I think you misunderstood me. I don't think it's a good idea for blocks not to have their own scope. However, in C++, as in JavaScript, it's not a problem to declare variables at any point within a function (e.g: in the middle of the function body).
No, it is a problem, because declared variables have block scope. That is, { int c = 3; int x = 4;... } is exactly the same as (let* ((c 3) (x 4))...). If you declare a variable deep inside a bunch of loops in C, it won't be visible outside those loops, even if you're still within the same function. So Scheme is really not that different from Java or C in this respect. Yes, it is different from JavaScript, but JavaScript's behavior is largely considered a misfeature.
I much prefer to declare variables when I need them rather than being forced to predeclare them at the top of the current block.
I much prefer to declare variables so that they have as small a scope as possible, rather than having them leak out to the rest of a function even though they're only used inside a single for loop.
Essentially, I with that, in Scheme, you could write a define anywhere in a function, and it would add a new variable to the current scope.
Well, essentially you can, so long as you put the rest of the code that needs to use that variable inside the let block. So in practice, lexical scope doesn't need to result in all the variables being declared at the top of the function; only the ones you need to access throughout the function. In my Java code, typically the only variable declared at the top of a method is one called 'result' that will contain the value to pass back to the caller when the return statement on the last line is executed.
What I meant is that in JS you can declare a new variable at any point in a function by writing "var foo =...".
That's true, but the variable still has function scope no matter where it's declared, so it's generally considered bad style. JSLint will flag it as an error, for example.
In Scheme, the only way to do that without a let block is to write a bunch of defines at the beginning of a function, and initialize them to some bogus value, so that you can redefine them later.
If you actively want function scoping then you and I are going to have very different ideas about what constitutes a good programming language. I'd say that you're probably in the minority on this one as well; C, C++, Java, Ruby, Python, none of them have function scope.
That, or nobody in the LISP world bothered to really put the work into a proper optimizing JIT.
Well, there are a bunch of Lisp and Scheme variants that target the JVM, such as ABCL. But yeah, until recently the speedy Lisps have been CMUCL and SBCL, which compile to native code rather than using a JIT in the VM. I think that's because it's pretty recent that VMs using JIT have started to be viewed as offering acceptable performance. (And on Slashdot, you'll still find a lot of people who dispute even that, and still believe in static native compilation as the only way to get fast code.)
It became annoying, however, because Scheme lacks convenient ways of creating local variable declarations without nesting the whole thing into a let form.
Most languages lack convenient ways to create local variable declarations without nesting the whole thing in a code block. In fact, JavaScript lacks a convenient way to create local variable declarations without using an anonymous function.
Another issue was the complete lack of a useful standard library. No hash maps, no sorting function. Even the most basic string functions were lacking. If I'm not mistaken, the Scheme spec doesn't even specify a way to get the numerical code for a character...
And then the performance of the code I'd written (pretty straightforwardly) was abysmal.
Well, it's true that Guile has horrible performance, even compared to other Scheme implementations'; but then, JavaScript had pretty horrible performance until Google and Mozilla started putting serious effort into speeding it up.
Yes and no. Add-ons were the reason for using Firefox back when the competition was IE, but they were also the reason I stopped using Firefox.
My normal Firefox setup was Web Developer, NoScript, CookieSafe, and Firebug. The browser was unreliable and crashy. Chrome offered me all of the same functionality without needing any add-ons. So I switched.
Apple are removing everything that's GPLv3 from OS X (e.g. Samba). I think that's why they removed MySQL. I don't think it has anything to do with Oracle.
The EFF has no say over the GPL3's terms or Apple's terms. Their opinion is just opinion, nothing more.
The fact that GPL3 software can't be locked down isn't a problem, it's a feature. The fact that Apple is gradually locking down the Mac, that's the problem, speaking as a Mac user.
I used plain TeX for my dissertation. These days I use ConTeXt for documents. But actually, yes, there are tools to generate tables of contents for markdown documents.
Also, the root account is enabled by default. I know you can do this in ubuntu also, but it's one of a long list of annoyances I have with that distribution.
Why? There's nothing you can do with root enabled that you can't do with sudo -i except log in directly to the GUI as root, and you shouldn't be doing that anyway.
Not the person you're asking, but personally, if the terms are rental terms because of DRM, I expect rental pricing, or pricing less than the cost of buying the thing, reading/playing/watching it, and then selling it used.
So typically I buy PSN titles, but only if they're $5 or less. One recent exception was Assassin's Creed II for $9.99, because that was cheaper than buying it, playing it, and selling it on used.
Yes. The real reason Apple have dropped MySQL is that they're purging everything that's licensed under GPL3, because of their new business model of patent trolling.
lulz clearly you haven't played Metroid Prime 3. It's unplayable with wagglemotion and there's no option to return it to 'real' controls.
On the contrary, turn on expert mode in the menus. The nunchuck then controls movement, and the Wii remote aims the gun cursor; no wagglemotion. It's then a much better experience than FPSs with a PS3 DualShock controller, in my experience. It's not quite as precise as a mouse for aiming, but it's better than a keyboard for movement.
If Nintendo was a sheepherder then it would not just kill the sheep for its wool, it would machine gun the entire herd to do a blood sampling. By all means, bleed me for every nickle I got but do you got to bugger me at the same time while dancing on my mothers grave?
Slashdot Mirror
1994 is when they decided not to use Tcl, so that's the date on the rationale. When would you like them to have made the decision?
Also cypherpunk2:cypherpunk
I think you misunderstood me. I don't think it's a good idea for blocks not to have their own scope. However, in C++, as in JavaScript, it's not a problem to declare variables at any point within a function (e.g: in the middle of the function body).
No, it is a problem, because declared variables have block scope. That is, { int c = 3; int x = 4; ... } is exactly the same as (let* ((c 3) (x 4)) ...). If you declare a variable deep inside a bunch of loops in C, it won't be visible outside those loops, even if you're still within the same function. So Scheme is really not that different from Java or C in this respect. Yes, it is different from JavaScript, but JavaScript's behavior is largely considered a misfeature.
I much prefer to declare variables when I need them rather than being forced to predeclare them at the top of the current block.
I much prefer to declare variables so that they have as small a scope as possible, rather than having them leak out to the rest of a function even though they're only used inside a single for loop.
Essentially, I with that, in Scheme, you could write a define anywhere in a function, and it would add a new variable to the current scope.
Well, essentially you can, so long as you put the rest of the code that needs to use that variable inside the let block. So in practice, lexical scope doesn't need to result in all the variables being declared at the top of the function; only the ones you need to access throughout the function. In my Java code, typically the only variable declared at the top of a method is one called 'result' that will contain the value to pass back to the caller when the return statement on the last line is executed.
What I meant is that in JS you can declare a new variable at any point in a function by writing "var foo = ...".
That's true, but the variable still has function scope no matter where it's declared, so it's generally considered bad style. JSLint will flag it as an error, for example.
In Scheme, the only way to do that without a let block is to write a bunch of defines at the beginning of a function, and initialize them to some bogus value, so that you can redefine them later.
If you actively want function scoping then you and I are going to have very different ideas about what constitutes a good programming language. I'd say that you're probably in the minority on this one as well; C, C++, Java, Ruby, Python, none of them have function scope.
That, or nobody in the LISP world bothered to really put the work into a proper optimizing JIT.
Well, there are a bunch of Lisp and Scheme variants that target the JVM, such as ABCL. But yeah, until recently the speedy Lisps have been CMUCL and SBCL, which compile to native code rather than using a JIT in the VM. I think that's because it's pretty recent that VMs using JIT have started to be viewed as offering acceptable performance. (And on Slashdot, you'll still find a lot of people who dispute even that, and still believe in static native compilation as the only way to get fast code.)
Not Invented Here Syndrome is why they didn't use Tcl either (which is still being actively developed).
Tcl being a joke of a language is why they didn't use Tcl; see Stallman's original explanation.
It became annoying, however, because Scheme lacks convenient ways of creating local variable declarations without nesting the whole thing into a let form.
Most languages lack convenient ways to create local variable declarations without nesting the whole thing in a code block. In fact, JavaScript lacks a convenient way to create local variable declarations without using an anonymous function.
Another issue was the complete lack of a useful standard library. No hash maps, no sorting function. Even the most basic string functions were lacking. If I'm not mistaken, the Scheme spec doesn't even specify a way to get the numerical code for a character...
Well, we're talking about Guile here, which has sort functions and regexp string manipulation.
And then the performance of the code I'd written (pretty straightforwardly) was abysmal.
Well, it's true that Guile has horrible performance, even compared to other Scheme implementations'; but then, JavaScript had pretty horrible performance until Google and Mozilla started putting serious effort into speeding it up.
"Add-ons are the reason people use Firefox."
Yes and no. Add-ons were the reason for using Firefox back when the competition was IE, but they were also the reason I stopped using Firefox.
My normal Firefox setup was Web Developer, NoScript, CookieSafe, and Firebug. The browser was unreliable and crashy. Chrome offered me all of the same functionality without needing any add-ons. So I switched.
I use rsync, which practically shits rainbows.
Except they're dropping support for VIA Unichrome, such as the low power VIA EPIA series. My M10000 only uses 10W of power.
Yes, Apple's excellent HTML 5 support enables web apps to simply install an icon on the iPhone home screen and behave like real apps.
Oh, wait...
HTML 5 apps can be first-class apps on the phone with icons and everything, and don't require bundling and placing in the app store, eh?
No, didn't think so.
Yeah, I'm sure Apple will get right onto that.
Otherwise everyone is still suck using Oracle's version, especially Enterprise users (which I'd imagine accounts for most of Java's use).
IBM's enterprise Java products all use IBM's Java runtime, not Oracle's, as far as I know.
Apple are removing everything that's GPLv3 from OS X (e.g. Samba). I think that's why they removed MySQL. I don't think it has anything to do with Oracle.
The EFF has no say over the GPL3's terms or Apple's terms. Their opinion is just opinion, nothing more.
The fact that GPL3 software can't be locked down isn't a problem, it's a feature. The fact that Apple is gradually locking down the Mac, that's the problem, speaking as a Mac user.
I used plain TeX for my dissertation. These days I use ConTeXt for documents. But actually, yes, there are tools to generate tables of contents for markdown documents.
Apple chose their app store terms to prohibit GPL3 software distribution. They could easily have not done so for OS X.
pretty obvious: he will write his own Desktop Environment.
Inside the monolithic kernel.
Also, the root account is enabled by default. I know you can do this in ubuntu also, but it's one of a long list of annoyances I have with that distribution.
Why? There's nothing you can do with root enabled that you can't do with sudo -i except log in directly to the GUI as root, and you shouldn't be doing that anyway.
Not the person you're asking, but personally, if the terms are rental terms because of DRM, I expect rental pricing, or pricing less than the cost of buying the thing, reading/playing/watching it, and then selling it used.
So typically I buy PSN titles, but only if they're $5 or less. One recent exception was Assassin's Creed II for $9.99, because that was cheaper than buying it, playing it, and selling it on used.
Indeed. For a PhD thesis on the politics of Nicaragua, I'd have started with markdown and then converted that to ePub and LaTeX.
You know that Apple has promotions for specific apps, and limited time free app offers, right?
Yes. The real reason Apple have dropped MySQL is that they're purging everything that's licensed under GPL3, because of their new business model of patent trolling.
lulz clearly you haven't played Metroid Prime 3. It's unplayable with wagglemotion and there's no option to return it to 'real' controls.
On the contrary, turn on expert mode in the menus. The nunchuck then controls movement, and the Wii remote aims the gun cursor; no wagglemotion. It's then a much better experience than FPSs with a PS3 DualShock controller, in my experience. It's not quite as precise as a mouse for aiming, but it's better than a keyboard for movement.
If Nintendo was a sheepherder then it would not just kill the sheep for its wool, it would machine gun the entire herd to do a blood sampling. By all means, bleed me for every nickle I got but do you got to bugger me at the same time while dancing on my mothers grave?
Thanks for keeping things in proper perspective.