If I were given odds of 80% to get there, and 1% to get back, I wouldn't hesitate to go. Heck, I might even go at 0% to get back.
Bad luck for you, I guess. I would think the "I would go no matter what the risks are" lunatics are among the first to be kicked out during recruitment tests:-).
Yeah well, when you're an FEM-based car crash simulation software running on a Cray T3E, Rails looks like ASP to you, Ruby looks like VBScript, Photoshop looks like MS Paint, and Elisha Cuthbert looks like your mother in law.
Of course MSDN uses ASP.NET -- what did you expect? They also used ASP back when that was en vogue, and ASP is, well, not better than PHP. Livejournal uses Perl/MySQL, Wikipedia uses PHP/MySQL, for example. *Your* point was that ASP.NET or J2EE are the only way to build large sites. Enumerating large sites that use ASP.NET or J2EE doesn't prove this point; enumerating large sites that use something else disproves it. Most web apps are naturally scalable because the only shared parts are the database (cluster) and the session data (for which fast methods exist for sharing the data among machines in your web server farm), so if everything else fails, you can just throw hardware at the problem to speed it up as long as your database scales.
Rails has multiple possibilities for caching web content at different granularities. You could at least have googled for "Rails caching" and incorporated the results in your posting to make it look as if you'd made a comparison.
Essentially everything but assembler and C/C++ uses some variant of "managed code that protects from errors", and Ruby is no exception.
The Ruby and Rails communities are very competent and helpful, and Rails itself is actually easy enough that you can read and understand the source code in a few days. I strongly dispute the claim that big real-world applications (that's those containing large amounts of business code, not just a dragged'n'dropped bunch of databound controls) require less code in ASP.NET than in Ruby. Ruby's OOP features and its meta-object protocol (i.e. runtime "hackability" of artifacts of the language itself) make it easy to write very clean, high-level, declarative code.
Let's just say that you may know something about J2EE and ASP.NET but not much else:-p
If you are making a large scale site or a corporate inter/intranet site you really have two choices ASP.NET and J2EE.
Yeah, proof by assertion. That must be why Slashdot, Amazon, Livejournal and zillions of other "large scale site" don't use them, or what? Amazon and Slashdot use Perl, btw...
Either one is a vast improvement over Ruby/PHP/Perl is both performance and features.
What features of ASP.NET and J2EE are so unique and essential that they make them a "vast improvement" and the only viable choices for larger web apps? Ruby and Rails have features that simply let programmers do the same thing with less code, faster, and with more fun, and make programming in Java/C# look boring and tedious.
This is intriguing. Work with me for a second here, OK?
Mozilla for example is based on ideas, technology and a codebase developed by Netscape. How does Mozilla innovate?
[...]
What point R U trying to make? The parent was talking about Microsoft ripping off competitor's products. Enumerating OSS products that supposedly aren't innovative either won't invalidate that point.
And MySQL was conceptualized as a simple, lean and fast data backend for web sites where many of the "advanced" features are unneeded and slow. You also don't complain that Excel doesn't have transactions. Just because two products support SQL in some way doesn't mean they aim for the exact same feature set.
And then there is Apache, Perl, Python, Ruby, Ports, Apt, Eclipse etc., to name a few.
If they accidentally deliver a patch to IE that makes the browser send 256 requests per second to randomly chosen servers, something that's indistinguishable from "breaking the Internet" will happen.
Just letting an object travel out of the solar system is probably easier than, say, landing it on Mars (or Titan) or having it orbit Saturn. It just takes more time (and some additional swingby maneuvers)...
...and disable the feature by default for all accounts, including admin.
I mean, on other occasions you hear them blather about Windows' totally stellar, fine-grained security architecture, and now they want to prevent Joe Average user from accidentally using raw sockets by, uh, removing the feature altogether?
Paul & john group 1
lisa mel - group 2
john & sue group 3
FUD
ACL
g1 r---
g2 -w---
g3 r-x--
I don't understand. I was talking about implementing this using standard *ix permissions. How do you assign permissions for three groups to a single file then?
(btw, I just noticed that the name "john" occured twice in my example. Those were meant to be two different johns, of course:) )
You can do everything with the *NIX permission model that you can with ACL
Now that's certainly untrue -- you can only assign at most three different permission sets for three different groups of users to a given file. "rwx" would allow eight different permission sets though. You can't, for example, assign "r--" to paul and john, "-w-" to lisa and mel, "r-x" to john and sue, and "---" to anybody else.
How often this is really needed is another question though.
When using fcgi? How does that work? You have multiple Ruby interpreters, and each request may be handled by a different one (or not?). So there's no shared memory in which to store the session data...
If your app must support user sessions, where do you store the sessions' data? I know that rails (or cgi.rb) by default reads and dumps the session data on each request using Ruby's object marshalling/unmarshalling facilities. Does that scale well to your 1000 req/s scenario? do you use NFS or something among the fcgi server machines then? Or do you use a database?
Methinks that storing the sessions in memory wouldn't work because each request may be handled by a different Ruby interpreter running on a different machine.
...well, at least for certain professions like politicians, press offices, financial advisors etc.
It's just too much hassle (and too expensive) to persuade everybody that everything that was publicly announced on April 1st was meant completely seriously (if it was).
What do these "reference" numbers in the explorer-like app mean? Looks like Inode numbers to me... But it's a good thing they still hide file extensions as those are just too confusing for the casual user, uh-huh.
If you're so concerned about "the children," I would suggest you focus on a much more significant threat to their ability to pursue the happiness a famous old document promises them: the gigantic, spiralling deficit. A deficit is a tax on your kids.
Nobody said that one shouldn't (or couldn't) focus on more than one thing at the same time.
Slashdot Mirror
Bad luck for you, I guess. I would think the "I would go no matter what the risks are" lunatics are among the first to be kicked out during recruitment tests :-).
Yeah well, when you're an FEM-based car crash simulation software running on a Cray T3E, Rails looks like ASP to you, Ruby looks like VBScript, Photoshop looks like MS Paint, and Elisha Cuthbert looks like your mother in law.
This is lame. Why doesn't the backside read "deae"?
Rails has multiple possibilities for caching web content at different granularities. You could at least have googled for "Rails caching" and incorporated the results in your posting to make it look as if you'd made a comparison.
Essentially everything but assembler and C/C++ uses some variant of "managed code that protects from errors", and Ruby is no exception.
The Ruby and Rails communities are very competent and helpful, and Rails itself is actually easy enough that you can read and understand the source code in a few days. I strongly dispute the claim that big real-world applications (that's those containing large amounts of business code, not just a dragged'n'dropped bunch of databound controls) require less code in ASP.NET than in Ruby. Ruby's OOP features and its meta-object protocol (i.e. runtime "hackability" of artifacts of the language itself) make it easy to write very clean, high-level, declarative code.
Let's just say that you may know something about J2EE and ASP.NET but not much else :-p
Yeah, proof by assertion. That must be why Slashdot, Amazon, Livejournal and zillions of other "large scale site" don't use them, or what? Amazon and Slashdot use Perl, btw...
Either one is a vast improvement over Ruby/PHP/Perl is both performance and features.
What features of ASP.NET and J2EE are so unique and essential that they make them a "vast improvement" and the only viable choices for larger web apps? Ruby and Rails have features that simply let programmers do the same thing with less code, faster, and with more fun, and make programming in Java/C# look boring and tedious.
It takes three to five years to bang out a scripting language that scripts some "commandlets" on top of an already existing platform? I don't get it.
Mozilla for example is based on ideas, technology and a codebase developed by Netscape. How does Mozilla innovate? [...]
What point R U trying to make? The parent was talking about Microsoft ripping off competitor's products. Enumerating OSS products that supposedly aren't innovative either won't invalidate that point.
And MySQL was conceptualized as a simple, lean and fast data backend for web sites where many of the "advanced" features are unneeded and slow. You also don't complain that Excel doesn't have transactions. Just because two products support SQL in some way doesn't mean they aim for the exact same feature set.
And then there is Apache, Perl, Python, Ruby, Ports, Apt, Eclipse etc., to name a few.
If they accidentally deliver a patch to IE that makes the browser send 256 requests per second to randomly chosen servers, something that's indistinguishable from "breaking the Internet" will happen.
Aww come on, God would *at least* have made it work with something other than text...
Hm, last time I checked, a majority of the American public supported publicly financed space exploration. So there...
And if Mr. Allen would "love to throw a few mil at a project like this", then why doesn't he do so?
Just letting an object travel out of the solar system is probably easier than, say, landing it on Mars (or Titan) or having it orbit Saturn. It just takes more time (and some additional swingby maneuvers)...
I mean, on other occasions you hear them blather about Windows' totally stellar, fine-grained security architecture, and now they want to prevent Joe Average user from accidentally using raw sockets by, uh, removing the feature altogether?
lisa mel - group 2
john & sue group 3
FUD
ACL
g1 r---
g2 -w---
g3 r-x--
I don't understand. I was talking about implementing this using standard *ix permissions. How do you assign permissions for three groups to a single file then?
(btw, I just noticed that the name "john" occured twice in my example. Those were meant to be two different johns, of course:) )
Aha. Looks like a design flaw to me. mount points belong into the VFS layer, not into specific FS drivers.
Now that's certainly untrue -- you can only assign at most three different permission sets for three different groups of users to a given file. "rwx" would allow eight different permission sets though. You can't, for example, assign "r--" to paul and john, "-w-" to lisa and mel, "r-x" to john and sue, and "---" to anybody else.
How often this is really needed is another question though.
IT: Mabir.A Virus Targets Symbian Phones
Science: The End of Mathematical Proofs by Humans?
Hardware: Homemade Mecha Walks in Japan
Hardware: Finally ... RoboShark!
Lunar Dust: A Major Worry for Moon Visitors
So "Lunar Dust" will soon be a new /. category. I certainly appreciate that.
Wow. I see. Thanks a lot.
When using fcgi? How does that work? You have multiple Ruby interpreters, and each request may be handled by a different one (or not?). So there's no shared memory in which to store the session data...
If your app must support user sessions, where do you store the sessions' data? I know that rails (or cgi.rb) by default reads and dumps the session data on each request using Ruby's object marshalling/unmarshalling facilities. Does that scale well to your 1000 req/s scenario? do you use NFS or something among the fcgi server machines then? Or do you use a database?
Methinks that storing the sessions in memory wouldn't work because each request may be handled by a different Ruby interpreter running on a different machine.
It's just too much hassle (and too expensive) to persuade everybody that everything that was publicly announced on April 1st was meant completely seriously (if it was).
Frictionless gearings.
Combustion engines without exhaust heat.
Lighter airplanes.
Faster cars.
Anti-cancer vaccines.
Warp drives.
What do these "reference" numbers in the explorer-like app mean? Looks like Inode numbers to me... But it's a good thing they still hide file extensions as those are just too confusing for the casual user, uh-huh.
Nobody said that one shouldn't (or couldn't) focus on more than one thing at the same time.