The extension idea is naturally quite clever. The only real problem (which is huge enough IMO) is that Q&A sucks for most of these things. And the whole versioning issue also.
Scouring the net, to find just the extension you want, in the correct version, hoping it works is rather shitty, compared to Opera, where you get it as one coherent (often... ) package which has been tested and works.
And not to talk about inconsistent user interfaces. You get options in a million places. Each extension defining it's own UI in a certain way. Most will just be in the extension list, right-click to get options. Others (TBE I think) are so mammoth they add their own seperate menus to the system.
The end result is chaos and cluttered interfaces, poorly Q&A'ed extensions. Extensions which clash (that is one that bites), and so on...
You actually search for ".NET" ? My searches are always tagged with parameters (the code snippet you want, the problem you're trying to solve). Write those first, and then add C# or.NET, it always works.
Importing the datastructure into the language is very easy. In JS you can basicly go "eval(datastring)". No need to either a custom XML walker, or invoke some ActiveX, or create a document/fragment.
Other obvious things, is that the "menuitem" notation in JSON is actually an array, this is not the case in the XML case. There we just have siblings, with coincidentally the same names. So JSON has richer information (this obvious alot of duplication, only one menuitem, instead of three).
It's called "eating your own dog-food", and it's something Microsoft is famous for. And it's of course a good strategy (in most ways). Don't RMS run GCC/Emacs himself perhaps?
Reading others mistakes DOES NOT equate having gone through them yourself. When you make a mistake, in your effort to solve it, you will often head in the wrong direction, and explorer. Maybe time wasted now, but the broader understanding you're hopefully picking up, while looking for the correct answer, will more then pay back any time lost.
The target group for any MS Office for Linux would be used to paying. The largest group is the corporate sector, where having your licenses in order is actually the norm, then not. How it fares for homeusers is something else. Some people do buy Office standalone, others pirate it, others can get it through their Office (that deal where you can install it on your home PC).
If Office for Linux was out, I'd bet good money it would sell well.
It comes down to semantics. Are all programs commercially exploitable, even if we give them away for free? Sure, to some degree. But that's not the point. Apart from small, in the public unheard of stuff, there had never been a OSS "convert", like Netscape. Everyone know (knew?) Netscape, and while maybe far from everyone knew it went OSS, and all that jazz, it did go OSS, and that was the first time ever.
RMS didn't write GCC, to make money. It's not against the his philosophy, and all that. But he basicly wanted the software to be free. That's hardly the same things that went through Netscapes board of directors, or whoever got the idea to go OSS (I have it in a book somewhere, but hey, I'm lazy). That's the difference. I'm sure even RMS can see this.
And FSF has never, to my knowlegde, engaged in actual commerce? Various copyright holders, and others, may have. But I don't think the FSF has. But I may be wrong.
Re:Mozilla was not the first.
on
Netscape Turns 10
·
· Score: 3, Insightful
GCC is not commercial. Netscape was owned by a commercial entity, which released the source. That was, AFAIK the first time that ever happened with a big profile product.
GCC may have provided other people with a living, but that doesn't make it "commercial", in the same sense Netscape was commercially owned.
of course. the demo was of a tight int/string loop. I'm not expecting Windows to run 70% faster, but in special cases, subrutines, it can give a good deal of speed.
And yes, it's dependent on the VM knowing what it's dealing with, as you say, the JVM doesn't, so it probably doesn't get as good (if any) boost.
Generics usually IMPROVE runtime performance, because you don't need to cast, and unbox/box stuff going into whatever. Hejlsberg did a demo at my school, showing between 50 and 70 percent decrease in execution times.
That doesn't make sense. As you browse the web, you download the image. Unless the program is something strange (Moz), then it would probably use Windows libraries to display the image, and bam.
Yeah look. They're a great company, with one assert, search. It's obviously clear how they're scrambling to counter the threat from Microsoft, or lest they be snubbed out.
They're expanding out to community building. This is clear from the unified login you can share between Google groups, and Gmail. Aren't they also doing some friendster like thing? Orkut? They've reached into the OS, via the toolbar, but if it stops there, you can rest assured, Google will be no more in ten years.
Either you're being facetecious, or you don't know how to write a language, and a compiler. The extreme simplicity of s-expressions, versus the full grammar of the java language should alone tell you that string parsing the java language is going to be 100 times as difficult.
Now, I've never done much (if anything) with JSP, so please tell me how Java can easily, and naturally, without much string mucking, can both a generate, and interpreter of itself? Some simple code examples would be great, since of course, I can't really spend time learning all of JSP.
Now, generating bytecode objects, or stuff like that, isn't the idea. It's generating plain text files, with Java. From Java. With my limited knowlegde of Java, I don't see anyway, this can easily be done, without the use of very complex abstracting libraries. Lisp, and S-expressions on the other hand, are supremely suited for this.
S-expressions are a easy to parse notation, and is so simple, that "standard" could surely describe it. At any rate, they mention that while you could write the whole program yourself, having code, that generates your program, would be optimal. Java code, that generates java code? I would hate to see that. S-expressions are easy to generate, consider Lisp code, that generates, or evaluates Lisp code.
The concept of storing files, much like how we do in filing cabinets is cute, but old fashioned. Abstracting the file system, such as into a DB, would allow any view on the data (files) stored, that you may desire. The directory concept is good, but possibly there may be other views, more advantagous at different times.
Such as... well, your music collection. Why be forced to sort it (assuming you do, I do) in one way. You could present multiple views of the music data, one totally flat, one by albums, one by genrer, and so on.
We already see all of this in many different types of apps. Either music managers, or disk information viewers, showing space taken up by this or that file.
We're seeing this in email clients too. Opera's M2 (which sucks otherwise) does a great job of this. Gmail does it too now, though not as well (IMO).
At least, assuming this WinFS is what it sounds like.
I do know what codeweavers is and what wine is. And it's exactly like vmware, it's emulation of a interface. Wine emulates the windows interface, also known as the win32 api (unless wine has changed alot since I last messed with it).
The keypoint of your post, is in the "as if" part. Because no matter what, it'll only ever be "as if" it's running on linux...
Interface design. Knowing exactly what the code you're calling does, will let you do some nasty stuff. Essentially, you're stripping away all the information hiding, and what not.
Sure, "rabbit" is a common term, but does that mean, that real users will actually search for it? Maybe the algorithm relies on more information, such as "rabbit breeding", or whatever. It's like searching for "computers", a virtually useless search. Creating a algorithm that works on common input, instead of ANY input, to improve results, might not be such a dumb thing. But who knows.
It's also likely, that the search returns so many hits, but they're all values so low, that no one stands out. Much like how the new MySQL searching works.
It's of course also possible that the algorithm is crap.
the LaTex package "FiXme" ( http://www.tug.org/tex-archive/macros/latex/contri b/fixme/ ), which uses the command "\fixme{text}" have had this, since 1998.
It, per default will not allow final compilation of your document, if any fixmes are still in the document. The fixmes are also summaried in the stdout of the document (both in the dvi part, and the message part).
(but of course, I have not read the above patent)
Slashdot Mirror
I'm not a buzzword hater, but express your idea without XML, and it will be far more interesting. XML is not a solution to anything, it is a tool.
The extension idea is naturally quite clever. The only real problem (which is huge enough IMO) is that Q&A sucks for most of these things. And the whole versioning issue also.
... ) package which has been tested and works.
...
Scouring the net, to find just the extension you want, in the correct version, hoping it works is rather shitty, compared to Opera, where you get it as one coherent (often
And not to talk about inconsistent user interfaces. You get options in a million places. Each extension defining it's own UI in a certain way. Most will just be in the extension list, right-click to get options. Others (TBE I think) are so mammoth they add their own seperate menus to the system.
The end result is chaos and cluttered interfaces, poorly Q&A'ed extensions. Extensions which clash (that is one that bites), and so on
You actually search for ".NET" ? My searches are always tagged with parameters (the code snippet you want, the problem you're trying to solve). Write those first, and then add C# or .NET, it always works.
Importing the datastructure into the language is very easy. In JS you can basicly go "eval(datastring)". No need to either a custom XML walker, or invoke some ActiveX, or create a document/fragment.
Other obvious things, is that the "menuitem" notation in JSON is actually an array, this is not the case in the XML case. There we just have siblings, with coincidentally the same names. So JSON has richer information (this obvious alot of duplication, only one menuitem, instead of three).
It's called "eating your own dog-food", and it's something Microsoft is famous for. And it's of course a good strategy (in most ways). Don't RMS run GCC/Emacs himself perhaps?
Sony also makes quite a few more products then Apple.
A witty saying proves nothing.
Reading others mistakes DOES NOT equate having gone through them yourself. When you make a mistake, in your effort to solve it, you will often head in the wrong direction, and explorer. Maybe time wasted now, but the broader understanding you're hopefully picking up, while looking for the correct answer, will more then pay back any time lost.
The target group for any MS Office for Linux would be used to paying. The largest group is the corporate sector, where having your licenses in order is actually the norm, then not. How it fares for homeusers is something else. Some people do buy Office standalone, others pirate it, others can get it through their Office (that deal where you can install it on your home PC).
If Office for Linux was out, I'd bet good money it would sell well.
Ah ... cd's and t-shirts. The true hallmark of a great corporation ;)
It comes down to semantics. Are all programs commercially exploitable, even if we give them away for free? Sure, to some degree. But that's not the point. Apart from small, in the public unheard of stuff, there had never been a OSS "convert", like Netscape. Everyone know (knew?) Netscape, and while maybe far from everyone knew it went OSS, and all that jazz, it did go OSS, and that was the first time ever.
RMS didn't write GCC, to make money. It's not against the his philosophy, and all that. But he basicly wanted the software to be free. That's hardly the same things that went through Netscapes board of directors, or whoever got the idea to go OSS (I have it in a book somewhere, but hey, I'm lazy). That's the difference. I'm sure even RMS can see this.
And FSF has never, to my knowlegde, engaged in actual commerce? Various copyright holders, and others, may have. But I don't think the FSF has. But I may be wrong.
GCC is not commercial. Netscape was owned by a commercial entity, which released the source. That was, AFAIK the first time that ever happened with a big profile product.
GCC may have provided other people with a living, but that doesn't make it "commercial", in the same sense Netscape was commercially owned.
of course. the demo was of a tight int/string loop. I'm not expecting Windows to run 70% faster, but in special cases, subrutines, it can give a good deal of speed.
And yes, it's dependent on the VM knowing what it's dealing with, as you say, the JVM doesn't, so it probably doesn't get as good (if any) boost.
Generics usually IMPROVE runtime performance, because you don't need to cast, and unbox/box stuff going into whatever. Hejlsberg did a demo at my school, showing between 50 and 70 percent decrease in execution times.
That doesn't make sense. As you browse the web, you download the image. Unless the program is something strange (Moz), then it would probably use Windows libraries to display the image, and bam.
Yeah look. They're a great company, with one assert, search. It's obviously clear how they're scrambling to counter the threat from Microsoft, or lest they be snubbed out.
... is synergy and convergence ...!
They're expanding out to community building. This is clear from the unified login you can share between Google groups, and Gmail. Aren't they also doing some friendster like thing? Orkut? They've reached into the OS, via the toolbar, but if it stops there, you can rest assured, Google will be no more in ten years.
The keyword
Either you're being facetecious, or you don't know how to write a language, and a compiler. The extreme simplicity of s-expressions, versus the full grammar of the java language should alone tell you that string parsing the java language is going to be 100 times as difficult.
Ok ... so I did ;)
Now, I've never done much (if anything) with JSP, so please tell me how Java can easily, and naturally, without much string mucking, can both a generate, and interpreter of itself? Some simple code examples would be great, since of course, I can't really spend time learning all of JSP.
Now, generating bytecode objects, or stuff like that, isn't the idea. It's generating plain text files, with Java. From Java. With my limited knowlegde of Java, I don't see anyway, this can easily be done, without the use of very complex abstracting libraries. Lisp, and S-expressions on the other hand, are supremely suited for this.
S-expressions are a easy to parse notation, and is so simple, that "standard" could surely describe it. At any rate, they mention that while you could write the whole program yourself, having code, that generates your program, would be optimal. Java code, that generates java code? I would hate to see that. S-expressions are easy to generate, consider Lisp code, that generates, or evaluates Lisp code.
The concept of storing files, much like how we do in filing cabinets is cute, but old fashioned. Abstracting the file system, such as into a DB, would allow any view on the data (files) stored, that you may desire. The directory concept is good, but possibly there may be other views, more advantagous at different times.
... well, your music collection. Why be forced to sort it (assuming you do, I do) in one way. You could present multiple views of the music data, one totally flat, one by albums, one by genrer, and so on.
Such as
We already see all of this in many different types of apps. Either music managers, or disk information viewers, showing space taken up by this or that file.
We're seeing this in email clients too. Opera's M2 (which sucks otherwise) does a great job of this. Gmail does it too now, though not as well (IMO).
At least, assuming this WinFS is what it sounds like.
I do know what codeweavers is and what wine is. And it's exactly like vmware, it's emulation of a interface. Wine emulates the windows interface, also known as the win32 api (unless wine has changed alot since I last messed with it).
The keypoint of your post, is in the "as if" part. Because no matter what, it'll only ever be "as if" it's running on linux...
You mean that he purchased from iTMS using iTunes on Windows on Linux?
Interface design. Knowing exactly what the code you're calling does, will let you do some nasty stuff. Essentially, you're stripping away all the information hiding, and what not.
Python lacks enterprise backing.
Sure, "rabbit" is a common term, but does that mean, that real users will actually search for it? Maybe the algorithm relies on more information, such as "rabbit breeding", or whatever. It's like searching for "computers", a virtually useless search. Creating a algorithm that works on common input, instead of ANY input, to improve results, might not be such a dumb thing. But who knows.
It's also likely, that the search returns so many hits, but they're all values so low, that no one stands out. Much like how the new MySQL searching works.
It's of course also possible that the algorithm is crap.
the LaTex package "FiXme" ( http://www.tug.org/tex-archive/macros/latex/contri b/fixme/ ), which uses the command "\fixme{text}" have had this, since 1998.
It, per default will not allow final compilation of your document, if any fixmes are still in the document. The fixmes are also summaried in the stdout of the document (both in the dvi part, and the message part).
(but of course, I have not read the above patent)