The problem is they don't play well together. I want a tool that uses the information already present in my programs to take out the tedious and error-prone stuff. Why do I have to tell my computer how to map my objects to tables? Why do I have to write so much support code to use CORBA? Why do I have to worry about include paths and library linkage and jar files? Why do I embed javascript in HTML, which is in turn embedded in a string in a script file?
Yeah, I hope they're already thinking about the personal version, because I've been dreaming of Google on my machine for a long time. An intelligent search beats just about any other kind of infrequent interaction: menus, directory navigation, dialog boxes with lots of little pages on them. I want to hit ctrl-g, type in what I'm interested in, and get the right thing.
The output was not in alphabetical order. The LaTeX version was huge - 1000's of pages for a large project, even after I messed with the options to slim it down. It didn't handle templates very well, either. There was a copy of the template's documentation for every instance of the template! I don't remember which version I used, and I only devoted a few days to tuning it, but overall it was not useful for our project.
It seems like a good idea, but I think typing and clicking little command bursts is faster than explaining to my computer what I want it to do. Maybe for searches, like 'find me a picture of a cow', or 'where was that document I was working on yesterday'?
First of all, pipes are stupid stupid stupid. A pipe is a degenerate function: f(bunch of bytes) -> bunch of different bytes. Why aren't we allowed to write (and document and eforce the syntax of) different functions? We can in our prgramming languages, just not our OS. And then people start yammering about CORBA, which is wayyyy to general in the other direction. I want my compiler document-and-enforce-syntax magic on all my applications.
And then why would there be any symmetry between text processing and graphical user interfaces? I mean, it's worthwhile to look, but I checked and there isn't any. Hook my 'hierarchy generator' up to my 'explorer interface to hierarchies' sounds like a winner until you realize you need type safety to prevent the whole thing blowing up, and then the expressive power of functions to make more useful stuff happen (and documentation, so the GUI can prompt you to fill in the occasional blank).
At this point I sound like I want to replace pipes with something a bit more strict, which is true. But not for GUIs. Oh no. I want TeX.
Why have I ever in my life had to position or resize a window? That's like manually pushing text around in a terminal. And then menus are a step in the right direction, but they hide too much information. Icons are a little helpful in my file browser, but then graphic artists run off and think I can understand what they mean by just looking at them.
I want a text formatter to lay out my entire computer. I want to control how many pages I see at a time, and be able to navigate between the pages. I want a help system that shows me what hotkeys are available right this second. Instead of menus where I get to hunt in little 30x60 text lists, I want to go to help mode and maybe get a little text search to find the page of the user interface that lets me set margins or whatever. I want documents, not applications. I want an object database, not files. I want interfaces with conformance tests, not standards.
Our development environment is twenty-five years old. Our GUI comes straight from the first 'well maybe we should make these things easy to use' thoughts a few (dozen) academic leaders had. We need some independent thinkers to try something different.
Oh- no mouse. There's GOT to be a way to send the computer commands with having to take my hands off of my 101-key keyboard. Buttons remind me what I can do; that doesn't mean I actually have to walk over there and push the things.
I called one of my local Clear Channel stations, and the lady told me she hadn't seen a list. And yeah, songs that mention jets and explosions might be a little innapropriate right now.
But meeting the requirements is a complicated goal, and well-designed software is an important part of the solution. I think well-designed software has a lot in common with more human designs, like music, and so to make well-designed software it has to be 'beautiful'.
A system's archtitecture is the programmer's user interface. With that in mind, the goal of the interface is the same as any other: be pleasant to use. We usually say 'maintainable', but it's the same goal. To help achieve that goal, software architecture should be beautiful, because a beautiful architecture is easy on the mind. A well-designed system is easier to remember and reason about.
Consistency and symmetry are the first two design guidlines that come to mind when people think about well-designed software, and of course software has to actually do what it was written to do. But beyond that, there are thousands of choices that depend on the engineer's aesthetics, and hundreds of choices where bending the rules gets better results. Classification is much more difficult that consistent variable names.
A lot of these 'beautiful displays of complex information' problems were solved by old-school drafting fifty years ago.
What does Moore's Law have to do with software? Moore's law just says that about every year, the number of transistors you can fit in a chip will double. He changed it in 1995, to say about every two years. 18 months to be exact. And he originally said this in 1965. And I do believe the software industry existed before microsoft entered the picture. See his bio here
Computers got faster, and that enabled better software. As opposed to software getting better with each version, which was the author's opinion.
Since the software industry as he paints it started with Microsoft in 1975, every program subject to his law straddles almost half of the industry's lifetime! Over this time period computers got much faster and more people bought computers, which might have more to do with sales figures than aging.
And to indulge in a little Microsoft bashing: products that last ten years usually suck compared to the competition, but come out earlier and have better marketing. Once they're the last ones standing, they finally turn their attention to making the product work like it says on the box. I think the reason it takes ten years to write good software is you have to spend the first 8 beating your competition.
Yeah, if you extend a GPL'ed product it has to be GPL'ed, but if you extend a non-GPL'ed product you get sued, right? And you can even use GPL'ed code in proprietary code under the LGPL. I think Bill is committing false dilemma, because you don't have to extend an GPL'ed project to use free software. Or maybe just bullshit?
There are developers and distributors. The developers must convince a distributor that they can develop a quality product on time, at which point the distributor will fund their development. The distributor pays money for successfully completed milestones, and possibly some money up front to begin development. I'm pretty sure (this wasn't my department) we used design docs and concept art to pitch our concepts to potential distributors.
There are also contracts floating around- distributors like Hasboro will have an idea that they want developed. This might be a good place to start. Either that, or be as good as Blizzard or Valve:)
Remember that the distributor could lose millions on your game through duplication, warehousing, shelf space, and marketing (which costs probably double the total development price). To these people, you are a lottery ticket!
So code is the formal expression of truth. When matematicians go looking for truth, they don't search for a convincing argument. Instead, they develop the truth (a theorem) through the application of very precise rules (axioms and production rules) into a proof.
The language of mathematical proofs is so well-defined that you can instruct a machine to process the instructions, eg carry out the proof. This makes a proof a program, so the boundary between absolute, universal truth and copyright-circumventing code is very blurry. In one respect, deCSS is the proof of the theorem that, for any CSS-encrypted byte sequence there is an unencrypted byte sequence. But sin(x)**2 + cos(x)**2 = 1 is also a program; can the DMCA outlaw that?
Now obviously there is a big difference between a program that, to extend the DA's example, blew up the whole world versus a program that calculates digits of PI. But as a developer, the fact that code can be outlawed at all makes me very scared and unhappy, and I would the like the boundary of evil code and good code to be well defined. And I would really like for the MPAA and RIAA to have nothing at all to do with the establishment of that boundary.
Go read Godel, Escher, Bach, it has lots of good stuff on formal systems.
I was in the same spot- going for my CS degree, learning everything I could about game development and programming (first book: Teach Yourself Game Programming in 21 Days. Not exactly a classic, but still near and dear to my heart). My girlfriend at the time moved from Atlanta to Greensboro to pursue her education. While she was out walking her dogs at the park, she started up a conversation with someone who happened to be dating the VP of development at a local game company, and they were hiring. My gf told me about her encounter when I came up to visit, I made the call, got the interview (in jeans and a t-shirt and no resume), and got the job. One week later, I left my crappy help desk position in Atlanta, loaded all of my stuff into my car, and went up north to write video games. It was a dream come true, and it all happened over a couple of days.
So now I'm a firm believer in personal networking, and of being ready for your chosen career even if your chosen career isn't ready for you.
On a more practical note, most of the designers there worked their way up from testing. On an even more practical note, I was laid off a year later and left the industry, but I could still go back with the experience on my resume.
By submitting bids that request PC systems without an
Operating System due to a Microsoft site license, you
can earn points and win!
Emphasis mine. They're going after people who are misusing a Microsoft license, not regular customers. Microsoft's email is a bit confusing, but no more so than slashdot. Thanks for keeping me on my toes. You guys are unresponsible, unaccountable morons. More predictable than a computer.
It's an annoying and weird problem. I'm not really keen on the idea of switching the method handler around for the special case of construction. You've really got two phases: phase one is the trad C++ make-my-memory-useful phase, and phase two is the 'stuff you always need to do with me every time I'm created' phase.
The subtle difference is that phase one gets an incomplete base class, while phase two gets the finished object, and hence virtual method calls in the base class get passed to the derived class handler.
Back when PC's sucked and mainframes sucked less, it made sense to pay a hefty premium for anything approaching reliable. But because of economies of scale and improving technology, today's hardware is in a sweet spot: it's good enough to depend on, and cheap enough to throw away.
Inertia, terrible software, and marketing gusto have kept the expensive-reliable hardware guys afloat, but it won't last. I completely disagree with '[t]he days of running your company on your kids Apple IIc (or its modern day equivilant - the X-toy) are over.' You could never run a business on an IBM XT, but you can power an enterprise with a bunch of Pentiums.
Computers will stop being hard to configure, which some people mistake for powerful, and start being simple squares that you plug into the wall. Consoles, in other words.
Slashdot Mirror
The problem is they don't play well together. I want a tool that uses the information already present in my programs to take out the tedious and error-prone stuff. Why do I have to tell my computer how to map my objects to tables? Why do I have to write so much support code to use CORBA? Why do I have to worry about include paths and library linkage and jar files? Why do I embed javascript in HTML, which is in turn embedded in a string in a script file?
Yeah, I hope they're already thinking about the personal version, because I've been dreaming of Google on my machine for a long time. An intelligent search beats just about any other kind of infrequent interaction: menus, directory navigation, dialog boxes with lots of little pages on them. I want to hit ctrl-g, type in what I'm interested in, and get the right thing.
The output was not in alphabetical order. The LaTeX version was huge - 1000's of pages for a large project, even after I messed with the options to slim it down. It didn't handle templates very well, either. There was a copy of the template's documentation for every instance of the template! I don't remember which version I used, and I only devoted a few days to tuning it, but overall it was not useful for our project.
PERL's a Saggitarius. The most philisophical of all the signs.
A week before Christmas isn't too bad.
It seems like a good idea, but I think typing and clicking little command bursts is faster than explaining to my computer what I want it to do. Maybe for searches, like 'find me a picture of a cow', or 'where was that document I was working on yesterday'?
And then why would there be any symmetry between text processing and graphical user interfaces? I mean, it's worthwhile to look, but I checked and there isn't any. Hook my 'hierarchy generator' up to my 'explorer interface to hierarchies' sounds like a winner until you realize you need type safety to prevent the whole thing blowing up, and then the expressive power of functions to make more useful stuff happen (and documentation, so the GUI can prompt you to fill in the occasional blank).
At this point I sound like I want to replace pipes with something a bit more strict, which is true. But not for GUIs. Oh no. I want TeX.
Why have I ever in my life had to position or resize a window? That's like manually pushing text around in a terminal. And then menus are a step in the right direction, but they hide too much information. Icons are a little helpful in my file browser, but then graphic artists run off and think I can understand what they mean by just looking at them.
I want a text formatter to lay out my entire computer. I want to control how many pages I see at a time, and be able to navigate between the pages. I want a help system that shows me what hotkeys are available right this second. Instead of menus where I get to hunt in little 30x60 text lists, I want to go to help mode and maybe get a little text search to find the page of the user interface that lets me set margins or whatever. I want documents, not applications. I want an object database, not files. I want interfaces with conformance tests, not standards.
Our development environment is twenty-five years old. Our GUI comes straight from the first 'well maybe we should make these things easy to use' thoughts a few (dozen) academic leaders had. We need some independent thinkers to try something different.
Oh- no mouse. There's GOT to be a way to send the computer commands with having to take my hands off of my 101-key keyboard. Buttons remind me what I can do; that doesn't mean I actually have to walk over there and push the things.
I called one of my local Clear Channel stations, and the lady told me she hadn't seen a list. And yeah, songs that mention jets and explosions might be a little innapropriate right now.
A system's archtitecture is the programmer's user interface. With that in mind, the goal of the interface is the same as any other: be pleasant to use. We usually say 'maintainable', but it's the same goal. To help achieve that goal, software architecture should be beautiful, because a beautiful architecture is easy on the mind. A well-designed system is easier to remember and reason about.
Consistency and symmetry are the first two design guidlines that come to mind when people think about well-designed software, and of course software has to actually do what it was written to do. But beyond that, there are thousands of choices that depend on the engineer's aesthetics, and hundreds of choices where bending the rules gets better results. Classification is much more difficult that consistent variable names.
A lot of these 'beautiful displays of complex information' problems were solved by old-school drafting fifty years ago.
Please check the user's id next time you fetch a message. Thanks!
To get trade secrets.
A virus that infects a computer and installs the patch for itself.
Computers got faster, and that enabled better software. As opposed to software getting better with each version, which was the author's opinion.
I didn't read your stuff, so maybe I'm wrong, but... Wouldn't the logical specification be just as hard to write as the program?
And to indulge in a little Microsoft bashing: products that last ten years usually suck compared to the competition, but come out earlier and have better marketing. Once they're the last ones standing, they finally turn their attention to making the product work like it says on the box. I think the reason it takes ten years to write good software is you have to spend the first 8 beating your competition.
And this guy got into Harvard?
There are also contracts floating around- distributors like Hasboro will have an idea that they want developed. This might be a good place to start. Either that, or be as good as Blizzard or Valve :)
Remember that the distributor could lose millions on your game through duplication, warehousing, shelf space, and marketing (which costs probably double the total development price). To these people, you are a lottery ticket!
I work for an medical software company, and we're thinking about developing a highly available database using a peer-to-peer style architecture.
The language of mathematical proofs is so well-defined that you can instruct a machine to process the instructions, eg carry out the proof. This makes a proof a program, so the boundary between absolute, universal truth and copyright-circumventing code is very blurry. In one respect, deCSS is the proof of the theorem that, for any CSS-encrypted byte sequence there is an unencrypted byte sequence. But sin(x)**2 + cos(x)**2 = 1 is also a program; can the DMCA outlaw that?
Now obviously there is a big difference between a program that, to extend the DA's example, blew up the whole world versus a program that calculates digits of PI. But as a developer, the fact that code can be outlawed at all makes me very scared and unhappy, and I would the like the boundary of evil code and good code to be well defined. And I would really like for the MPAA and RIAA to have nothing at all to do with the establishment of that boundary.
Go read Godel, Escher, Bach, it has lots of good stuff on formal systems.
So now I'm a firm believer in personal networking, and of being ready for your chosen career even if your chosen career isn't ready for you.
On a more practical note, most of the designers there worked their way up from testing. On an even more practical note, I was laid off a year later and left the industry, but I could still go back with the experience on my resume.
Emphasis mine. They're going after people who are misusing a Microsoft license, not regular customers. Microsoft's email is a bit confusing, but no more so than slashdot. Thanks for keeping me on my toes. You guys are unresponsible, unaccountable morons. More predictable than a computer.
This is just insane, time to start writing congressmen.
What a stupid stupid way to import stuff: gotta protect the whole thing with #ifdefs, think about forward declarations, urg
The subtle difference is that phase one gets an incomplete base class, while phase two gets the finished object, and hence virtual method calls in the base class get passed to the derived class handler.
To mourn a valueless economy with an artless art project. I dot-commiserate.
Inertia, terrible software, and marketing gusto have kept the expensive-reliable hardware guys afloat, but it won't last. I completely disagree with '[t]he days of running your company on your kids Apple IIc (or its modern day equivilant - the X-toy) are over.' You could never run a business on an IBM XT, but you can power an enterprise with a bunch of Pentiums.
Computers will stop being hard to configure, which some people mistake for powerful, and start being simple squares that you plug into the wall. Consoles, in other words.