you could do this if Gecko came anywhere close to bing compatible with mshtml's interfaces, but it doesn't - it uses it's own set of interfaces. Sure, you could write a 'compatibility' layer to interface between the two, but it would problaby be as complex as a whole new browser.
they provide the components by default in the OS distribution to help developers (external and internal) with the distribution of their applications that use those components.
Imagine if every X application had to supply the font rendering libraries necessary to run. It would be a mess - DLL hell! Integrating a feature into the base distributions does 2 things:
helps uses by not requiring them to download that feature every time they need to install an application that requires it. (the windows CD comes with the necessary HTML libraries for applications)
helps developers by not requiring them to provide that feature, either by including the standard feature distribution with their product, or a custom alternative. (you don't need to download IE/MSHTML if you download an application that uses it)
helps stability by aleviating the DLL hell problem. If the application can depend on a standard set of interfaces instead of some distribution that they provide with their application, then they are more likely to test against the various HTML distributions (ie4/ie5/ie6, etc...) and so their application will be less likely to break when the user updates these common components
Yes, you could except I just created 'tomcruise@actor.com' as a passport account, so actually you can't. Why don't you just go to the passport site and try for yourself?
actually you can use any one of your existing email addresses as a Passport identifier, if you want. Initially it was limited to '@hotmail.com', '@msn.com' and '@passport.com', but this is no longer the case. As i said in my original posting, you could use 'foo@foo.com' if you wanted (and it wasn't already taken). They said that they were going to add some feature to make sure that you didn't 'steal' other people's email addresses (based on domains) but that's not going to be in the first release.
I think MS will stop supporting win98 as soon as XP ships. The same will happen to winME when blackcomb ships. By that time we'll all be arguing about who's.NET implementation is the best...
hmm... i think there's a WinCE Terminal Services client that might do what you want. You'd need to switch devices, though, so that may not be much use to you.
I just got back from Microsoft's ISV summit that they hosted in Redmond. It was very heavy on Windows XP and.NET (not surprisingly) and they also focused on Passport/Hailstorm. Here's a couple of points that they made that I think people here should know:
The requirements for a passport identity are:
a unique identifier. eg: and this is the example they gave: "foo@foo.com". it doesn't need to be your email address. it doesn't even need to be a real email address, just a unique string that looks kinda like an email address.
a password
Optional data includes things like:
your actual email address
zip
age
city
state
etc..
Microsoft won't use any of the information you give them for marketing purposes.
the user can control exactly what information is given to which sites (through use of ACLs)
microsoft won't do verification of any info that you send them - that is up to the 3rd party sites that you use.
they made it pretty clear that security is a big concern for them.
Is it just me or would ARM processors be a great platform for high-power computing? Surely their low power consumption, which is one of the main things slowing down the mega-processors of today (read: x86-base monsters) gives them much greater headroom than their large cousins.
I'm no processor expert. But it would seem to me that if intel can build a 20-stage pipeline for the P4, they're going to have to waste quite a bit of that on misses of the branch-prediction process. The conditional execution of the ARM instruction set would provide an excellent means of reducing the need for much of this branch prediction. Its low power consumption and small size should be an excellent base for addition of further optimizations, such as a deeper pipeline, super-scalar ALUs, larger instruction caches and faster clock speeds. I'm sure if you were to flesh one of these out to the extent that the P4/Palomino have been, it would make quite a processor.
The press-release says that OEMs will be allowed to remove the icons for IE. I'm sure that IE itself (mshtml.dll at al) will still be there. You'll probably still be able to type URLs in the Start->Run dialog to bring up IE.
My main question is this: is anyone really going to buy such a machine from an OEM?
I use apt-get and dselect on my debian system, and it's a dream. I remember installing slackware from scratch back in '93 - upgrading linux has certainly come a long way.
I guess you should probably use the package installer that's native to the distribution you're using. That way you can be sure to get all the right bits, with the correct libs, etc...
The good thing with debian is that almost everything you're likely to need is right there on their servers, you don't have to go and find the packages on some server somewhere. For example I wanted to install apache and squid last week, I typed one command, it downloaded all the stuff, and within a minute or two of waiting for the downloads (I didn't have to type anything) both were configured and running. If your system is also that easy, then stick with it.
I was told that when Microsoft's cisco box crashed (it couldn't handle the load) Foundry sent them a box and had it up and running within 4 hours. Apparently the next day, cisco sent someone up with a special patch for the IOS code and they sent the Foundry box back. We're running Foundry boxes here on our network and I'm really impressed. They're much cheaper and the support is excellent (cheers Smitty).
Yeah, and RedHat just turned a profit. On their site they list it as $600,000 net adjusted income.
I would estimate that since Adam Barr started working at Microsoft (1990) his personal net adjusted income is on the order of TEN TIMES THAT. That may not interest some people, but if you're interested in raising a family, job security, etc... then it can be a good place to start.
Sure windows isn't POSIX compatible, but be fair, the posix calls on Windows are EMULATED. For example, not only is the malloc call on windows implemented with it's own set of memory management structures (on top of VirtualAlloc) it is also thread-safe, which means that it has to lock its internal structures on every call. Very expensive.
Imagine if you wrote the original test in Win32 then ported to the other operating systems using a portability library (such as Mainsoft's MainWin). Would you expect as good performance as if you had written it using POSIX in the first place? No, you'd be stupid to.
In order to do a benchmark like this you need to write different programs, one for each platform, that make the best use of the APIs on those platforms. Only then can you know the true performance of the platform itself. Of course, you have to make sure that the non API-specific parts of your programs are as similar as possible to reduce any discrepancies introduced.
As far as this benchmark goes, I have found significant performance and stability increases by moving my code from the POSIX-style sockets API on Windows to the Winsock2 API. Yeah, it's not portable, but hey, I'm not porting it.
The 'Windows is detecting your hardware' bit of the win95 install was probably one of the more complex (and most reworked) parts of the whole operating system. Mostly due to the fact that probing for one type of card can easily cause another card to get into a weird state. No other operating system has this kind of hardware support, even six years later.
Slashdot Mirror
you could do this if Gecko came anywhere close to bing compatible with mshtml's interfaces, but it doesn't - it uses it's own set of interfaces. Sure, you could write a 'compatibility' layer to interface between the two, but it would problaby be as complex as a whole new browser.
Being informed is a wonderful thing. Being uninformed sure must suck ??
yes, but so did the original release of netscape. but users complained and they both stopped the practice.
Imagine if every X application had to supply the font rendering libraries necessary to run. It would be a mess - DLL hell! Integrating a feature into the base distributions does 2 things:
sure it's cheap. so is unsubstantiated FUD.
Yes, you could except I just created 'tomcruise@actor.com' as a passport account, so actually you can't. Why don't you just go to the passport site and try for yourself?
actually you can use any one of your existing email addresses as a Passport identifier, if you want. Initially it was limited to '@hotmail.com', '@msn.com' and '@passport.com', but this is no longer the case. As i said in my original posting, you could use 'foo@foo.com' if you wanted (and it wasn't already taken). They said that they were going to add some feature to make sure that you didn't 'steal' other people's email addresses (based on domains) but that's not going to be in the first release.
imagine that!
maybe, but the goverment tried to get microsoft to ship windows without a web browser, which by your analogy would be less than 99%.
I think MS will stop supporting win98 as soon as XP ships. The same will happen to winME when blackcomb ships. By that time we'll all be arguing about who's .NET implementation is the best...
actually JScript is the name of microsoft's implementation of ECMAscript, the ECMA-262 scripting language. JavaScript preceeded the ECMA standard.
hmm... i think there's a WinCE Terminal Services client that might do what you want. You'd need to switch devices, though, so that may not be much use to you.
Optional data includes things like:
- your actual email address
- zip
- age
- city
- state
etc..I'm no processor expert. But it would seem to me that if intel can build a 20-stage pipeline for the P4, they're going to have to waste quite a bit of that on misses of the branch-prediction process. The conditional execution of the ARM instruction set would provide an excellent means of reducing the need for much of this branch prediction. Its low power consumption and small size should be an excellent base for addition of further optimizations, such as a deeper pipeline, super-scalar ALUs, larger instruction caches and faster clock speeds. I'm sure if you were to flesh one of these out to the extent that the P4/Palomino have been, it would make quite a processor.
Imagine a beowulf, err...
why are you using VNC? you can do all these things from another win2k machine using the mmc.p.
My main question is this: is anyone really going to buy such a machine from an OEM?
"NEW!!!! Now without Internet Explorer!!!"
Wohoo!
I guess you should probably use the package installer that's native to the distribution you're using. That way you can be sure to get all the right bits, with the correct libs, etc...
The good thing with debian is that almost everything you're likely to need is right there on their servers, you don't have to go and find the packages on some server somewhere. For example I wanted to install apache and squid last week, I typed one command, it downloaded all the stuff, and within a minute or two of waiting for the downloads (I didn't have to type anything) both were configured and running. If your system is also that easy, then stick with it.
actually your heart-rate will be conveniently configurable from any of your mobile devices through HailStorm.
I was told that when Microsoft's cisco box crashed (it couldn't handle the load) Foundry sent them a box and had it up and running within 4 hours. Apparently the next day, cisco sent someone up with a special patch for the IOS code and they sent the Foundry box back. We're running Foundry boxes here on our network and I'm really impressed. They're much cheaper and the support is excellent (cheers Smitty).
Go Girl! err... your honor.
I would estimate that since Adam Barr started working at Microsoft (1990) his personal net adjusted income is on the order of TEN TIMES THAT. That may not interest some people, but if you're interested in raising a family, job security, etc... then it can be a good place to start.
Imagine if you wrote the original test in Win32 then ported to the other operating systems using a portability library (such as Mainsoft's MainWin). Would you expect as good performance as if you had written it using POSIX in the first place? No, you'd be stupid to.
In order to do a benchmark like this you need to write different programs, one for each platform, that make the best use of the APIs on those platforms. Only then can you know the true performance of the platform itself. Of course, you have to make sure that the non API-specific parts of your programs are as similar as possible to reduce any discrepancies introduced.
As far as this benchmark goes, I have found significant performance and stability increases by moving my code from the POSIX-style sockets API on Windows to the Winsock2 API. Yeah, it's not portable, but hey, I'm not porting it.
The 'Windows is detecting your hardware' bit of the win95 install was probably one of the more complex (and most reworked) parts of the whole operating system. Mostly due to the fact that probing for one type of card can easily cause another card to get into a weird state. No other operating system has this kind of hardware support, even six years later.
Who is Bill Gates?
beware, though. doing this will clearly demonstrate IE's OS integration.