At the same time that IE was the default browser and far and away the most used browser for Windows. Further IE 5 for Mac was a better browser than IE 6 for Windows. Microsoft's mac division often writes better software since Microsoft can treat that platform as a test platform.
1) Does the distribution provide a GUI for the package manager? 2) Does the package manager allow an install from CD.
The distributions that provide a GUI for most things provide good quality GUIs for package management. So I'd assume if you want one and pick the right kind of distribution the answer to (1) is yes. But if you are scared of the command line why are you using a Linux?
As for (2) every package manager I know of allows you to set directories as sources for files. So/mnt/CD or where ever your CD mounts to is either already added or can be added with no problems, assuming you consider this a feature. Mandriva which is the home distribution I'm most familiar with ships with this on.
Microsoft with NT claimed to be writing a secure, enterprise OS. They didn't claim it was a home OS which could run useful business software.
-- Crappy drivers should simply fail. Or they could take on driver responsibility directly, drivers go through a Microsoft QA process. Apple has bad 3rd party apps break all the time. -- The security model that was originally created for NT is excellent. Microsoft didn't implement it in their own apps. The crappy 3rd party programs shouldn't have worked -- Microsoft has for a long time maintained a "we will get it to run" attitude. The result is developers engage in bad practices
The fact is there is no reason an ignorant home user should be able to install a rootkit even if they tried. Capability architecture is far far safer than the permissions architecture of the Unixes; Windows security should have blown Unix security out of the water.
Microsoft has business practices that are technically illegal. Anti trust law isn't enforced the way it was in the 1970s. Microsoft has avoided engaging in easy solutions that would get it out of anti trust problems. For example: -- A simple uniform OEM windows price is public and available to all OEMs. That removes the whole argument about Microsoft "intimidating" vendors -- Clear policies about 3rd party modifications to Windows, that apply regardless of component or corporate strategic direction (i.e. then bundling doesn't become a strategic move) -- Making their office suite available for multiple platforms thus diminishing the argument they are using these two monopolies to reinforce one another
etc...
Sony conversely during the rootkit fiasco within days removed the rootkit. They didn't fight this in court for years and years.
No it just means that language isn't algebra. A word may not mean what its component parts mean and an expression may not mean what the combination of words mean. The rules for expressions and grammar are complicated.
It depends on which European country but customer service standards can often be much higher in Europe than in America. Think about the range of customer service standards in various types of venues in different countries or even within countries (Paris vs. Rivera). I don't know if you can make comparisons across countries like that for this reason. In the home computing market America basically has one level of service with a pretty narrow range. In clothing for example we have a huge range in service standards. In restaurants we have a huge range (though not as large as the range in Europe).
Unless you mean marketing in a very expansive way it had very little do with marketing. Active X was at that time something like 30x faster than java. The active X gui was the windows GUI which looked much better than the Java 1.1 GUIs.
If Microsoft's lawyers couldn't get the case thrown out, what chance does a free browser have?
Much higher. Take a look at Groklaw and the information they dug up for IBM. The open source community can do open source research and find prior art. They can find employees who implemented the technologies that will site where they got the ideas from. Further they don't have anything to hide so they discovery will be public. I'm not sure what Eolas claims was their unique idea but I know of specific quotes (some Bill Gates interesting enough) from the late 1980s which basically outline the notion of pluggins.
So worst case, which is a very big assumption, Firefox doesn't use plugins. Its open source it can do things like load binary libraries. Same with Konq and Seamonkey. Opera might be hurt but that's an easy change for everyone.
Yeah. I remember the days of gopher (like HTML but text based) when the content to crap ratio was 20:1 (in favor of content). I think.gifs were a net benefit but I'd pay to have a net free of all that stuff
I don't think it was a huge mistake from the get go. Microsoft has always advocated smart client interfaces. HTML is essentially a dumb terminal / mainframe like system. The ability to have full scale applications using all the power of the desktop distributed via the net was a great idea. Where they failed was:
1) The windows NT security model failed and instead became a user model. Had a full fledged capability model been in place the permissions on the browser could have been set much lower than the permission on the user. Active X should not have been supported at all for the Windows 98 family due to poor security model.
2) Microsoft marketing didn't explain to customers that active x was dangerous and should always be turned off except for trusted sites. The warnings when reducing permissions should have been much more harshly worded (and the active X should have had to be changed up front.
3) They should have made available low security active X users for untrusted apps.
This gives you a setup like this:
a) default for most sites is active x is off b) for some sites active X is on but runs with very low privs c) for other sites active x is on runs as you but with lower privs d) for the most trusted sites active x is a normal app
I happen to agree with the position you are presenting in this post regarding point numbers. The person who brought up component upgrades was you Windows XP was a 1 point upgrade from 2000 Pro. Some would argue that it was also a minor upgrade. But what you fail to grasp is that the 10.2,10.3,10.4 is the "release" version which has nothing to do with the versioning of core components such as the kernel and frameworks. In Jaguar, the kernel was 6.x but in Tiger it is 8.x. The same thing applies for major frameworks.. I was arguing those kernel upgrades and additional functionalities were not major upgrades but point upgrades.
They weren't advertised as breaking APIs. In spite of that however apps do still have problems.
So anyway with respect to this post there is no disagreement.
Yes it probably was more like 3% for 10.2 -> 10.3. OTOH well more than 10% broke on 10.1 -> 10.2, I think 10% is about right for 10.3->10.4.... Anyway I agree that 10.3 -> 10.4 is pretty close to 2K to XP. But then again less broke for the 2K -> XP move which was my point.
I don't disagree the open model with recompilation is the better model. What I am point out, is that if you are simply comparing apples to apples (binary compatibility) then
1) Microsoft takes this very seriously 2) They do this with non apps not produced by Microsoft
For example Maxis's Sim City had a whole separate memory subsystem in NT 4.0 to get it to work properly.
As for the point of slowness I don't agree with the NY times article. Here is my position
1) Binary compatibility increases Microsoft's development costs massively. This results in much longer product release cycles and bug fix cycles 2) It doesn't slow down the OS too much (say 30%) 3) The real damage is done because a) Microsoft does not want to leave behind low end desktops. Longhorn as originally conceived would have been a much better product but would have required a $2k + machine to run b) Security and convenience cannot be achieved without changing application developer behavior and that requires Microsoft putting some pressure on developers (i.e. breaking naughty apps).
Between Windows 98 and Windows 2k the entire kernel was replaced, the entire programs model was replaced most interfaces were replaced the administrative system was replaced the whole driver model was changed. Windows 98 literally has almost the same relationship to 2k that classic does to OSX. Its a much bigger change.
10.0 -> 10.4 are evolutionary. Each one is a step forward. Regardless I'm not arguing 10% break between 10.3.4 and 10.3.5 but those changes happen every few months.
We are discussing binary compatibility. This would be the debian equivalent of using options like ignore-breakage, not having apt resolve dependencies issues, etc... I agree with you regarding the Linux model being a much better one in terms of control (and problem resolution). But it doesn't even try to offer binary compatibility. You can't take a debian 1.3.deb file force and install and expect it to work.
There are. You can have a situation where you can't sue unless you case is sort of pre-vetted by a judge. You see this in Europe and it would definitely be worth considering for the US. Incidentally you also have it at the criminal level.
OK I think we may be agreeing then. I was presenting Microsoft's position on business legacy hardware support. I agree that for home its much less important.
You are thinking home users, business users don't upgrade their whole infrastructure. So yes upgrades will be common. Corporate users is really the group that Microsoft is worried about regarding compatibility they are the ones with very high standards in this regard. Home users are pretty forgiving. As an aside, free printers are really really expensive per page; you are paying for them.
Re:same as hardware really, ms laziness?
on
Why Windows is Slow
·
· Score: 1
we're not talking about ms compatibility though, we're talking about culling code for legacy hardware & software. most of compatibility testing, etc. will be done with modern software/hardware i would have thought.
No most of the problems are with the older legacy code base. For example the developers themselves do a good job making sure VB.NET works fine. The problem is weird functions in VB 4.
cleaning >50M lines of code? are you sure that wouldn't be expensive? and time consuming?
Yep lets say 200M. Lets say it costs $20 per line to cull. Ain't even close to what its costing Microsoft to keep old stuff and bad hardware working.
You train your users as to expectations. GNU/Linux's unwillingness to offer an easy platform for binary developers has made it very clear to people that binary applications won't just work out of the box. Therefore the expectation is OK source compatibility with very low binary compatibility. Apple has convinced developers that each OS release will require them to check it and possibly offer a patch.
Microsoft and Sun basically promise that if you get it running once it will very likely keep running
Slashdot Mirror
At the same time that IE was the default browser and far and away the most used browser for Windows. Further IE 5 for Mac was a better browser than IE 6 for Windows. Microsoft's mac division often writes better software since Microsoft can treat that platform as a test platform.
Those are two different issues.
/mnt/CD or where ever your CD mounts to is either already added or can be added with no problems, assuming you consider this a feature. Mandriva which is the home distribution I'm most familiar with ships with this on.
1) Does the distribution provide a GUI for the package manager?
2) Does the package manager allow an install from CD.
The distributions that provide a GUI for most things provide good quality GUIs for package management. So I'd assume if you want one and pick the right kind of distribution the answer to (1) is yes. But if you are scared of the command line why are you using a Linux?
As for (2) every package manager I know of allows you to set directories as sources for files. So
Microsoft with NT claimed to be writing a secure, enterprise OS. They didn't claim it was a home OS which could run useful business software.
-- Crappy drivers should simply fail. Or they could take on driver responsibility directly, drivers go through a Microsoft QA process. Apple has bad 3rd party apps break all the time.
-- The security model that was originally created for NT is excellent. Microsoft didn't implement it in their own apps. The crappy 3rd party programs shouldn't have worked
-- Microsoft has for a long time maintained a "we will get it to run" attitude. The result is developers engage in bad practices
The fact is there is no reason an ignorant home user should be able to install a rootkit even if they tried. Capability architecture is far far safer than the permissions architecture of the Unixes; Windows security should have blown Unix security out of the water.
Microsoft has business practices that are technically illegal. Anti trust law isn't enforced the way it was in the 1970s. Microsoft has avoided engaging in easy solutions that would get it out of anti trust problems. For example:
-- A simple uniform OEM windows price is public and available to all OEMs. That removes the whole argument about Microsoft "intimidating" vendors
-- Clear policies about 3rd party modifications to Windows, that apply regardless of component or corporate strategic direction (i.e. then bundling doesn't become a strategic move)
-- Making their office suite available for multiple platforms thus diminishing the argument they are using these two monopolies to reinforce one another
etc...
Sony conversely during the rootkit fiasco within days removed the rootkit. They didn't fight this in court for years and years.
No it just means that language isn't algebra. A word may not mean what its component parts mean and an expression may not mean what the combination of words mean. The rules for expressions and grammar are complicated.
It depends on which European country but customer service standards can often be much higher in Europe than in America. Think about the range of customer service standards in various types of venues in different countries or even within countries (Paris vs. Rivera). I don't know if you can make comparisons across countries like that for this reason. In the home computing market America basically has one level of service with a pretty narrow range. In clothing for example we have a huge range in service standards. In restaurants we have a huge range (though not as large as the range in Europe).
Although if anyone does know of a Debian+KDE distro that's stable, well documented, and polished, I'm in the market.)
Ah you get the king of polished. Though you sound a little advanced for Xandros. Another alternative which may be a better fit Mandriva.
Unless you mean marketing in a very expansive way it had very little do with marketing. Active X was at that time something like 30x faster than java. The active X gui was the windows GUI which looked much better than the Java 1.1 GUIs.
If Microsoft's lawyers couldn't get the case thrown out, what chance does a free browser have?
Much higher. Take a look at Groklaw and the information they dug up for IBM. The open source community can do open source research and find prior art. They can find employees who implemented the technologies that will site where they got the ideas from. Further they don't have anything to hide so they discovery will be public. I'm not sure what Eolas claims was their unique idea but I know of specific quotes (some Bill Gates interesting enough) from the late 1980s which basically outline the notion of pluggins.
So worst case, which is a very big assumption, Firefox doesn't use plugins. Its open source it can do things like load binary libraries. Same with Konq and Seamonkey. Opera might be hurt but that's an easy change for everyone.
Yeah. I remember the days of gopher (like HTML but text based) when the content to crap ratio was 20:1 (in favor of content). I think .gifs were a net benefit but I'd pay to have a net free of all that stuff
I don't think it was a huge mistake from the get go. Microsoft has always advocated smart client interfaces. HTML is essentially a dumb terminal / mainframe like system. The ability to have full scale applications using all the power of the desktop distributed via the net was a great idea. Where they failed was:
1) The windows NT security model failed and instead became a user model. Had a full fledged capability model been in place the permissions on the browser could have been set much lower than the permission on the user. Active X should not have been supported at all for the Windows 98 family due to poor security model.
2) Microsoft marketing didn't explain to customers that active x was dangerous and should always be turned off except for trusted sites. The warnings when reducing permissions should have been much more harshly worded (and the active X should have had to be changed up front.
3) They should have made available low security active X users for untrusted apps.
This gives you a setup like this:
a) default for most sites is active x is off
b) for some sites active X is on but runs with very low privs
c) for other sites active x is on runs as you but with lower privs
d) for the most trusted sites active x is a normal app
I happen to agree with the position you are presenting in this post regarding point numbers. The person who brought up component upgrades was you Windows XP was a 1 point upgrade from 2000 Pro. Some would argue that it was also a minor upgrade. But what you fail to grasp is that the 10.2,10.3,10.4 is the "release" version which has nothing to do with the versioning of core components such as the kernel and frameworks. In Jaguar, the kernel was 6.x but in Tiger it is 8.x. The same thing applies for major frameworks .. I was arguing those kernel upgrades and additional functionalities were not major upgrades but point upgrades.
They weren't advertised as breaking APIs. In spite of that however apps do still have problems.
So anyway with respect to this post there is no disagreement.
Yes it probably was more like 3% for 10.2 -> 10.3. OTOH well more than 10% broke on 10.1 -> 10.2, I think 10% is about right for 10.3->10.4.... Anyway I agree that 10.3 -> 10.4 is pretty close to 2K to XP. But then again less broke for the 2K -> XP move which was my point.
I don't disagree the open model with recompilation is the better model. What I am point out, is that if you are simply comparing apples to apples (binary compatibility) then
1) Microsoft takes this very seriously
2) They do this with non apps not produced by Microsoft
For example Maxis's Sim City had a whole separate memory subsystem in NT 4.0 to get it to work properly.
As for the point of slowness I don't agree with the NY times article. Here is my position
1) Binary compatibility increases Microsoft's development costs massively. This results in much longer product release cycles and bug fix cycles
2) It doesn't slow down the OS too much (say 30%)
3) The real damage is done because
a) Microsoft does not want to leave behind low end desktops. Longhorn as originally conceived would have been a much better product but would have required a $2k + machine to run
b) Security and convenience cannot be achieved without changing application developer behavior and that requires Microsoft putting some pressure on developers (i.e. breaking naughty apps).
Sorry I stand by my point that OS 8 -> OS 9 -> OS 10.0 were much bigger changes than 10.2 -> 10.3 -> 10.4
Between Windows 98 and Windows 2k the entire kernel was replaced, the entire programs model was replaced most interfaces were replaced the administrative system was replaced the whole driver model was changed. Windows 98 literally has almost the same relationship to 2k that classic does to OSX. Its a much bigger change.
10.0 -> 10.4 are evolutionary. Each one is a step forward. Regardless I'm not arguing 10% break between 10.3.4 and 10.3.5 but those changes happen every few months.
We are discussing binary compatibility. This would be the debian equivalent of using options like ignore-breakage, not having apt resolve dependencies issues, etc... I agree with you regarding the Linux model being a much better one in terms of control (and problem resolution). But it doesn't even try to offer binary compatibility. You can't take a debian 1.3 .deb file force and install and expect it to work.
Jaguar to Tiger is a 2 point minor upgrade (10.2 to 10.4). As for sources just google for it you'll get lots of info (quite a bit on Apple's site) and for example Tiger Review: Incompatibilities and Workarounds, or updates required.
There are. You can have a situation where you can't sue unless you case is sort of pre-vetted by a judge. You see this in Europe and it would definitely be worth considering for the US. Incidentally you also have it at the criminal level.
OK I think we may be agreeing then. I was presenting Microsoft's position on business legacy hardware support. I agree that for home its much less important.
You are thinking home users, business users don't upgrade their whole infrastructure. So yes upgrades will be common. Corporate users is really the group that Microsoft is worried about regarding compatibility they are the ones with very high standards in this regard. Home users are pretty forgiving. As an aside, free printers are really really expensive per page; you are paying for them.
we're not talking about ms compatibility though, we're talking about culling code for legacy hardware & software. most of compatibility testing, etc. will be done with modern software/hardware i would have thought.
.NET works fine. The problem is weird functions in VB 4.
No most of the problems are with the older legacy code base. For example the developers themselves do a good job making sure VB
cleaning >50M lines of code? are you sure that wouldn't be expensive? and time consuming?
Yep lets say 200M. Lets say it costs $20 per line to cull. Ain't even close to what its costing Microsoft to keep old stuff and bad hardware working.
You train your users as to expectations. GNU/Linux's unwillingness to offer an easy platform for binary developers has made it very clear to people that binary applications won't just work out of the box. Therefore the expectation is OK source compatibility with very low binary compatibility. Apple has convinced developers that each OS release will require them to check it and possibly offer a patch.
Microsoft and Sun basically promise that if you get it running once it will very likely keep running
In terms of hardware: old scanners, old printers, old monitors, old speciality devices (especially old cards).