actual developer (me) from the making the decisions on the early stages of the project, and holding me responsible for implementing their stupid decisions, based on the lack of understanding of what programmers do!
My rant has the following relation to the quote: in 60-th the cost of a computer was enormous comparing to a programmer's salary. Computers were not entirely pathetic, but any significant system should have been designed and implemented not just well but perfectly; otherwise, computers would not be able to handle it.
In the modern world it is cheaper to buy another server than to hire another couple of programmers; they just throw more iron into a problem to compensate for the lack of quality and efficiency.
So, the problem lies with the clueless managers and executives.
Plus, try thinking about the image of a scientist in the popular culture.
;-) if not evil maniac Fu Manchu.
Scientist is Einstein who cannot make a step in the real life without help and forgets everything unrelated to the science (and does not know that Pepsi is better
With such an attitude it is not surprising that a lot of Math and Physics graduate students have Chinese, Indian or Russian surnames. The same is true about programmers too.
Have you heard about Fiancee Visa (K-1, IIRC) ???
e +visa&kl=XX&pg=q&Translate=on
You stay in the US, she comes on that visa and you get married here.
http://www.infodomain.com/american.htm
http://www.altavista.com/cgi-bin/query?q=Fiance
... publish patent applications on Slashdot. This will get a lot of stupid patents out of the gene pool ;-)
They have a major issue with Sun, and they have even better relationship with RedHat than IBM.
Actually, a bus cannot move faster than traffic. And it has a lot of stops! ;-)
... moderated it down ???
If you look at the TPC or similar test, you'll see that most of the cost for the setup is the storage, followed by the cost of software.
;-)
Actually, workstation performance depends on the floating point whether server one depends on the throughput and bandwidth.
So, if you want to edit documents in Word, get an Intel, otherwise get something that solves the problem the best.
BTW, Buick sucks, and Lexus sucks too (I have Nissan 300zx Turbo
Scott McNealy told once that the reason he does not want NT on Sparc is that Sun does not want to become a low-margin assembler of the boxes.
;-)
Also, M$ would never want to sell anything that can give Sun a dime of revenue
It is an irk, but an understandable and forgivable one.
If you look at the Intel platform, they change the packaging too, so that you don't have an upgrade path.
US-I and US-II were in the same packaging since 1995; IIRC, Sun changes the packaging in every odd-numbered version of their chip. This time they needed it in order to increase the bandwidth.
There is a company, called HAL. They belong to Fujitsu. They make UltraSPARC-compatible processors and servers. I've seen somewhere on their site that their next generation of chips will be PIN-compatible with the US-II, so they might become a source of upgrade.
http://www.hal.com/
http://mpd.hal.com/
According to the SPEC numbers,
900Mhz US-III is about the same as 750Mhz Alpha-21264A or 1Ghz Pentium-III with Rambus in OREGON-840 motherboard on Integer point.
On floating point, 900Mhz US-III is about 90% of the 750Mhz Alpha-21264A and 130% of the 1Ghz Pentium-III with Rambus in OREGON-840 motherboard.
WARNING: I used Base numbers; if using the Peak ones, Intel bites the dust even more.
His ego will meet its end then ;-)
Palm is Apple, i.e. successful proprietary product that became a segment-buster. The company uses superior software to sell more expensive hardware.
;-)
As this, it has attracted the attention of the big boys. They have decided to enter the segment. Every Apple has its moment of truth, and this is the moment when they still have a choice of becoming Microsoft. It means becoming mostly software company at the expense of the ability to move overpriced hardware
As for the Palm, they were given more breathing room because WinCE devices were too expensive; Microsoft could not use its main marketing method that is dumping.
Don't forget about the tuition. Foreign students pay full price, and they are not eligible to education loans that are taken as granted by the Americans.
What planet are you on? Even the fortune 500 richest people is flooded with programmers, something NEVER seen in earlier years
;-)
They are not just programmers; they are PROPRIETORS of the BUSINESSES they have founded.
The only exception is probably Bill Gates who secretly does Visual Basic coding in the nights
They are the richest people in the world because of the Wall Street; their money are the paper ones. If Micro$oft to have a bad quarter, Billy's worth will be decimated.
... are the companies thinking long-term in this country if they are judged by the beast of Wall Street on the quarterly basis ??? :-(
... they will probably put these inexpensive UltraSPARCs into the next generation Cubes. No one cares what processor is in, especially in the Linux world where you can compile anything for the target architecture.
;-) Another thing is that they will probably release a closed-source C/C++ compiler for Linux on SPARC for free, the way DEC did.
As for the performance, it will be pretty good. These boxes are not processor-bound anyway.
I also think they will keep Linux as the OS: can anyone spell UltraPenguin ?
Lucky Bunny (sic), having 3-4 months off ;-)
;-)))
Don't you have problems explaining the hiring managers why you have these gaps ???
As for making permanent employment worthwhile, they've invented stock options.
Contracting IS a career choice; you make a decision to get more money, but not to become rich when the stock jumps over the roof. Another good thing for contractors is being paid for the overtime
Also, 60K/year will bring you an entry-level permanent specialist whether 240K/year contractor will be a top-notch one. His productivity might be (and will be) much higher.
OTOH, you won't be able to squeeze more than 40 hrs/week from the contractor without paying overtime. Even though, the overtimes might not be necessary with a good specialist.