True. However, AFAIK that's not how current OpenID sites work (LiveJournal comes to mind, though I've never tried to use OpenID on there; it's just what I remember seeing when somebody commented).
.NET (when I say.NET, I refer primarily to VB.NET and C#, not the stuff on the DLR and the like) is a strongly typed, very structured environment. It's a lot more conducive to writing "clean" code--while it comes at the price of terseness,.NET code tends to flow better, be more readable (I hate "self-documenting" because it's a stupid term, but it almost fits here), and, at least in my experience, tends to be more easily picked up by another developer. It's hard to explain unless you've had it happen, but I'll try.
Whenever I've come onto a new PHP project or had somebody new jump into one of mine, there's been a rather long period of acclimatization; there's a lot of different ways something can be done for a lot of basic tasks, even as basic as logins. With.NET, there's only really One Good Way to perform most tasks, which means that developers generally share a development style that isn't the same in a lot of other languages. (The holy wars such as 1TBS versus Allman don't exist in C# mostly because the Visual C# IDE enforces indenting, bracing, etcetera. It's one of the reasons, I think, that the Mono guys have trouble getting and keeping contributors--people who already know C# look at their coding standards and cringe.)
It also helps that.NET has a ton of stuff built-in to prevent needless code reuse. PEAR has a bunch of it too, but PEAR isn't built right in and integrated with IDEs' intellisense.
I've got about four years of experience with.NET and have been using PHP since PHP3. If I'm working by myself, the choice actually goes to PHP (because I am comfortable with my own style and would rather use my Linux server for applications for myself). If I'm working with a group of people who I've worked with for a long time and I know won't bail, it's a toss-up--either works for us, depending on the situation. But if it's a project where I may be pulling in new developers on a semi-regular basis, like at a "big-name" corporation where there's some regular turnover? Give me.NET any day.
-The APIs suck. (Not the.NET one so much, but I prefer PHP for my web development. And that one is fucking atrocious.)
-I want to use a different password and username everywhere. I don't necessarily want somebody on Ars Technica, for example, being able to go "hey, that guy's the same guy I saw on Slashdot!".
-I use a different password for every account I possess and save them in an encrypted password file, along with my browser; I enter a password maybe once a week, and if one's compromised, nothing else is. If my OpenID provider is compromised (even one I'm running for myself, because while I'm arguably more likely than most people to keep the software updated, everybody has lapses), everything is compromised.
Not interested. The current system is not broken; this is a solution looking for a problem that's already been solved.
When the authentication server is your home server, you can pretty well guard against fishing.
Not always. They know your home server, presumably, and as such could, without too much effort, duplicate the look of it--which is enough to fool a depressingly large percentage of people.
I don't know about you, but I can't remember the last time I didn't have my laptop or phone with me; the former runs Firefox already and has my bookmarks, extensions, and passwords, and the latter my passwords.
IIS7 is remarkably performant, even compared to Apache2, and I haven't seen a significant security update come down the pipe for it or ASP.NET in quite some time.
The initial cost outlay of a Windows machine is higher, of course, but did you consider the other costs? Maintaining an application written in.NET is a lot easier than doing the same in, say, Perl or PHP (unless your staff is universally comprised of language virtuosos, which strikes me as unlikely). Their development teams may already be proficient at.NET. I know that the common wisdom around here is that a good programmer should be able to pick whatever the hell they need to use and be able to get up to speed, but a lot of programmers aren't good but are nominally productive with what they already know. And the support Microsoft offers to big companies is really, really good. (Red Hat and Novell offer good support too, but neither are the same kind of behemoths that Microsoft is.)
There are good reasons to pick something that isn't what you personally like, whether it's proprietary or not.
You're a fucking moron, you know that? It's quite possible to favor closed-source software because it works better for me as a user than Linux.
Linux works fine on my server (well, no, Apache does, and I can run that on BSD or Solaris or OS X or even Windows, if I wanted). It sucks on the desktop. So I use it on a server and use it on the desktop. Open standards make it possible for people (I use the word loosely in your case) like you to use whatever you'd like on your Linux desktop (how's your audio stack working today?) while communicating with the rest of us who use what we want to use.
The post you replied to originally didn't fucking mention Microsoft once--it fucking mentioned an Apple product, for god's sake. And he implied that he's using OpenOffice instead of that Apple product because the Apple product doesn't support ODF!
Christ, you're stupid. You make the people who actually are talking about things in a rational manner look bad. Kill yourself.
There may, in the future, be a benefit to switching everything over to Linux and open source software. That will occur when it's both as good as the closed-source offering from a user perspective (meaning, for the people obsessed with the wrongheaded gnulot ideology and with "but our stuff is technically better", fuck off) and when the cost of continued use of Microsoft is more expensive than the costs of switching.
Nor does the necessary infrastructure even exist with regard to training. (MCSEs aren't worth a lot, no, but nothing both comparable and as available exist; I've never seen a Linux-based cert that would be worthwhile in its current form. For some apps, sure, but you don't hire an MCSE to run a MSSQL database server, you hire someone who's got MSSQL certifications, as you would an Oracle or MySQL box.)
Why would Vietnam actually open source anything, exactly? They're not distributing it outside their own offices, so even the rotten GPL (and yes, gnulots, it IS rotten--MPL, CDDL, or BSD for the win) can't make it so you get code.
(And they're not saying anything in the article about switching to Linux. Read the fucking article next time.)
And that's fine. You don't have to view Flash. But you don't have any real grounds to complain that you can't use a Flash-based site, because it's there.
(That, and as I posted upthread, Flash is an open spec, etc. etc.--Gnash just kind of sucks at more than Flash video right now. I assume it'll improve. Maybe.)
Really? Optimal freedom for everyone, users and developers alike, is "stupid and arbitrary"?
My aim is to have my code used by as many people as possible while still being usable in whatever form a developer wants to use it in. The aim of the GPL is to jam a philosophy down the throat of anyone who wants to develop with their code (the LGPL less so, but I won't use it because it can be relicensed to GPL). Who's aiming to preserve freedom (instead of Stallman's fucked-up "Freedom") here?
That has nothing to do with the (perceived) problem of youths in the UK going out armed with knives. A disturbingly high percentage of these are from middle and upper-class families.
You're not "excluded," you are choosing not to use a de facto standard. Yes, there are de facto standards apart from the de jure ones. Life sucks. Get a helmet.
Yes, as I said: "stupid and arbitrary restrictions." The particulars of the LGPL serve as more of a hassle than anything. One case in a thousand it might be useful, many more than that it's an annoyance.
(Admittedly I overstate its stupidity, as I suppose it's nice that one-in-a-thousand times, but as it's GNU-backed, it really does immediately go in the "distrust and shun" category for me. I do not approve of the GNU licensing scheme at all, though the LGPL is less cretinous than the GPL.)
I intentionally couple the two, because for a while KDE 3.5 was worth the hassle of WINE because it was awfully nice. But if it's end-of-lifed, without the nagging bugs being fixed anymore, there's no reason to stick with it. At least XP still gets bug fixes.
I disagree on the "work decently" part, especially when it comes to non-Java languages. PHP is OK (either with Zend or another one that I kind of liked, but the name of the latter escapes me right now). But C++? Either KDevelop on Linux or VS on Windows blow it out of the water.
And VS2005 does support Java (sort of) with J#, although not excellently anymore. I know I've seen Java toolkits for VS2005/2008, too, though maybe not with integrated Java debugger engines (I didn't look closely enough).
-Kludgy documentation. Microsoft's documentation is very nice and is logically organized; the Sun JDK docs are a bit of a mess.
-Bad, and I mean bad, GUI frameworks. Compare Swing to Windows.Forms and you'll see what I mean pretty quickly.
-Arbitrary and stupid value/reference type division. There is no good reason to have int and Integer be different things. In.NET, the C# "int" is the same as System.Int32. While there are types that are by default non-nullable (such as Int32), this can be overruled by using the Nullable generic, aka the "?" operator (so a nullable Int32 is written as "Int32? foo").
-Crappy platform-specific programming. Yes, yes, I know, Java is cross-platform--but Mono handles this very nicely; you can put libFoo.so and libFoo.dll in the same directory, import functions from "libFoo", and it will do the Right Thing.
-ActionListeners. I mean...seriously..NET delegates make the ActionListener paradigm look sad and pathetic.
That said, there are a number of problems with.NET too (the collections API is somewhat retarded), but the only place I use Java these days is when I'm fooling around with GWT.
You can include MPL or CDDL licensed files in a proprietary application. Instead of the boundary being at the process level, it's at the source file level. If Foo.c is MPL and Bar.c is proprietary, you can include them in the same application; you only have to distribute changes to Foo.c. Sort of like the LGPL in the sense that it's modular, but without the stupid and arbitrary restrictions.
It'd be nicer to be able to actually make a living off it, if we're supposed (so sayeth The Great Stallman) to make all our software GPL. Only people who seem to be making make-a-living money are people working for big corporations.
You're very right about the tech advances making it a bad idea, but not so right about the details of the Saturn Vs. We have the blueprints and, for most of the esoteric bits, actual examples "lying around". It wouldn't be easy, but it'd be doable.
Common misconception about the Saturn V. The plans and resources are still there.
That said, given the changes in technology since then, you're right--it doesn't make sense. But the idea that we don't have the data necessary to build a Saturn V if we need to is a joke. It would not be easy, but there are examples and/or blueprints for, quite literally, everything in one of those.
Slashdot Mirror
True. However, AFAIK that's not how current OpenID sites work (LiveJournal comes to mind, though I've never tried to use OpenID on there; it's just what I remember seeing when somebody commented).
.NET (when I say .NET, I refer primarily to VB.NET and C#, not the stuff on the DLR and the like) is a strongly typed, very structured environment. It's a lot more conducive to writing "clean" code--while it comes at the price of terseness, .NET code tends to flow better, be more readable (I hate "self-documenting" because it's a stupid term, but it almost fits here), and, at least in my experience, tends to be more easily picked up by another developer. It's hard to explain unless you've had it happen, but I'll try.
Whenever I've come onto a new PHP project or had somebody new jump into one of mine, there's been a rather long period of acclimatization; there's a lot of different ways something can be done for a lot of basic tasks, even as basic as logins. With .NET, there's only really One Good Way to perform most tasks, which means that developers generally share a development style that isn't the same in a lot of other languages. (The holy wars such as 1TBS versus Allman don't exist in C# mostly because the Visual C# IDE enforces indenting, bracing, etcetera. It's one of the reasons, I think, that the Mono guys have trouble getting and keeping contributors--people who already know C# look at their coding standards and cringe.)
It also helps that .NET has a ton of stuff built-in to prevent needless code reuse. PEAR has a bunch of it too, but PEAR isn't built right in and integrated with IDEs' intellisense.
I've got about four years of experience with .NET and have been using PHP since PHP3. If I'm working by myself, the choice actually goes to PHP (because I am comfortable with my own style and would rather use my Linux server for applications for myself). If I'm working with a group of people who I've worked with for a long time and I know won't bail, it's a toss-up--either works for us, depending on the situation. But if it's a project where I may be pulling in new developers on a semi-regular basis, like at a "big-name" corporation where there's some regular turnover? Give me .NET any day.
I wouldn't.
-The APIs suck. (Not the .NET one so much, but I prefer PHP for my web development. And that one is fucking atrocious.)
-I want to use a different password and username everywhere. I don't necessarily want somebody on Ars Technica, for example, being able to go "hey, that guy's the same guy I saw on Slashdot!".
-I use a different password for every account I possess and save them in an encrypted password file, along with my browser; I enter a password maybe once a week, and if one's compromised, nothing else is. If my OpenID provider is compromised (even one I'm running for myself, because while I'm arguably more likely than most people to keep the software updated, everybody has lapses), everything is compromised.
Not interested. The current system is not broken; this is a solution looking for a problem that's already been solved.
When the authentication server is your home server, you can pretty well guard against fishing.
Not always. They know your home server, presumably, and as such could, without too much effort, duplicate the look of it--which is enough to fool a depressingly large percentage of people.
I don't know about you, but I can't remember the last time I didn't have my laptop or phone with me; the former runs Firefox already and has my bookmarks, extensions, and passwords, and the latter my passwords.
IIS7 is remarkably performant, even compared to Apache2, and I haven't seen a significant security update come down the pipe for it or ASP.NET in quite some time.
The initial cost outlay of a Windows machine is higher, of course, but did you consider the other costs? Maintaining an application written in .NET is a lot easier than doing the same in, say, Perl or PHP (unless your staff is universally comprised of language virtuosos, which strikes me as unlikely). Their development teams may already be proficient at .NET. I know that the common wisdom around here is that a good programmer should be able to pick whatever the hell they need to use and be able to get up to speed, but a lot of programmers aren't good but are nominally productive with what they already know. And the support Microsoft offers to big companies is really, really good. (Red Hat and Novell offer good support too, but neither are the same kind of behemoths that Microsoft is.)
There are good reasons to pick something that isn't what you personally like, whether it's proprietary or not.
You're a fucking moron, you know that? It's quite possible to favor closed-source software because it works better for me as a user than Linux.
Linux works fine on my server (well, no, Apache does, and I can run that on BSD or Solaris or OS X or even Windows, if I wanted). It sucks on the desktop. So I use it on a server and use it on the desktop. Open standards make it possible for people (I use the word loosely in your case) like you to use whatever you'd like on your Linux desktop (how's your audio stack working today?) while communicating with the rest of us who use what we want to use.
The post you replied to originally didn't fucking mention Microsoft once--it fucking mentioned an Apple product, for god's sake. And he implied that he's using OpenOffice instead of that Apple product because the Apple product doesn't support ODF!
Christ, you're stupid. You make the people who actually are talking about things in a rational manner look bad. Kill yourself.
One of the few good posts I've seen.
There may, in the future, be a benefit to switching everything over to Linux and open source software. That will occur when it's both as good as the closed-source offering from a user perspective (meaning, for the people obsessed with the wrongheaded gnulot ideology and with "but our stuff is technically better", fuck off) and when the cost of continued use of Microsoft is more expensive than the costs of switching.
RedHat and Novell are big, yes.
Neither even approach the size of Microsoft.
Nor does the necessary infrastructure even exist with regard to training. (MCSEs aren't worth a lot, no, but nothing both comparable and as available exist; I've never seen a Linux-based cert that would be worthwhile in its current form. For some apps, sure, but you don't hire an MCSE to run a MSSQL database server, you hire someone who's got MSSQL certifications, as you would an Oracle or MySQL box.)
Why would Vietnam actually open source anything, exactly? They're not distributing it outside their own offices, so even the rotten GPL (and yes, gnulots, it IS rotten--MPL, CDDL, or BSD for the win) can't make it so you get code.
(And they're not saying anything in the article about switching to Linux. Read the fucking article next time.)
And that's fine. You don't have to view Flash. But you don't have any real grounds to complain that you can't use a Flash-based site, because it's there.
(That, and as I posted upthread, Flash is an open spec, etc. etc.--Gnash just kind of sucks at more than Flash video right now. I assume it'll improve. Maybe.)
Good call. We should ban all xylemic substances.
Really? Optimal freedom for everyone, users and developers alike, is "stupid and arbitrary"?
My aim is to have my code used by as many people as possible while still being usable in whatever form a developer wants to use it in. The aim of the GPL is to jam a philosophy down the throat of anyone who wants to develop with their code (the LGPL less so, but I won't use it because it can be relicensed to GPL). Who's aiming to preserve freedom (instead of Stallman's fucked-up "Freedom") here?
That has nothing to do with the (perceived) problem of youths in the UK going out armed with knives. A disturbingly high percentage of these are from middle and upper-class families.
Your ideas are intriguing to me and I wish to subscribe to your newsletter.
You're not "excluded," you are choosing not to use a de facto standard. Yes, there are de facto standards apart from the de jure ones. Life sucks. Get a helmet.
And to put more holes in your stupid, stupid whine, the Adobe SWF (Flash file) specification is available, no strings attached. (It's missing RTMP, but Gnash has reverse engineered it and published their findings.)
Yes, as I said: "stupid and arbitrary restrictions." The particulars of the LGPL serve as more of a hassle than anything. One case in a thousand it might be useful, many more than that it's an annoyance.
(Admittedly I overstate its stupidity, as I suppose it's nice that one-in-a-thousand times, but as it's GNU-backed, it really does immediately go in the "distrust and shun" category for me. I do not approve of the GNU licensing scheme at all, though the LGPL is less cretinous than the GPL.)
You will never "prevent" crime. You can make it less appealing at best.
"Knife crime" will turn into "screwdriver crime" if you ban pointed knives.
Or they'll sharpen points on their knives themselves. It's not hard.
I intentionally couple the two, because for a while KDE 3.5 was worth the hassle of WINE because it was awfully nice. But if it's end-of-lifed, without the nagging bugs being fixed anymore, there's no reason to stick with it. At least XP still gets bug fixes.
I disagree on the "work decently" part, especially when it comes to non-Java languages. PHP is OK (either with Zend or another one that I kind of liked, but the name of the latter escapes me right now). But C++? Either KDevelop on Linux or VS on Windows blow it out of the water.
And VS2005 does support Java (sort of) with J#, although not excellently anymore. I know I've seen Java toolkits for VS2005/2008, too, though maybe not with integrated Java debugger engines (I didn't look closely enough).
Some things that come to mind:
-Kludgy documentation. Microsoft's documentation is very nice and is logically organized; the Sun JDK docs are a bit of a mess.
-Bad, and I mean bad, GUI frameworks. Compare Swing to Windows.Forms and you'll see what I mean pretty quickly.
-Arbitrary and stupid value/reference type division. There is no good reason to have int and Integer be different things. In .NET, the C# "int" is the same as System.Int32. While there are types that are by default non-nullable (such as Int32), this can be overruled by using the Nullable generic, aka the "?" operator (so a nullable Int32 is written as "Int32? foo").
-Crappy platform-specific programming. Yes, yes, I know, Java is cross-platform--but Mono handles this very nicely; you can put libFoo.so and libFoo.dll in the same directory, import functions from "libFoo", and it will do the Right Thing.
-ActionListeners. I mean...seriously. .NET delegates make the ActionListener paradigm look sad and pathetic.
That said, there are a number of problems with .NET too (the collections API is somewhat retarded), but the only place I use Java these days is when I'm fooling around with GWT.
You can include MPL or CDDL licensed files in a proprietary application. Instead of the boundary being at the process level, it's at the source file level. If Foo.c is MPL and Bar.c is proprietary, you can include them in the same application; you only have to distribute changes to Foo.c. Sort of like the LGPL in the sense that it's modular, but without the stupid and arbitrary restrictions.
It'd be nicer to be able to actually make a living off it, if we're supposed (so sayeth The Great Stallman) to make all our software GPL. Only people who seem to be making make-a-living money are people working for big corporations.
You're very right about the tech advances making it a bad idea, but not so right about the details of the Saturn Vs. We have the blueprints and, for most of the esoteric bits, actual examples "lying around". It wouldn't be easy, but it'd be doable.
Also very right about kerosene over hydrogen.
Common misconception about the Saturn V. The plans and resources are still there.
That said, given the changes in technology since then, you're right--it doesn't make sense. But the idea that we don't have the data necessary to build a Saturn V if we need to is a joke. It would not be easy, but there are examples and/or blueprints for, quite literally, everything in one of those.