You don't put training wheels on a 1100cc Honda do you?
Re:Am I the only person who is hesitent about this
on
.NETly News
·
· Score: 1
Ever so true...
But a good bot for me to have would be one that I can instruct to go to work for me.
I say: "go to work and earn me $85,000 per year after taxes" (i'm not enormously greedy) and the thing says "okey dokey" and money just falls into my bank account.
Until that happens, Computers will just be something else to pay for and maintain.
I guarantee you I'm gonna be at the local watering hole meeting people the old-fashioned way.
Absolutely, and that's why I vote with the dollar bill every day.
Think of it: Corporations with money are the ones that pay lobbyists, who in turn pressure politicians into doing what the CORPS want regardless of their campain promises.
Much more effective way to grease the wheel that squeaks than vote one every 4 years for a bunch of flag-waving, baby-kissing liars.
Congratulations. you did exactly what the old resume (with no accent even though I'm French and I should know better) is supposed to do: You modified it slightly to match the particular ad, but you sent a lot of them.
But did you send 20,000 exactly alike? No. So you didn't SPAM.
I want less infrastructure (less taxes) not more. I want fuel-efficient vehicles (hello Honda!) not electricity-driven public transportation systems (burn petroleum in central station, then carry the electricity 20+km to train), I want freedom to go to Vegas at 4:22 AM, when I FEEL like it, not because it's convenient on the bureaucracy-mandated "she-dule" (skedual for US speakers).
I want convenience, and low cost... In UK, gasoline is out of this work, but in US, it's not, so no public transportation for me.
What? our freeways are congested? So what. telecommute, move minutes from your job. there are many more, better alternatives. But the fact is, you have the choice. With Public transportation and high taxes, you lose the choice, becuase you're GOING to get taxed, so you might as well use it, even if it's less convenient.
In the article, someone--am not sure who, I read it last night, slept since--mentioned that building software for a client is like building a building for a client.
How would the architect like it if the client kept calling and saying things like that:
--I've decided to add a moat.
--Can we build in in France for the same cost?
--I've decided to add 30 stories... How long will that take?
--Each office suite must have a view of the Ocean.
--I don't like glass on the outside, let's go with granite.
--I don't like granite for the outside, let's go with marble.
--I don't like marble for the outside, let's go back to glass.
--Why are you charging me so much?
--Why is the design of the project taking so long?
Bang Bang Bang...
My point is:
I don't write a single line of code until the client has told me, specifically, what he wants the program to do.
If the client is unable to formulate that in the form a concise and well-written spec, then I write the spec for them (in English, not pseudocode), and I bill them for that.
If they don't know what they want to software to do to begin with, I do an analysis of their business process, with flowcharts, goals, SWAT and all. Then I write the proposal for the software to streamline their operation and allow them to leverage their competitive advantages. And I bill them for that.
I tell them up front that that is what I'm going to do. If they don't like it or they don't want to pay for it, I say: Call me back when you do.
By the way: it really helps to have taken upper division business courses, and have a healthy (not biased) dose of knowledge about the current business world (Wall Street Journal or the like).
That way, when your client balks at the cost, time, and effort involved, you say: Are you willing to lose any market share to (insert their competitor's name)?
Anyway. Writing software is not really about coding, it's about what you can do to make businesses more productive and competitive. If you can do that, it doesn't matter if you use COBOL or C++ or C# or Java or what.
Extreme Programming (the name's gotta go) and Interaction Design both use different methods to address the lack of proper business analysis and requirements.
The same goes for human languages. The language itself influences the thoughts of the author, so that a particular expression in one language can never be quite the same in another.
Can't argue with you on that... But the question begs again: Why is that important for a computer job?
A good carpenter needs to know carpentry, period. He does not need to be able to recite german poetry in german.
It might make him a better human being, granted, but that seems completely unrelated to the task.
When there is a meeting, and 6 people in expensive suits ask point-blank: How long will it take to implement this? They want an answer in number of days, not a philosophy discussions on the merits of this or that.
The fact is there is no acceptable alternative to college/university education. It means nothing that a man may be self-taught. It used to in this country.
You know, it comes across sometimes like this: I don't have a piece of paper that proves I can speak English. You can, because you're listening to me, realize that I do not need it, because I have acquired the skill some other way. Do I need to go to school to get a piece of paper that says I can speak english? As a frenchman, I should. Likewise, I just recently took my first official computer science class (a waste of time if you want to know the truth), yet I can program in half a dozen languages, and write it the corporate way: specs, docs, testing, etc...
Do I need a piece of paper that says I can program? In america, I do...
Why?
I'll tell you why... Because there's money involved. University professors and deans and such stiffs do not want an alternative to develop. They have a lucrative monopoly on education in america, and they're not about to let a few smart and self-motivated people ruin their plans. The idea that a man today cannot be well-educated without going to college has become ingrained into the corporate psyche to a degree that precludes self-taught people from being recognized as contributing members of society.
BTW, I understand about culture and exposure to different ways of thinking. I can tell you that immigrants with HS diplomas have better understanding of that than 3rd generation american citizens who graduate from prestigious universities. It's called living in a foreign country for more than 6 months. It's something that few americans ever experience, college or not.
Same here: Told my teacher I didn't want to take introduction to C++ in order to get into Introduction to Java (it was a requirement--she turned down 3 students before because they didn't meet the prerequisite).
I told her I worked at a bank as a DBA, and did custom programming in ASP, VB, XML, XMLRPC, XHTML, ADO, SQLDMO, SQL, OO, COM, and a few other fancy acronyms that went way over her head, with all documentation and all that, and she said, you'll do fine. I took the java class and aced it (which isn't to say I love java now, I don't, it's horrible--although, compared to ASP it's almost elegant).
But the result is that I saved a semester.
Now, all the other stupid classes is what sucks (like who cares about the phases of the moon or the composition of rocks).
I think also that the foreign language requirement should be waved, since programming languages are foreign languages, and that, like pilots, english IS the language of computer science (ask the french or Japanese--if they can't read english, they can't read the comments in code... Forget translations...)
Thinking of that, why don't you set up something with the vendor where you will send them (fedex, ups) an encrypted CDrom with the info. That way, if the CD is lost, you don't care, and it doesn't go through the internet at all.
But the better solution is to get together with your bank (you are a normal business wuth a commerical account, of course) and get their solution on board.
Besides, it would be good if some of that code was actually being reused and refitted from older progs. Thus, you could do 1000 lines of code in an afternoon.
5000 lines of code per week, counting 60 hours (being generous), is 83 lines per hour, or 1.4 lines per minutes. assuming 40 bytes per line (including the spaces (no tabs--of course)) that's 55 bytes per minutes. Assuming 7 bytes per word, that's 8 words per minutes.
I'm assuming that the aveerage programmer can manage to type that fast.
The reality is that the programmer needs to be highly skilled, highly motivated, and have a very good understanding of what he/she is trying to achieve.
But it is possible. Also, the programmer may need to be single and independently wealthy (living at parent's or have exercised interesting stock options in the summer of 2000)
I have a saying: There should be 5 lines of comment for each line of code. (i'm including the lines of ************ in there too though)
Each program should have a complete changelog, with dates and emails and www pages of authors, a complete breakdown of all functions, at the top, in english, not a mix of english and pseudocode.
Also, a second document should exist, in html (or whatever) that shows the purpose of the program, the strengths and shortcomings of the program, the runtime requirements (dependencies, OS versions, hardware, memory, HDD access permissions, etc), and any outstanding issues, requests for change, etc.
Also, there should be a third document, created before, outlining in english all the expectations of the program to be written. Specs, in fact. This can be html or text, as the programming team prefers.
I program in ASP, so it's very easy to spaggettize. I have learned to enforce iron discipline. Because there is nothing like trying to read through 900 lines of html-producing code for that one thing management wants to change (I call those manabugs--things management considers bugs)).
I work at a bank. We have mean excel workbooks, with 20+ worksheets, dozens of vba scripts, links to other workbooks, and OpenOffice6 just *refuses* to open them (illegal operation errors).
They work in Excel 97, Excel 2000 and Excel XP. They don't work in StarOffice6 (granted, Beta). There's no compatibility.
And I guarantee you something, everyone in accounting will have my head on a shiny platter if I cause last year's management reports to be in an incompatible format...
So, I'm not even going to mention it.
On the other hand, I'm working on a neato little program to xml-rpc excel spreadsheet data to databases... more on that later.
The book publishing industry has generally capitalized on that, by publishing well-written and well-edited works (YMMV) and selling them at $40+
The problem is that you have to own the book in order to get useful information out of it, and since most people in IT work on a wide variety of software and hardware, you end up having to buy the equivalent of a law library just so you can have the info at your fingertip.
Print 100,000 books, 100,000 people can get info. Publish 1 web page, 1,000,000 people can get info (barring/. effect OC).
Since arguably it is more expensive to document and explain software well than to code the software, the documenters/manual writers should get money. I think a lot more people code for the love of the craft than document for the love of the craft. I may be wrong, but that's my opinion.
How does one allocate money for manual and documentation? By billing companies.
If there was a project to document the whole of open-source software, including tricks and shortcuts, troubleshooting, complete with screenshots and pdf manuals (sort of like PostGreSQL docs), and companies would... Arrgghhh I realized: companies would never do that... Programmers have to fund it themselves.
Dangit!
Why would companies not pay for it? Budgets.
In budget, you allocate money for service contracts, not voluntary payments to the community.
Besides, everything a company pays for is either invoiced or throught employee reimbursments.
Companies don't pay invoices without service contracts, and they don't reimburse employees for that kind of expenses (imagine the MSCE-Linux-wannabe telling his boss that he's sending $400 bux to an open-source consortium for a linux documentation project--he'll be lucky if he gets his next paycheck).
I just thought of something... if there was a gunmaker that made a gun with a bug, and the bug was that if the gun is shot straight up in the air (not often but it happens), the bullet would actually come out of the back of the gun and kill the shooter.
Now, imagine the same gunmaker telling people that there's nothign wrong with their guns, that the fault is with the people telling the shooters to shoot straight up in the air...
True, these people would be malicious, and EVIL in the Bush sense, but they are utilizing a flaw in the gun that shouldn't be there to begin with.
I would bet that only the incompetent and ignorant would buy that sort of gun (not that incompetent and ignorant people should be buying guns, mind you) and then shoot straight up in the air when the l33t says to.
This is to say that both the virus writer and the software writer are to blame (in criminal court with fines and jail time), one for exploiting a flaw, the other for providing the means to.
There should be a government mandated warning on all microsoft software, and all computers with microsoft software preistalled that reads, in 72 point red letters "Consumer Warning: This software can be used by criminals to steal and destroy your personal information. Exercise caution."
That way, people who got a virus would blame microsoft ALONG with the virus writer.
==
The reality is that nothing like that will ever happen. (sig-in-training)
so how many people do you figure it takes to organize this sort of thing?
And since companies would be willing to pay for support, then the authors of articles/resources should also be paid...
how different is that from linuxdoc.org?
So then there should be efforts somewhere on writing really good documentation... Because in the end, I believe that any software that is extremely well documented will have fewwer problems and more users than those who aren't.
Slashdot Mirror
You don't put training wheels on a 1100cc Honda do you?
Ever so true...
But a good bot for me to have would be one that I can instruct to go to work for me.
I say: "go to work and earn me $85,000 per year after taxes" (i'm not enormously greedy) and the thing says "okey dokey" and money just falls into my bank account.
Until that happens, Computers will just be something else to pay for and maintain.
I guarantee you I'm gonna be at the local watering hole meeting people the old-fashioned way.
VB was my first language, and I do like it, as that. It's a bike with training wheels. That's it.
Laptops.
And laptop manufacturers need to have the option of turning off the monitor (maybe a BUTTON somewhere) without affecting the processes on the machine.
Absolutely, and that's why I vote with the dollar bill every day.
Think of it: Corporations with money are the ones that pay lobbyists, who in turn pressure politicians into doing what the CORPS want regardless of their campain promises.
Much more effective way to grease the wheel that squeaks than vote one every 4 years for a bunch of flag-waving, baby-kissing liars.
As far as the Aussies: Way to ding those Japs.
[note to self: check sarcasm at the door]
Congratulations. you did exactly what the old resume (with no accent even though I'm French and I should know better) is supposed to do: You modified it slightly to match the particular ad, but you sent a lot of them.
But did you send 20,000 exactly alike? No. So you didn't SPAM.
How much you want to bet that if they don't turn a profit, our taxes will bail them out?
I want less infrastructure (less taxes) not more. I want fuel-efficient vehicles (hello Honda!) not electricity-driven public transportation systems (burn petroleum in central station, then carry the electricity 20+km to train), I want freedom to go to Vegas at 4:22 AM, when I FEEL like it, not because it's convenient on the bureaucracy-mandated "she-dule" (skedual for US speakers).
I want convenience, and low cost... In UK, gasoline is out of this work, but in US, it's not, so no public transportation for me.
What? our freeways are congested? So what. telecommute, move minutes from your job. there are many more, better alternatives. But the fact is, you have the choice. With Public transportation and high taxes, you lose the choice, becuase you're GOING to get taxed, so you might as well use it, even if it's less convenient.
I hear you.
The important thing is for people to realize that the more fries you want, the more the bill is going to be. There are no free refills.
In the article, someone--am not sure who, I read it last night, slept since--mentioned that building software for a client is like building a building for a client.
How would the architect like it if the client kept calling and saying things like that:
--I've decided to add a moat.
--Can we build in in France for the same cost?
--I've decided to add 30 stories... How long will that take?
--Each office suite must have a view of the Ocean.
--I don't like glass on the outside, let's go with granite.
--I don't like granite for the outside, let's go with marble.
--I don't like marble for the outside, let's go back to glass.
--Why are you charging me so much?
--Why is the design of the project taking so long?
Bang Bang Bang...
My point is:
I don't write a single line of code until the client has told me, specifically, what he wants the program to do.
If the client is unable to formulate that in the form a concise and well-written spec, then I write the spec for them (in English, not pseudocode), and I bill them for that.
If they don't know what they want to software to do to begin with, I do an analysis of their business process, with flowcharts, goals, SWAT and all. Then I write the proposal for the software to streamline their operation and allow them to leverage their competitive advantages. And I bill them for that.
I tell them up front that that is what I'm going to do. If they don't like it or they don't want to pay for it, I say: Call me back when you do.
By the way: it really helps to have taken upper division business courses, and have a healthy (not biased) dose of knowledge about the current business world (Wall Street Journal or the like).
That way, when your client balks at the cost, time, and effort involved, you say: Are you willing to lose any market share to (insert their competitor's name)?
Anyway. Writing software is not really about coding, it's about what you can do to make businesses more productive and competitive. If you can do that, it doesn't matter if you use COBOL or C++ or C# or Java or what.
Extreme Programming (the name's gotta go) and Interaction Design both use different methods to address the lack of proper business analysis and requirements.
Absolutely.
The same goes for human languages. The language itself influences the thoughts of the author, so that a particular expression in one language can never be quite the same in another.
Can't argue with you on that... But the question begs again: Why is that important for a computer job?
A good carpenter needs to know carpentry, period. He does not need to be able to recite german poetry in german.
It might make him a better human being, granted, but that seems completely unrelated to the task.
When there is a meeting, and 6 people in expensive suits ask point-blank: How long will it take to implement this? They want an answer in number of days, not a philosophy discussions on the merits of this or that.
The fact is there is no acceptable alternative to college/university education. It means nothing that a man may be self-taught. It used to in this country.
You know, it comes across sometimes like this: I don't have a piece of paper that proves I can speak English. You can, because you're listening to me, realize that I do not need it, because I have acquired the skill some other way. Do I need to go to school to get a piece of paper that says I can speak english? As a frenchman, I should. Likewise, I just recently took my first official computer science class (a waste of time if you want to know the truth), yet I can program in half a dozen languages, and write it the corporate way: specs, docs, testing, etc...
Do I need a piece of paper that says I can program? In america, I do...
Why?
I'll tell you why... Because there's money involved. University professors and deans and such stiffs do not want an alternative to develop. They have a lucrative monopoly on education in america, and they're not about to let a few smart and self-motivated people ruin their plans. The idea that a man today cannot be well-educated without going to college has become ingrained into the corporate psyche to a degree that precludes self-taught people from being recognized as contributing members of society.
BTW, I understand about culture and exposure to different ways of thinking. I can tell you that immigrants with HS diplomas have better understanding of that than 3rd generation american citizens who graduate from prestigious universities. It's called living in a foreign country for more than 6 months. It's something that few americans ever experience, college or not.
Same here: Told my teacher I didn't want to take introduction to C++ in order to get into Introduction to Java (it was a requirement--she turned down 3 students before because they didn't meet the prerequisite).
I told her I worked at a bank as a DBA, and did custom programming in ASP, VB, XML, XMLRPC, XHTML, ADO, SQLDMO, SQL, OO, COM, and a few other fancy acronyms that went way over her head, with all documentation and all that, and she said, you'll do fine. I took the java class and aced it (which isn't to say I love java now, I don't, it's horrible--although, compared to ASP it's almost elegant).
But the result is that I saved a semester.
Now, all the other stupid classes is what sucks (like who cares about the phases of the moon or the composition of rocks).
I think also that the foreign language requirement should be waved, since programming languages are foreign languages, and that, like pilots, english IS the language of computer science (ask the french or Japanese--if they can't read english, they can't read the comments in code... Forget translations...)
Anyway... I digress.
for sure...
:)
I put my productive level at 100 lines per day.
More, and there's a bug in there somewhere...
Less, and I was writing docs...
granted, but you *have* to be able to type to write programs (unless you have a very good typist or are still using punch cards...)
Thinking of that, why don't you set up something with the vendor where you will send them (fedex, ups) an encrypted CDrom with the info. That way, if the CD is lost, you don't care, and it doesn't go through the internet at all.
But the better solution is to get together with your bank (you are a normal business wuth a commerical account, of course) and get their solution on board.
ever heard of copy and paste?
Besides, it would be good if some of that code was actually being reused and refitted from older progs. Thus, you could do 1000 lines of code in an afternoon.
5000 lines of code per week, counting 60 hours (being generous), is 83 lines per hour, or 1.4 lines per minutes. assuming 40 bytes per line (including the spaces (no tabs--of course)) that's 55 bytes per minutes. Assuming 7 bytes per word, that's 8 words per minutes.
I'm assuming that the aveerage programmer can manage to type that fast.
The reality is that the programmer needs to be highly skilled, highly motivated, and have a very good understanding of what he/she is trying to achieve.
But it is possible. Also, the programmer may need to be single and independently wealthy (living at parent's or have exercised interesting stock options in the summer of 2000)
Ever so true...
I have a saying: There should be 5 lines of comment for each line of code. (i'm including the lines of ************ in there too though)
Each program should have a complete changelog, with dates and emails and www pages of authors, a complete breakdown of all functions, at the top, in english, not a mix of english and pseudocode.
Also, a second document should exist, in html (or whatever) that shows the purpose of the program, the strengths and shortcomings of the program, the runtime requirements (dependencies, OS versions, hardware, memory, HDD access permissions, etc), and any outstanding issues, requests for change, etc.
Also, there should be a third document, created before, outlining in english all the expectations of the program to be written. Specs, in fact. This can be html or text, as the programming team prefers.
I program in ASP, so it's very easy to spaggettize. I have learned to enforce iron discipline. Because there is nothing like trying to read through 900 lines of html-producing code for that one thing management wants to change (I call those manabugs--things management considers bugs)).
The money uses SDI. the fin analysis uses excel.
And please, i'm not trolling. this is real-world application.
I work at a bank. We have mean excel workbooks, with 20+ worksheets, dozens of vba scripts, links to other workbooks, and OpenOffice6 just *refuses* to open them (illegal operation errors).
They work in Excel 97, Excel 2000 and Excel XP. They don't work in StarOffice6 (granted, Beta). There's no compatibility.
And I guarantee you something, everyone in accounting will have my head on a shiny platter if I cause last year's management reports to be in an incompatible format...
So, I'm not even going to mention it.
On the other hand, I'm working on a neato little program to xml-rpc excel spreadsheet data to databases... more on that later.
Answering myself, the faux-pas of faux pas :)
/. effect OC).
The book publishing industry has generally capitalized on that, by publishing well-written and well-edited works (YMMV) and selling them at $40+
The problem is that you have to own the book in order to get useful information out of it, and since most people in IT work on a wide variety of software and hardware, you end up having to buy the equivalent of a law library just so you can have the info at your fingertip.
Print 100,000 books, 100,000 people can get info. Publish 1 web page, 1,000,000 people can get info (barring
Since arguably it is more expensive to document and explain software well than to code the software, the documenters/manual writers should get money. I think a lot more people code for the love of the craft than document for the love of the craft. I may be wrong, but that's my opinion.
How does one allocate money for manual and documentation? By billing companies.
If there was a project to document the whole of open-source software, including tricks and shortcuts, troubleshooting, complete with screenshots and pdf manuals (sort of like PostGreSQL docs), and companies would... Arrgghhh I realized: companies would never do that... Programmers have to fund it themselves.
Dangit!
Why would companies not pay for it? Budgets.
In budget, you allocate money for service contracts, not voluntary payments to the community.
Besides, everything a company pays for is either invoiced or throught employee reimbursments.
Companies don't pay invoices without service contracts, and they don't reimburse employees for that kind of expenses (imagine the MSCE-Linux-wannabe telling his boss that he's sending $400 bux to an open-source consortium for a linux documentation project--he'll be lucky if he gets his next paycheck).
How to fund?
I just thought of something... if there was a gunmaker that made a gun with a bug, and the bug was that if the gun is shot straight up in the air (not often but it happens), the bullet would actually come out of the back of the gun and kill the shooter.
Now, imagine the same gunmaker telling people that there's nothign wrong with their guns, that the fault is with the people telling the shooters to shoot straight up in the air...
True, these people would be malicious, and EVIL in the Bush sense, but they are utilizing a flaw in the gun that shouldn't be there to begin with.
I would bet that only the incompetent and ignorant would buy that sort of gun (not that incompetent and ignorant people should be buying guns, mind you) and then shoot straight up in the air when the l33t says to.
This is to say that both the virus writer and the software writer are to blame (in criminal court with fines and jail time), one for exploiting a flaw, the other for providing the means to.
There should be a government mandated warning on all microsoft software, and all computers with microsoft software preistalled that reads, in 72 point red letters "Consumer Warning: This software can be used by criminals to steal and destroy your personal information. Exercise caution."
That way, people who got a virus would blame microsoft ALONG with the virus writer.
==
The reality is that nothing like that will ever happen. (sig-in-training)
so how many people do you figure it takes to organize this sort of thing?
And since companies would be willing to pay for support, then the authors of articles/resources should also be paid...
how different is that from linuxdoc.org?
So then there should be efforts somewhere on writing really good documentation... Because in the end, I believe that any software that is extremely well documented will have fewwer problems and more users than those who aren't.
Who will audit you?