I suppose the fact that IE has all sorts of nice direct access to the Windows code with god-knows-what tricks embedded to speed it up helps. Firefox is bound by what any non-MS program can do with the API.
Nice try, but no. Mozilla/Firefox made the decsion to add a heap load of bloat as a portability/skins layer. So they run everywhere, slowly.
It's pretty funny that you accuse MS of using "tricks" when the GUI of your browser is written in frickin JavaScript. (Secret MS Speed Trick #1: Don't use a scripting language).
If they wanted to, they could have coded a native Windows app right to the documented Win32 API, and it's very unlikely that you'd see any difference in startup speed. (See Opera)
Since none of the tests apparently simulated client load, this seemed more like an OS benchmark (for MySQL) than a database benchmark. (despite however everyone is spinning it).
In my view, this wasn't something that Mozilla got right until recently. Previously they tended to release security fixes in QA releases such as "1.5a2", because they expected someone else to "productize" their dev work. And yes, there's been huge bugs they haven't been too urgent about fixing because they didn't affect the "stable branch" (1.0 or whatever). In other words, for people using Mozilla/Firefox as their day-to-day browser, the process sucked.
Plus, you're wrong about IE -- it's not that they went a long time between critical fixes, its that there's a new batch every month and they never ever ever catch up. The sandbox architecture is obviously totally misdesigned. But the fact that Microsoft screwed it up doesn't necessarily mean that Mozilla's approach is "superior"... their track record isn't a whole lot better.
> lack of commercial XPI extentions
Gladly I was wrong about this. But if Yahoo can get you to install a toolbar, anyone else can too...
> The XPI security improvements
Even if they didn't copy MS, this was a known issue that wasn't fixed until people used it as a malware vector. That's reactive and not superior design.
> Secunia
IMO, most of these are Spoofs and not security flaws. Point conceeded, but you might want to type in your URLs instead of clicking on links:)
Yes, but common Javascript actions like Alert boxes and popup windows are not part of any official standard. Even for common HTML/Form stuff, the W3C DOM spec is not very complete.
I'd bet that "Win.NET" is really Windows 2003 Server (which was called.NET Server in the beta period)..NET does report itself in the IE user agent, but it wouldn't make sense to put it in that chart.
And you've tried it? Because it works fine here on Apple Mail 10.3. No warning message even.
(PS: Unlike HTTP, email has been able to handle forked Mac files for a long time) (PPS: OS X doesn't really need forked files or Mac metadata, a.scpt file or a "Mach-O" executable is good enough.) (PPPS: If you send a.app directory, Apple Mail even ZIPs it up for you behind the scenes. However it doesn't run from Mail like a Carbon app)
Summary of the AC posts, you may not have been notified about:
+ Where's the evidence "superior process and architecture"? (think this refers to security bugs hiding in secret areas of bugzilla for many months).
+ Marketshare, Firefox has not "run the gauntlet". (this is debatable, but the lack of commercial XPI extentions in general attests to this)
+ Secunia shows a number of unpatched Firefox flaws.
+ Whether people will upgrade promptly. (the firefox auto-upgrade seems broken here)
+ The XPI security improvements were copied straight from Internet Explorer (somewhat true, XPSP2 features were known before Firefox even thought about this)
Anyway, after all that handwaving, it would be interesting to hear you respond to these points.
That was a buffer overflow bug in Outlook and nothing to do with the script engine - it was just a VBS payload instead of an EXE.
Everytime this comes up on Mac boards, people accept that ILOVEYOU:Mac is possible, but there probably won't be any urgency to fix it until it happens.
Applescripts don't run automatically, you can't email them to somebody and have it delete their home directory without their participation
You know, the famous ILOVEYOU.vbs and other script viruses didn't run automatically either -- one had to click on it and dimiss a warning. Furthermore, those viruses rarely required Administrator rights, they ususally just sent mail and copied themselves wilynily.
I see no difference between OS X and Windows in this department. EXECPT for the fact that people are scanning for Windows trojans, so when one hits the Mac world, they'll be unprotected.
Keep in mind that every GUI action on the Mac has a 2 second animiation go with it. So while the Mac-using grandparent is marveling over the "close window" feedback animation that he doesn't get with Windows, Windows users are thinking "why does it take so long for a window to disappear?".
Plus there are some things that are just plainly slower on the Mac, like scrolling.
Yes, although many SGI fanboys blame the symptom and not the disease. (Understandable because everything was hunky-dory and then 3-4 years later they're dead.)
SGI was a big enough company that a software project like Fahrenheit wasn't that big of a deal.
Apache being used more than IIs and yet being more secure proved that one false fucking years ago
This is the standard ESRism that gets trotted in response to the "popularity" argument, but it fails to stand up the slightest bit of critical thinking. We don't even know if Apache is more popular than IIS or not! (You have to seriously misread the Netcraft survey to even come to this conclusion.) And "more secure" seems to assume an unpatched box, which is a meaningless comparison point for people with competant sysadmins.
Furthermore, it may end up proving the opposite point:
Apache is by far the most popular Unix webserver. Does Apache have more vulnrabilties discovred than other Unix webservers? Is Apache attacked more often than other Unix webservers? I suspect it might be, purely due to popularity.
Another good example is SGI. Their unix workstation business was doing great. Two years later it had totally cratered and they were pushing WinTel out of despiration.
The whole OSX-Transition/Slow G4 thing a few years ago *may* have caused serious damage in their professional markets. However, Apple was able to reposition itself as a high-end 'yuppie' home/consumer brand. However, that market is much more fickle.
I dunno man, I couldn't make a Word 2003 file incompatitible with Word 97 even if I wanted to. I've got 15 year old Word 4.0 files that open fine. The entire "insightful" thread lacks information.
But apparently this is some huge problem that y'all deal with on a daily basis or something (and therefore OpenOffice is urgently needed to "solve" it), so I thought I comment about the lack of track record there.
It's unsolvable because content goes into the edonkey network but never leaves (even if it is incomplete). To use a legal example, the guy downloading Fedora 4 is going to have to wait for the guy downloding RedHat 9, and visa-versa.
But OTOH, at least it is possible to get rare/old files from eDonkey, where as BitTorrent basically becomes useless as soon as the file isn't hot. For legal content (such as RedHat 9) this isn't a problem, but good luck finding a random 5 year old movie otherwise.
Let's put 8 different versions of OpenOffice Writer on millions of machines (10% of which have defective hardware, viruses, etc), and see how well works.
This really seems like a "grass is greener" issue. MSOffice has been everywhere for a long time and of course problems sometime crop up. But nobody really knows if OpenOffice interoperates better with itself because it has never been tried.
(And yes, I know about the XML format, but that doesn't prevent intrepetation/implementaiton issues.)
Except "standards" tend to fracture into different bits. People write to implementations without even knowing it. Maintaining perfect interoperability is hard work and requires tons of QA and formalized testing processes generally don't fit into the open source development model. Finding "100 different programs" that all do the exact same thing in the exact same way is probably impossible (except for simple protocols like HTTP). At some point everyone just gives up and converges on one vendor's implementation.
A prime example is NFS. *nix users sometimes find it easier to use the reverse-engineered SMB implementation to get their non-MS boxes talking rather than deal with the 100 different NFS variants (all of which are likely "standard").
Slashdot Mirror
I suppose the fact that IE has all sorts of nice direct access to the Windows code with god-knows-what tricks embedded to speed it up helps. Firefox is bound by what any non-MS program can do with the API.
Nice try, but no. Mozilla/Firefox made the decsion to add a heap load of bloat as a portability/skins layer. So they run everywhere, slowly.
It's pretty funny that you accuse MS of using "tricks" when the GUI of your browser is written in frickin JavaScript. (Secret MS Speed Trick #1: Don't use a scripting language).
If they wanted to, they could have coded a native Windows app right to the documented Win32 API, and it's very unlikely that you'd see any difference in startup speed. (See Opera)
Since none of the tests apparently simulated client load, this seemed more like an OS benchmark (for MySQL) than a database benchmark. (despite however everyone is spinning it).
> "superior process and architecture"
... their track record isn't a whole lot better.
:)
In my view, this wasn't something that Mozilla got right until recently. Previously they tended to release security fixes in QA releases such as "1.5a2", because they expected someone else to "productize" their dev work. And yes, there's been huge bugs they haven't been too urgent about fixing because they didn't affect the "stable branch" (1.0 or whatever). In other words, for people using Mozilla/Firefox as their day-to-day browser, the process sucked.
Plus, you're wrong about IE -- it's not that they went a long time between critical fixes, its that there's a new batch every month and they never ever ever catch up. The sandbox architecture is obviously totally misdesigned. But the fact that Microsoft screwed it up doesn't necessarily mean that Mozilla's approach is "superior"
> lack of commercial XPI extentions
Gladly I was wrong about this. But if Yahoo can get you to install a toolbar, anyone else can too...
> The XPI security improvements
Even if they didn't copy MS, this was a known issue that wasn't fixed until people used it as a malware vector. That's reactive and not superior design.
> Secunia
IMO, most of these are Spoofs and not security flaws. Point conceeded, but you might want to type in your URLs instead of clicking on links
> The "anti-spy" feature doesn't seem all that useful for Firefox
There's many vectors for spyware other than the browser. It's only a matter of time before Kazaa and the like start coming with Firefox add-ins.
I was going to guess Terminal Service users, but since its a Web Dev site, its probably mostly developers running their MSDN copies of 2003.
That's funny. I sent several from the AppleScript sample folder, and they all just ran. I'll try sending a few to my Mac-comrades.
Does FarCry have as much scripting as Doom3? I got the impression it is much more heavy on AI.
(Doom3 is full of Enter Room, Lights Go Out, Door Shuts type things. What do you do about the 2nd player when they get locked out?)
Yes, but common Javascript actions like Alert boxes and popup windows are not part of any official standard. Even for common HTML/Form stuff, the W3C DOM spec is not very complete.
I'd bet that "Win .NET" is really Windows 2003 Server (which was called .NET Server in the beta period). .NET does report itself in the IE user agent, but it wouldn't make sense to put it in that chart.
And you've tried it? Because it works fine here on Apple Mail 10.3. No warning message even.
.scpt file or a "Mach-O" executable is good enough.) .app directory, Apple Mail even ZIPs it up for you behind the scenes. However it doesn't run from Mail like a Carbon app)
(PS: Unlike HTTP, email has been able to handle forked Mac files for a long time)
(PPS: OS X doesn't really need forked files or Mac metadata, a
(PPPS: If you send a
Summary of the AC posts, you may not have been notified about:
+ Where's the evidence "superior process and architecture"?
(think this refers to security bugs hiding in secret areas of bugzilla for many months).
+ Marketshare, Firefox has not "run the gauntlet".
(this is debatable, but the lack of commercial XPI extentions in general attests to this)
+ Secunia shows a number of unpatched Firefox flaws.
+ Whether people will upgrade promptly.
(the firefox auto-upgrade seems broken here)
+ The XPI security improvements were copied straight from Internet Explorer
(somewhat true, XPSP2 features were known before Firefox even thought about this)
Anyway, after all that handwaving, it would be interesting to hear you respond to these points.
That was a buffer overflow bug in Outlook and nothing to do with the script engine - it was just a VBS payload instead of an EXE.
Everytime this comes up on Mac boards, people accept that ILOVEYOU:Mac is possible, but there probably won't be any urgency to fix it until it happens.
Applescripts don't run automatically, you can't email them to somebody and have it delete their home directory without their participation
You know, the famous ILOVEYOU.vbs and other script viruses didn't run automatically either -- one had to click on it and dimiss a warning. Furthermore, those viruses rarely required Administrator rights, they ususally just sent mail and copied themselves wilynily.
I see no difference between OS X and Windows in this department. EXECPT for the fact that people are scanning for Windows trojans, so when one hits the Mac world, they'll be unprotected.
Keep in mind that every GUI action on the Mac has a 2 second animiation go with it. So while the Mac-using grandparent is marveling over the "close window" feedback animation that he doesn't get with Windows, Windows users are thinking "why does it take so long for a window to disappear?".
Plus there are some things that are just plainly slower on the Mac, like scrolling.
> Is that the sequence of events though?
Yes, although many SGI fanboys blame the symptom and not the disease. (Understandable because everything was hunky-dory and then 3-4 years later they're dead.)
SGI was a big enough company that a software project like Fahrenheit wasn't that big of a deal.
Apache being used more than IIs and yet being more secure proved that one false fucking years ago
This is the standard ESRism that gets trotted in response to the "popularity" argument, but it fails to stand up the slightest bit of critical thinking. We don't even know if Apache is more popular than IIS or not! (You have to seriously misread the Netcraft survey to even come to this conclusion.) And "more secure" seems to assume an unpatched box, which is a meaningless comparison point for people with competant sysadmins.
Furthermore, it may end up proving the opposite point:
Apache is by far the most popular Unix webserver. Does Apache have more vulnrabilties discovred than other Unix webservers? Is Apache attacked more often than other Unix webservers? I suspect it might be, purely due to popularity.
Another good example is SGI. Their unix workstation business was doing great. Two years later it had totally cratered and they were pushing WinTel out of despiration.
The whole OSX-Transition/Slow G4 thing a few years ago *may* have caused serious damage in their professional markets. However, Apple was able to reposition itself as a high-end 'yuppie' home/consumer brand. However, that market is much more fickle.
I dunno man, I couldn't make a Word 2003 file incompatitible with Word 97 even if I wanted to. I've got 15 year old Word 4.0 files that open fine. The entire "insightful" thread lacks information.
But apparently this is some huge problem that y'all deal with on a daily basis or something (and therefore OpenOffice is urgently needed to "solve" it), so I thought I comment about the lack of track record there.
It's unsolvable because content goes into the edonkey network but never leaves (even if it is incomplete). To use a legal example, the guy downloading Fedora 4 is going to have to wait for the guy downloding RedHat 9, and visa-versa.
But OTOH, at least it is possible to get rare/old files from eDonkey, where as BitTorrent basically becomes useless as soon as the file isn't hot. For legal content (such as RedHat 9) this isn't a problem, but good luck finding a random 5 year old movie otherwise.
Wow, version 1.1.1 works with version 1.1.4! That really blew my argument out of the water.
As for "perfectly", you must have done the extensive regression testing to prove it. Feel free to post your data.
Beware! I Live. Looks like my ol UID has a little bit left :)
Which would work great, assuming every version of OOo is 100% bug-free.
You know, Office 97 was a long time ago -- the 20th century FUD recycled endless here gets boring aftera while. Contemporize!
Plus, I doubt you know what OOo may or may not do in the future.
Let's put 8 different versions of OpenOffice Writer on millions of machines (10% of which have defective hardware, viruses, etc), and see how well works.
This really seems like a "grass is greener" issue. MSOffice has been everywhere for a long time and of course problems sometime crop up. But nobody really knows if OpenOffice interoperates better with itself because it has never been tried.
(And yes, I know about the XML format, but that doesn't prevent intrepetation/implementaiton issues.)
Except "standards" tend to fracture into different bits. People write to implementations without even knowing it. Maintaining perfect interoperability is hard work and requires tons of QA and formalized testing processes generally don't fit into the open source development model. Finding "100 different programs" that all do the exact same thing in the exact same way is probably impossible (except for simple protocols like HTTP). At some point everyone just gives up and converges on one vendor's implementation.
A prime example is NFS. *nix users sometimes find it easier to use the reverse-engineered SMB implementation to get their non-MS boxes talking rather than deal with the 100 different NFS variants (all of which are likely "standard").