That's assuming a lumped capacitance model, which is false. In reality we're dealing with a convection system in a semi-infinite medium. The heat will not be convected away fast enough from a fusion system before it vapourises the liquid.
Well, that would be nice in theory if browsers were interchangeable, but instead we have a World Wide Web that is designed for IE, full of suicidal ActiveX controls and non-standard crap that other developers have to play catch-up with.
Look at the OpenOffice project - they can never actually develop a real office platform with any innovation as they have to focus entirely on feature-compatibility with Word in order to compete.
That's the friggin' point. THey have a vastly inferior product that owns the market anyways because of their monopoly. If IE was a separate download from FireFox and wasn't pre-installed, people wouldn't touch the thing with an 11-foot pole. This is what differs from the Moz debacle - FireFox is actually a superior product, which has lower market share entirely because of monopoly.
imho, while I usually hate to play the "blame microsoft" game, I put the blame squarely on them on this one - the combo box (and its modified form - the combolistbox) is their baby, and a good idea it was. The problem is that it is utterly stupid for keyboard input. I can't count the number of times that its frustrated me to find it in a program and I realize that, with 36000 entries, I'm gonna have to do a lot of searching for entries. If they'd just designed the widget to handle keyboard input better, this wouldn't be a problem.
But it is, and the majority of users use Win, so it shouldn't be used for long lists of entries where a single-letter input could be in conflict.
While I realise that many users have different experiences, I agree that I dislike grouping. It makes it much more difficult to quickly switch around apps by the taskbar. IMHO, the best approach is single-windowed apps where task switching is internal to the app (like tabbing in Firefox). Still, tabs aren't the best approach either, as they get crowded and there's no way to create multiple rows of tabs.
I'd like to see WM that just uses the title bar for tabs.
I think the problem people don't realise is that "Object" in modern languages means something different from "Object" in C++. In C++ an object was just a struct with extra optional features like privacy, inheritence, etc. In Python,.NET and Java, an object is part of the garbage collector.
For example, in Python (the only interpreter I know under-the-hood), a simple int has to deal with refcounting, next and previous objects in the garbage collection chain, and a host of other issues because it is an object.
Am I understanding this right: In the context of these newer languages, the feature that is desired is the ability to have local, generic structures rather than full blown "objects" to avoid all this overhead. So that a struct is just a shorthand collection of a few primitives?
My understanding was that underneath the CLR was a Lisp-style raw parse tree. If parse tree is the native form of the language (I don't know a thing about it) then Ruby should be just fine. Hell, there's a Scheme implementation for.NET, and if it can do Scheme then it can do anything.
How about getting GM pollen and seeds blown onto your crops to infect your fields, and then getting sued by Monsanto for "stealing" their IP? That'd be fun.
Say what you will about MS, but Visual Studio has always been an excellent product. Nice debugger, and VB is an excellent RAD language (particularly the GUI-drawing system).
Its a developer support company, which means they have very, very public e-mail addresses. Plus, they probably include their support requiests as SPAM.
In this veign, this could be even worse for crashing. For a good example, look at Palm OS software. On the Palm, many programs don't really have much of a "save" function - they just stay in memory. Restarting the machine is a very rare event. The problem is that a handful of programs barf at the thought of crash/restarting - when you restart, you lose your current state, because it goes through its initialization process all over again. Its stupid, but it happens.
So, think about this in terms of PCs - you have many programs on your PC that are no longer designed to reboot - they just expected you to keep the machine running. Then, when MS-Word 2010 crashes the machine and you restart it, you find that your IM has lost all its settings.
Actually, GiTS is a Shirow comic. So, while it may have been very dramatic and spiritual, the actual subject of it was sexy cyborg chicks dressed in tight leather dominatrix outfits with cables sticking out of them.
Well, look at Steam. Steam is ultimately intended to be a monthly service you pay for in exchange for getting oodles of free games and content. This just migth be the future:
Games following the same distribution system as "cable companies". You "subscribe" to the company, and they give you access to their MMOs and downloads of their games. So one company may run several MMOs, and MMOs that are very intermittent in play (like a game where you have to spend most of your time resting and waiting for stuff to complete) would be viable as MMOs, as it would be only one of the many things your gamers do.
Cruise missiles are better controlled. They're computer guided in-flight. Ultimately they can air-detonate if widely off-target. Meanwhile, ballistics are subject to airborne forces of wind and deformation. So if you fire one that looks like it will be of target but isn't, there's nothing you can do about it.
Hmm - a lot of plane 5s actually. Personally, my fave version of Adobe Photoshop was 5 as well, so maybe its just the major version number that's the trend. Wasn't DirectX 5 the big version when people officially stated "this no longer sucks"?
Not just that, but the hydrogen peroxide fuel is generally considered to be the reason Carmack's ship won't succeed. He has lots of good ideas, but from what my friendly-neighborhood physicists have told me, that wasn't one of them.
bah - hit enter accidentally. Anyhow, we had a custom object framework and defining classes using only standard type functions in python was annoying - what we really needed were custom type blocks, but couldn't do that in Python.
Yes, redefining standard objects is messy, but creating new definitions is not.
If it was A a = operatordot(b, "func")(operatorplus(c,d));
or however in whatever language, you'd still have to look up the implementations of said functions. Operator overriding does create the problem that you have _expected_ bahaviour, just as you have an _expected_ behaviour for an if block. But what about the places where there is no appropriate epectancy for what the operator does? Should it lie useless?
Lets say you have two sets. If a user does setA + setB it could do something intelligent like the union, or it coudl throw an error. Either way there is no expected behaviour, so the coder has to look it up anyhow.
Same with blocks. If you redefine "if" then yes, you're an asshat. But what if you want a block for "threadfork"?
Well, fine. I don't want to replace them, I just want to be able to create new ones. Like if I want to create a special treeclass system for non-multiple inheritence classes. This is a relevant problem for me - I was working on an embedded 3d engine, and the limitat
Well, Python+custom blocks is exactly what I'm looking for. Worked on an embedded platform that involved a lot of custom objects and Python's reliance on its core objects only in blocks was frustrating.
While your point is well made, the parent author does also have a good one. Lisp is basically a textual representation of a parse tree. All other languages end up converting your code into a parse tree (except for lisp-alikes). So why not try to work with a parse tree?
The point is this - in Lisp, you can implement the full featureset of another language, the only drawback being Lisp's clumsy legibility. While I probably could implement OOP with transistors, I would never want to. With Lisp, its not that bad. With C, it hurts, badly (see the source code for a c-based OOP language like Python to find out why).
Sounds good - I think Lisp might be just what I'm looking for. Except for the damn C conventions that plague our programming world (personal taste: I like Python's tab-based blocking and Pascal's ":=" assignment) it looks like it has the main feature I'm missing - custom blocks. Can those custom blocks be used for custom code-based object definitions? In Python the "class" is a kind of object. If I want to replace that with a wholly different object that implies a different class structure (like a single-inheritance-only opimized class concept), I can, but there's no statement to construct it - I have to code the functions separately and then pass them into a constructor, instead of being able to use a "class" block. Can that be done in Ruby, create a block not just to define custom flow control systems, but custom object types?
Slashdot Mirror
That's assuming a lumped capacitance model, which is false. In reality we're dealing with a convection system in a semi-infinite medium. The heat will not be convected away fast enough from a fusion system before it vapourises the liquid.
Well, that would be nice in theory if browsers were interchangeable, but instead we have a World Wide Web that is designed for IE, full of suicidal ActiveX controls and non-standard crap that other developers have to play catch-up with.
Look at the OpenOffice project - they can never actually develop a real office platform with any innovation as they have to focus entirely on feature-compatibility with Word in order to compete.
That's the friggin' point. THey have a vastly inferior product that owns the market anyways because of their monopoly. If IE was a separate download from FireFox and wasn't pre-installed, people wouldn't touch the thing with an 11-foot pole. This is what differs from the Moz debacle - FireFox is actually a superior product, which has lower market share entirely because of monopoly.
imho, while I usually hate to play the "blame microsoft" game, I put the blame squarely on them on this one - the combo box (and its modified form - the combolistbox) is their baby, and a good idea it was. The problem is that it is utterly stupid for keyboard input. I can't count the number of times that its frustrated me to find it in a program and I realize that, with 36000 entries, I'm gonna have to do a lot of searching for entries. If they'd just designed the widget to handle keyboard input better, this wouldn't be a problem.
But it is, and the majority of users use Win, so it shouldn't be used for long lists of entries where a single-letter input could be in conflict.
While I realise that many users have different experiences, I agree that I dislike grouping. It makes it much more difficult to quickly switch around apps by the taskbar. IMHO, the best approach is single-windowed apps where task switching is internal to the app (like tabbing in Firefox). Still, tabs aren't the best approach either, as they get crowded and there's no way to create multiple rows of tabs.
I'd like to see WM that just uses the title bar for tabs.
I think the problem people don't realise is that "Object" in modern languages means something different from "Object" in C++. In C++ an object was just a struct with extra optional features like privacy, inheritence, etc. In Python, .NET and Java, an object is part of the garbage collector.
For example, in Python (the only interpreter I know under-the-hood), a simple int has to deal with refcounting, next and previous objects in the garbage collection chain, and a host of other issues because it is an object.
Am I understanding this right: In the context of these newer languages, the feature that is desired is the ability to have local, generic structures rather than full blown "objects" to avoid all this overhead. So that a struct is just a shorthand collection of a few primitives?
My understanding was that underneath the CLR was a Lisp-style raw parse tree. If parse tree is the native form of the language (I don't know a thing about it) then Ruby should be just fine. Hell, there's a Scheme implementation for .NET, and if it can do Scheme then it can do anything.
How about getting GM pollen and seeds blown onto your crops to infect your fields, and then getting sued by Monsanto for "stealing" their IP? That'd be fun.
Say what you will about MS, but Visual Studio has always been an excellent product. Nice debugger, and VB is an excellent RAD language (particularly the GUI-drawing system).
Its a developer support company, which means they have very, very public e-mail addresses. Plus, they probably include their support requiests as SPAM.
In this veign, this could be even worse for crashing. For a good example, look at Palm OS software. On the Palm, many programs don't really have much of a "save" function - they just stay in memory. Restarting the machine is a very rare event. The problem is that a handful of programs barf at the thought of crash/restarting - when you restart, you lose your current state, because it goes through its initialization process all over again. Its stupid, but it happens.
So, think about this in terms of PCs - you have many programs on your PC that are no longer designed to reboot - they just expected you to keep the machine running. Then, when MS-Word 2010 crashes the machine and you restart it, you find that your IM has lost all its settings.
Actually, GiTS is a Shirow comic. So, while it may have been very dramatic and spiritual, the actual subject of it was sexy cyborg chicks dressed in tight leather dominatrix outfits with cables sticking out of them.
Well, look at Steam. Steam is ultimately intended to be a monthly service you pay for in exchange for getting oodles of free games and content. This just migth be the future:
Games following the same distribution system as "cable companies". You "subscribe" to the company, and they give you access to their MMOs and downloads of their games. So one company may run several MMOs, and MMOs that are very intermittent in play (like a game where you have to spend most of your time resting and waiting for stuff to complete) would be viable as MMOs, as it would be only one of the many things your gamers do.
Exactly. Plus a fat upload would allow better game servers, which need much more than 128kbps
Cruise missiles are better controlled. They're computer guided in-flight. Ultimately they can air-detonate if widely off-target. Meanwhile, ballistics are subject to airborne forces of wind and deformation. So if you fire one that looks like it will be of target but isn't, there's nothing you can do about it.
Hmm - a lot of plane 5s actually. Personally, my fave version of Adobe Photoshop was 5 as well, so maybe its just the major version number that's the trend. Wasn't DirectX 5 the big version when people officially stated "this no longer sucks"?
WP peaked at 5.1, Word peaked at 5.1 - any other products for which 5.1 was the magic version number?
Not just that, but the hydrogen peroxide fuel is generally considered to be the reason Carmack's ship won't succeed. He has lots of good ideas, but from what my friendly-neighborhood physicists have told me, that wasn't one of them.
bah - hit enter accidentally. Anyhow, we had a custom object framework and defining classes using only standard type functions in python was annoying - what we really needed were custom type blocks, but couldn't do that in Python.
Yes, redefining standard objects is messy, but creating new definitions is not.
If it was A a = operatordot(b, "func")(operatorplus(c,d));
or however in whatever language, you'd still have to look up the implementations of said functions. Operator overriding does create the problem that you have _expected_ bahaviour, just as you have an _expected_ behaviour for an if block. But what about the places where there is no appropriate epectancy for what the operator does? Should it lie useless?
Lets say you have two sets. If a user does setA + setB it could do something intelligent like the union, or it coudl throw an error. Either way there is no expected behaviour, so the coder has to look it up anyhow.
Same with blocks. If you redefine "if" then yes, you're an asshat. But what if you want a block for "threadfork"?
Well, fine. I don't want to replace them, I just want to be able to create new ones. Like if I want to create a special treeclass system for non-multiple inheritence classes. This is a relevant problem for me - I was working on an embedded 3d engine, and the limitat
Well, Python+custom blocks is exactly what I'm looking for. Worked on an embedded platform that involved a lot of custom objects and Python's reliance on its core objects only in blocks was frustrating.
While your point is well made, the parent author does also have a good one. Lisp is basically a textual representation of a parse tree. All other languages end up converting your code into a parse tree (except for lisp-alikes). So why not try to work with a parse tree?
The point is this - in Lisp, you can implement the full featureset of another language, the only drawback being Lisp's clumsy legibility. While I probably could implement OOP with transistors, I would never want to. With Lisp, its not that bad. With C, it hurts, badly (see the source code for a c-based OOP language like Python to find out why).
Whoops, I meant to say that I think Ruby might be what I'm looking for.
Sounds good - I think Lisp might be just what I'm looking for. Except for the damn C conventions that plague our programming world (personal taste: I like Python's tab-based blocking and Pascal's ":=" assignment) it looks like it has the main feature I'm missing - custom blocks. Can those custom blocks be used for custom code-based object definitions? In Python the "class" is a kind of object. If I want to replace that with a wholly different object that implies a different class structure (like a single-inheritance-only opimized class concept), I can, but there's no statement to construct it - I have to code the functions separately and then pass them into a constructor, instead of being able to use a "class" block. Can that be done in Ruby, create a block not just to define custom flow control systems, but custom object types?