I agree its not a big deal for n=100, that's why I mentioned test data sets. On the other hand n=10m is not uncommon in business programming. My working definition of "system's programming" is "when the constant term matters".
The "show me something you wrote independently" tests for:
a) Experience working on your own thing in isolation and a desire to do so. b) A genuine love of programming, seeing it as trying to get a job in your passion. c) The ability to use easy tools fluidly.
Now those are good characteristics for a start-up which is who the article is written for. But all of those are negatives in many enterprise jobs.
Isolated opinionated programmers are a definite determent in enterprises quite often. You want enthusiasm but not passion in most workplaces. You often don't want to test for easy tools, but the ability to use hard tools. Complex applications are orders of magnitude more confusing than simple android applets.
And finally the Microsoft brain teaser type problems are basically a computer IQ test. They are testing for: i) Do you know basic computer science ii) Are you smart.
You can fix skills deficits in employees. Generally you can't fix (i) or (ii), though with younger programers you can sometimes fix (i). You will fight those problems everyday forever. Quite often in programming you can construct two algorithms to solve a problem with times like: n^2 + 25n + 100, 1000n + 20000. If you hire the Android guy you often get the n^2 solution since it works so much better on test data sets.
Exactly, this is a political strategy that makes a lot of sense. The industry is unified behind reducing or eliminating software patents. All that's needed now is congressional action.
What broke the protection rackets was Robert Kennedy's war on the mob. Hopefully CEOs openly admitting they are paying a protection racket will get the Robert Kennedys involved.
I don't know about that. Once CEOs of computer companies are willing to say publicly that entities called "patent trolls" exist and that they are being paid off, essentially a tax congress or the courts may take action. Microsoft has had it with these guys and is doing the heavy lifting in the fight, I'm glad RedHat is helping.
I think you might be surprised how much the 48S could do. I had Mathematica 1.3 on my 386-40 with math coprocessor and the 28S (a step down from the 48). Except for symbolic manipulation and excluding speed (which was something like a factor of 1000x) the 48 could keep up. And in terms of ease of use and faster the 28 ran rings around Mathematica.
Good point aout Agilent and scientific equipment possibly. What made HP of the 1970s and 80s unique was they were true innovators.
Take for example RPL which was the programming language, this was a LISP. Which meant that on my 28s I was able to combine things in totally unique ways. For example I wrote a program to do contour integration: i.e. path integrals in complex spaces that evaluate to complex not real numbers. And it wasn't hard. I could effectively reprogram this calculator to do what I wanted.
Things like the innovations of the Alpha Chip, yes the innovations of VMS. I'm not mourning the passing of 1970s technology I'm mourning the technologies we don't have because HP isn't that company anymore. Apple is a real innovator imagine if we had another 5 Apple computers running around trying to totally change the world.
That's an algebraic equation. Obviously algebraic notation is better for expressing algebraic relationships. RPN is better for numerical methods.
I.E. RPN for arithmetic, vector calculation, matrix algebra....
Where RPN though would kick ass is '2+x-4=3*x' [STEQ] [shift x] would activate a solve for x function. You can see the whole equation just becomes a single argument in the RPN calculator which is jumping up a level of abstraction.
Everyone said it. She gutted HP. HP was an engineering company when she got there that made great devices in many areas. I can understand going where the money was, but I wish they had sold off the engineering divisions to someone who was interested in those more niche markets.
If only 20 years you got it as a hand me down. By the time you got it, it probably wasn't worth it. The 15C is nowhere near the 48's which is what was current then.
Spreadsheets have been getting worse computationally for almost 20 years. Algebra systems I haven't played with much since 2000 but they have definitely gotten better. And TI has finally brought an algebra system calculator combo out after all this time.
You would think given that calculators still sell pretty well and this one is doing good for 30 full years that HP would maybe consider that they made a mistake in essentially killing off this line. Wouldn't it be wonderful it HP put out hand device for engineers as far advanced a the HPs were then?
I still see a lot of DRM on Apple. As for video I have to plead ignorance. I do rentals to Tivo from Amazon and given that its a rental it has to have DRM.
If they aren't receiving a dime they aren't professional they aren't at the high end of amateur. What you are describing is a "creative hobbyist" market. I have no idea if that's even a distinct market but I've never heard of them acting like a distinct buying group.
That's not necessarily a problem. Apple doesn't have copyright and thus can't enforce that provision. If I explicitly waive that provision from Apple no GPL violation occurred.
Think about this situation. A writes a GPL app B distributes A's application C downloads from B. D tells C that he must get a giraffe to legally use A's application
And if the system has trouble resolving a dependency then what? What about if it triggers a long series of upgrades? What if that dependency forces an upgrade on another application that breaks it?
Linux apps handle this sort of thing well because they grow up in an environment of constantly changing libraries.
Slashdot Mirror
I agree its not a big deal for n=100, that's why I mentioned test data sets. On the other hand n=10m is not uncommon in business programming. My working definition of "system's programming" is "when the constant term matters".
The "show me something you wrote independently" tests for:
a) Experience working on your own thing in isolation and a desire to do so.
b) A genuine love of programming, seeing it as trying to get a job in your passion.
c) The ability to use easy tools fluidly.
Now those are good characteristics for a start-up which is who the article is written for. But all of those are negatives in many enterprise jobs.
Isolated opinionated programmers are a definite determent in enterprises quite often. You want enthusiasm but not passion in most workplaces. You often don't want to test for easy tools, but the ability to use hard tools. Complex applications are orders of magnitude more confusing than simple android applets.
And finally the Microsoft brain teaser type problems are basically a computer IQ test. They are testing for:
i) Do you know basic computer science
ii) Are you smart.
You can fix skills deficits in employees. Generally you can't fix (i) or (ii), though with younger programers you can sometimes fix (i). You will fight those problems everyday forever. Quite often in programming you can construct two algorithms to solve a problem with times like: n^2 + 25n + 100, 1000n + 20000. If you hire the Android guy you often get the n^2 solution since it works so much better on test data sets.
You are right I misread the original statement.
And after that correction I agree with your 4 statements as originally written.
Nope the first 2 equations aren't the same. Shifting to algebraic notation:
3*((4+5)+(6+7)) != ((4+5)+(6+7)) * 3.
You can't flip 3 like that for arbitrary matrices.
This equality is fine since you are commuting just the additions.
4 5 + 6 7 + * 3 + = X
5 4 + 7 6 + * 3 + = X
And your 4th statement is wrong. Those 2 (the 2nd and 3rd equation) are both equal to the 4th:
7 6 + 5 4 + * 3 + = 5 4 + 7 6 + * 3 +
You should tell that to my copy which still has lots of "Protected AAC audio file". Download the song of the week from iTunes store.
Exactly, this is a political strategy that makes a lot of sense. The industry is unified behind reducing or eliminating software patents. All that's needed now is congressional action.
What broke the protection rackets was Robert Kennedy's war on the mob. Hopefully CEOs openly admitting they are paying a protection racket will get the Robert Kennedys involved.
I don't know about that. Once CEOs of computer companies are willing to say publicly that entities called "patent trolls" exist and that they are being paid off, essentially a tax congress or the courts may take action. Microsoft has had it with these guys and is doing the heavy lifting in the fight, I'm glad RedHat is helping.
I think you might be surprised how much the 48S could do. I had Mathematica 1.3 on my 386-40 with math coprocessor and the 28S (a step down from the 48). Except for symbolic manipulation and excluding speed (which was something like a factor of 1000x) the 48 could keep up. And in terms of ease of use and faster the 28 ran rings around Mathematica.
Yep HP calculators are worth a lot. They have small group of dedicated users.
Good point aout Agilent and scientific equipment possibly. What made HP of the 1970s and 80s unique was they were true innovators.
Take for example RPL which was the programming language, this was a LISP. Which meant that on my 28s I was able to combine things in totally unique ways. For example I wrote a program to do contour integration: i.e. path integrals in complex spaces that evaluate to complex not real numbers. And it wasn't hard. I could effectively reprogram this calculator to do what I wanted.
Things like the innovations of the Alpha Chip, yes the innovations of VMS. I'm not mourning the passing of 1970s technology I'm mourning the technologies we don't have because HP isn't that company anymore. Apple is a real innovator imagine if we had another 5 Apple computers running around trying to totally change the world.
You aren't just adding you are also multiplying.
Don't forget chemistry. They had terrific tools for chemists.
That's an algebraic equation. Obviously algebraic notation is better for expressing algebraic relationships. RPN is better for numerical methods.
I.E. RPN for arithmetic, vector calculation, matrix algebra....
Where RPN though would kick ass is '2+x-4=3*x' [STEQ] [shift x]
would activate a solve for x function. You can see the whole equation just becomes a single argument in the RPN calculator which is jumping up a level of abstraction.
Everyone said it. She gutted HP. HP was an engineering company when she got there that made great devices in many areas. I can understand going where the money was, but I wish they had sold off the engineering divisions to someone who was interested in those more niche markets.
I still pretty regularly use Pcalc. I also use the command line "e" for that sort of mini calculation. If I want to step up GHCI is pretty awesome.
You are using the commutative law there. Replace the numbers with matrices (non commuting) and you would get 2 different answers.
If only 20 years you got it as a hand me down. By the time you got it, it probably wasn't worth it. The 15C is nowhere near the 48's which is what was current then.
Spreadsheets have been getting worse computationally for almost 20 years. Algebra systems I haven't played with much since 2000 but they have definitely gotten better. And TI has finally brought an algebra system calculator combo out after all this time.
You would think given that calculators still sell pretty well and this one is doing good for 30 full years that HP would maybe consider that they made a mistake in essentially killing off this line. Wouldn't it be wonderful it HP put out hand device for engineers as far advanced a the HPs were then?
Anyway the scientific version of the 12c is the 15c: http://www.hpmuseum.org/hp15.htm
And my love was the 28S. http://www.hpmuseum.org/hp28c.htm
I still see a lot of DRM on Apple. As for video I have to plead ignorance. I do rentals to Tivo from Amazon and given that its a rental it has to have DRM.
If they aren't receiving a dime they aren't professional they aren't at the high end of amateur. What you are describing is a "creative hobbyist" market. I have no idea if that's even a distinct market but I've never heard of them acting like a distinct buying group.
That's not necessarily a problem. Apple doesn't have copyright and thus can't enforce that provision. If I explicitly waive that provision from Apple no GPL violation occurred.
Think about this situation.
A writes a GPL app
B distributes A's application
C downloads from B.
D tells C that he must get a giraffe to legally use A's application
D's act doesn't cause B to violate the GPL.
And if the system has trouble resolving a dependency then what? What about if it triggers a long series of upgrades? What if that dependency forces an upgrade on another application that breaks it?
Linux apps handle this sort of thing well because they grow up in an environment of constantly changing libraries.
Do they have GUIs yes. Finkcommander darwinports GUI... But you have to know what you are doing to use Fink or DarwinPorts.
Homebrew OTOH would work well for an easy to use repository.