steve albini likes it.

Firefox is the best thing since sliced bread and condoms.

Install tab mix plus on it too, and configure it how you like it.

yut wrote:It has always been a nightmare for developers to cross develop, and honestly Microsoft has stuck with W3C recommendations more than Netscape or Mozilla.

Really? Where do you see evidence of this?

I certainly am not well versed in all the W3C standards, but I work pretty regularly with CSS, and every cross-browser snafu that requires a ghetto workaround is always in IE6.

unarmedman wrote:

yut wrote:It has always been a nightmare for developers to cross develop, and honestly Microsoft has stuck with W3C recommendations more than Netscape or Mozilla.

Really? Where do you see evidence of this?

I certainly am not well versed in all the W3C standards, but I work pretty regularly with CSS, and every cross-browser snafu that requires a ghetto workaround is always in IE6.

Seconded. IE6 is a page-breaking pig.
I wondered why he felt compelled to put the word "honestly" in there. It's a speech mannerism that says "to break from my normal bald-faced lie routine, this that I am about to say actually is true"

anybody here use opera? i have found it to be pretty nice. downloaded it for compatibility testing (turns out to be pretty similar to firefox), and use it occasionally.


anybody here happen to know how to test for IE compatibility on a mac?

anybody here happen to know how to test for IE compatibility on a mac?

My understanding is that IE for Mac is no longer going to be updated, and that Safari will be the standard Mac browser.

So if you're wanting to use it as a browser, it might not be the best choice. If you're wanting to use it to test web designs it still wouldn't be the best choice, as CSS 3 (and a some of CSS 2) isn't supported on it.

unarmedman wrote:

anybody here happen to know how to test for IE compatibility on a mac?

My understanding is that IE for Mac is no longer going to be updated, and that Safari will be the standard Mac browser.

So if you're wanting to use it as a browser, it might not be the best choice. If you're wanting to use it to test web designs it still wouldn't be the best choice, as CSS 3 (and a some of CSS 2) isn't supported on it.

you misunderstood what i asked. IE for mac has been at 5.2 for at least 5 years.

i mean is there any browser that does similar bullshit things that IE (say, 6 or 7 or whatever the newest one is for PC) does (like the width/padding problem), so that you can look at your code to see how cross-browser compatible it is.

i am led to believe the answer is no.

you misunderstood what i asked. IE for mac has been at 5.2 for at least 5 years.

i mean is there any browser that does similar bullshit things that IE (say, 6 or 7 or whatever the newest one is for PC) does (like the width/padding problem), so that you can look at your code to see how cross-browser compatible it is.

i am led to believe the answer is no.

Ah - my mistake. Yeah, you're right - IE for mac was a total rewrite and from my understanding is pretty standards-compliant.

stewie wrote:

yut wrote:They have no concept of a stack trace

J2EE is not an operating system, so you're comparing apples to oranges.

Besides, Windows does provide stack trace facilities. Google "minidumps" to learn about how to get Unix-like core dumps (and therefore, stack traces) on Windows. The .NET CLR also has full stack trace support. As does Sun's JVM on Windows.

yut wrote:and crappy logging.

Logging, and how you view log messages is really an application's concern. The OS can provide you with basic generic facilities, but it can never be as good as what can be offered by a tailored application.

yut wrote:We run J2EE on Windows Server, but are looking to move to Unix... Won't effect our code, because it is java and runs in an appserver that works on both platforms.

Good look with that. Get ready for a lot of testing and debugging.

Does your Unix appserver provide an identical level of J2EE compliance as your current one on Windows? Has your app code been written to be OS agnostic? Care to bet that the Sun JVM running on Win32/x86 is going to do the exact same thing with your Java code when run on an IBM-developed JVM on AIX with a PowerPC chipset?

Write Once, Test Everywhere.

If Microsoft stuff is so great, how come they can't even get their stuff to work? Their web applications are buggy, and whenever I go to their training sessions, I see their Senior VP demo-ing stuff that doesn't work. I mean, every time! I go to Java One, and their stuff works! Hotmail is crap ever since they migrated it. The Microsoft site has always had performance issues and loads of downtime.

In terms of logging, the event viewer is pure dookie. The links to issues on the Microsoft site are all broken. App servers like weblogic and websphere have logging out of the box. .NET does not.

A core dump is different than a stack trace. A core dump is the recorded state of working memory when the exception occured. A stack trace is the list of classes where the exception occured, which makes it a trillion times easier to trace back in the actual source code. The first line in the stack trace is 99% of the times where the problem is.

The point I was making with J2EE is not that it is an operating system, but that Microsoft does not do enterprise software well at all. Amazing how much they tried to rip off from J2EE, and how bad their stuff is in comparison.

As for J2EE portability, we have everything in .ear files and they plug right into the same brand of app server on Unix, and it all works fine. The configuration files are the same. The deployment descriptors are the same. QA has regression tested everything. You're looking way too low level here. Appservers provide a uniform platform for developing applications. A good appserver will run the same on Windows and Unix.

Your criticism of Java has some merit (especially with SWING UI components, randomization, system time, etc.), but with enterprise quality J2EE appservers, you will have more problems moving from version to version than from OS to OS. Appservers are all about not re-inventing the wheel and concentrating on logic more than system plumbing.

unarmedman wrote:I certainly am not well versed in all the W3C standards, but I work pretty regularly with CSS, and every cross-browser snafu that requires a ghetto workaround is always in IE6.

Exactly... So you develop for Mozilla, and it doesn't work in IE, which is the closer to the W3C standard. What ghetto workaround do you speak of? Firefox isn't as bad as it's Netscape ancestor (at least we have had no complaints and QA says our stuff works OK)...

I developed a very sophisticated form validation library in javascript, and Netscape/Mozilla was always the "ghetto workaround" case. There was one bug in IE 5 that I had to work around, but Netscape has been far worse.

As far as accessing the DOM, IE has such better ways to do it. I still have goofy stuff in my code to kludge for Netscape...

I think I am out of line in being critical of Firefox. It has become much better than it's Netscape ancestor since it took off after the Mozilla open source project. It's just that people here say they can't see this or that site, and YouTube comes up as html. Seems like it has problems with http headers or MIME types. I've never had that happen in IE.

It's just that, to me, IE is one of the few things MS does well. XP is a piece of crap (try searching for text in an sql file! Good luck, they broke it. I have to use a grep from a Korn shell)... After developing both client side and server side code for 10 years, I would have to say, IE is great, but steer clear of MS on the server side, unless you like ripping your hair out trying to figure out why something doesn't work...

One thing I am trying to do is build a .NET installer for a product from the command line. It builds an MSI, gives me no errors, and the installer is missing most of the application. So, with no info on how to fix it, I will have to begin the process of elimination, etc. to get it to work.

On the J2EE side, we have ANT and all these great build tools. ANT is even great for non-Java builds, and the irony is, I will probably end up using it to build these .NET installers...

yut wrote:If Microsoft stuff is so great, how come they can't even get their stuff to work? Their web applications are buggy, and whenever I go to their training sessions, I see their Senior VP demo-ing stuff that doesn't work. I mean, every time! I go to Java One, and their stuff works! Hotmail is crap ever since they migrated it. The Microsoft site has always had performance issues and loads of downtime.

I won't disagree with you that Microsoft's server-side ventures haven't been successful. I just disagreed with some of your criticisms of things which had nothing really to do with server-side enterprisey things like appservers, and more to do with the underlying OS.

yut wrote:App servers like weblogic and websphere have logging out of the box. .NET does not.

Not true. Have a look at the System.Diagnostics namespace. It's provided out of the box.

yut wrote:A core dump is different than a stack trace. A core dump is the recorded state of working memory when the exception occured. A stack trace is the list of classes where the exception occured, which makes it a trillion times easier to trace back in the actual source code. The first line in the stack trace is 99% of the times where the problem is.

Uhh...a core dump contains stack trace information. When the system dumps its memory state, you get everything, not just the heap. The stack is just another part of working memory. Why else would pstack take a core dump as an argument?

yut wrote:The point I was making with J2EE is not that it is an operating system, but that Microsoft does not do enterprise software well at all. Amazing how much they tried to rip off from J2EE, and how bad their stuff is in comparison.

I think the CLR is a much cleaner VM design than Sun's JVM. Garbage collection is not as retarded in .NET, for one thing. There are millions of client-side applications running just fine on it, but I agree they haven't been able to nail the server side.

However, I wouldn't start lauding praise onto J2EE, not with the crap I've had to deal with. Websphere and Weblogic are vastly overpriced resource hogs, more wasteful than anything I've ever seen, and I've had to deal with them falling over on many occasions.

yut wrote:As for J2EE portability, we have everything in .ear files and they plug right into the same brand of app server on Unix, and it all works fine. The configuration files are the same. The deployment descriptors are the same. QA has regression tested everything. You're looking way too low level here. Appservers provide a uniform platform for developing applications. A good appserver will run the same on Windows and Unix.

If all you're doing is vanilla J2EE stuff, sure, your .ear will deploy fine. But note that you said "the same brand of app server". One of the goals of J2EE was to have complete interchangeability of appservers for your applications, but this has just been thrown to the wayside as each vendor built more and more lock-in features, intentionally or not.

BTW, I don't work for M$, but I do work in the appserver area.