The difference between Solaris (and for that matter, Linux) and Mozilla is that you can take course in writing an OS. There is established litature. (And linux had something of a clear target). In theory, all CS graduates have OS developement experience. How many books on writing a web browser existed in 1998? Zero. How many people, anywhere, had experience writing a web browser? Probabaly less then 2000, half or more of them working on competitive browsers and thus not allowed to donate their time, even if they wanted to. What is the target, the goal, for developing a browser? The Mozilla folks dont even know, the existance of Firefox being proof enough that they were wrong for a while. (which is not to say that it is bad, but humanity simply does not have the experience to say that something significantly better will come along)
I wonder why that is... Is MS intentionally breaking OSS software? How would they know what to break?
As a thought, Firefox uses its own installer for Windows (or possibly a third party tool). Not.MSI format. On the Linux side, I wouldnt expect that, say, Anaconda (the RH/Fedora install/upgrate system) to properly deal with packages installed using.deb's, or especiall with things installed from source. How good is the MSI format... Is it good enough that it is reasonable for MS to not spend too much time on SP installs so that they can deal with unpackaged things? Or, has MS officialy or semi-officialy told app developers that MSI should be used, and if they dont, and a SP breaks their app, sucks to be you? Pragmaticly speeking those two things are the same.
To draw an analogy, 95 and NT had different dirver systems. 98 introduced WDM (windows driver model), though still supported 95 drivers as well. At the time, the plan was that the next relase of Windows would combine NT and 9x, and that that line would use WDM exclusvly, so harware manufactures should put out wdm's now. Since this new line was a way off (and took an additional cycle for the merge to happen), many hardware manufactures diddnt put in the effort. So when the drivers diddnt work for the new release (whenever it was that they dropped support for the old system, I happily left the Windows world before then), whose fault was that? MS breaking the old drivers, or the hardware companies for not making the new drivers? As much as it pains me to admit it, I think MS was right.
Maby all the tools that produce msi files cost money. Maby OSS developers are philosophically opposed to using msi. Maby they simply dont want to play the MS game. Maby this is not the problem at all.
Ok fair enough. How about:... if you want to do that exactly once, $GRAPHICALTOOL is a good idea. But if you possibly would ever do the same thing again, use $REALSYSTEM...
Using linux and using PivotTables are not exclusive. OOo has pivot tables. I dont know about charts.
But you dont understand what I said. I diddnt say: data analysis is useless. What I said was that if you are doing data analysis that could be solved with PivotTables, then you should use a different tool entirely. If you are stuck on a MS platform, that tool might be Access. I would probabaly build some web based system, with a db backend, plotting with GNUPlot, but thats just me.
After I posted that I fired up OOo, and saw what it did... A long time ago I did some Access work, so see it is very different. But they both do visual data manipulation. Pivot tables may be cool, but if you are doing that level of work in a spreadsheet, regardless of how good it is, you are using the wrong tool. (Not that Access would be the tool to use....). That said, Im the kinda guy who thinks a word processor should be used for word processing, and a desktop publishing system should be used for desktop publishing... Anything more complex then a letter should be done not in Word but Pagemaker.
Lets say that 90% of all modern UNIXes, and Linux, at the kernel level provide the same features. That is, only a small portion of a given system is unique. Some of that uniqueness is advanced features, and some is just unique, period. So 90% of the effort of maintaining an OS is duplication of what everyone else is doing... And no one cares about that 90% because it is the same everywhere.
Well, SGI has all but dropped their version of UNIX in favour of Linux. They took the 10% of Irix that was unique and cool, and ported it to linux. This allowed them to easily take their unique (and commercial) apps and port them to Linux. The now OSS SGI goodies are free for anyone, but they definitly still make money on them.. Some of them are paticularly usefull on very high end systems, like those SGI sells. Some of them allow Linux to better run high end graphics apps, like the ones SGI sells. Now SGI only has to wory about maintaining that 10% that is unique... And to a lesser extent since they are almost definitly getting outside contributions. The model works.
Back to Sun... As for having to compleatly redo all the work that Sun has done on that 10%, this may be the reality today, but it wasen't necessary. Had Sun embraced linux in 1998, they could have produced all their new and unique stuff in such a way to work with both Solaris and Linux. More work up front, but doable. Every day that goes by that Sun doesnt do this kind of thing makes the change harder. They have no one to blame but themselves for this.
Are you talking about sound PROduction, or sound REproduction?
If you are talking about production, what you say is true. Electric guitars sound "better" with tube amps, because thats how they sound. The player is not "distorting" the sound, the guitar+amp is the sound. A harmon mute "distorts" a trumpet sound, but when you are trying to make that sound, kick ass.
If you are talking about sound REproduction, bullshit. Discrete transistors distort the signal less, and you are trying to play back the recording AS CLOSE AS POSSIBLE TO HOW IT WAS RECORDED. Transistors will do that better then tubes, and have done so for decades. Tubes will fuck up the signal.
They have always been in/sbin; "System binaries", required before anything besides / is mounted. It is likely that either you had previously always logged in with root (as opposeed to su), did su -, or otherwise had your normal users parth including/sbin (which it shouldnt.
But thinking about it more, it is unlikely that you could emulate a mainframe at 10x the "speed", at least in reference to a payroll system. Sure you could do the math faster, but what about the IO? I cant find any benchmarks, but I suspect that it wasent untill at least SBus (on Sparcs) or MCA (on IBM PC servers) that microcomputer IO reached 1970 era mainframes.
That was a reflection on the difference between software and hardware.
But the unstated point is true: history has shown that sucuess and failures of both hardware and software systems have little to do with their capabilities, and their capabilities in relation to other things available. Its all about business and marketing. DEC in the 80s, Microsoft, the Super Foonly, OS/2, the Amiga, etc, etc, etc. All projects whose fate was unrelated to their quality.
First: Mozilla and Firefox are released under the Mozilla Public License. Second: the code is copyright Mozilla Foundation. The FSF has absolutly no standing concerning Mozilla.
It is quite acceptable to distribute a customized browser based on the Mozilla code. It is acceptable to include, as distinct components on the source level, proprietary components, and keep those proprietary components closed. The canonical example being the Netscape browser, which comes with the AOL IM component, and other proprietary extenstions. Since providing diffs from the base mozilla.org code qualify as releasing the changes to the MPLd components, you could release a compleatly customized mozilla based browser and only have to distribute a thousand or so line patch. There are other minor details, providing pointers to where the base code is, documenting the changes (which a diff does by itself), but not very much.
Thus it would be compleatly possible for Microsoft to use Mozilla as the core of the next version of IE. All that they would have to do is post some links to mozilla.org/src (whatever), and distribute a 1000 line patch. Microsoft gives away far more complex things on msdn.microsoft.com all the time.
Yep, using pam_auth_ncpfs to get a Windows desktop to login to a Novell server would be crazy amounts of work. Fortunatly, they have this thing called the "Novell Client" which is for windows. It takes all of about 5 minutes to install. Windows 3.11, Windows 95, and all later versions, are specificlly built to accept network client drivers, like say, those from Novell.
Or do you mean it is crazy amounts of work for Novell to write a widows client? I doubt it. For Windows, Novell currently has two clients, one for 95/98 and one for NT/2000/XP. The "log into the server" part is a minor component, trivial in comparision to every thing else it does, ZENWorks integration for one.
And this is not Windows NT or Netware 3.12 days. You login to the network not into a server.
Slashdot Mirror
*Bzzz* wrong. If you are in Europe you are not on a T1.
What else do you need?
The difference between Solaris (and for that matter, Linux) and Mozilla is that you can take course in writing an OS. There is established litature. (And linux had something of a clear target). In theory, all CS graduates have OS developement experience. How many books on writing a web browser existed in 1998? Zero. How many people, anywhere, had experience writing a web browser? Probabaly less then 2000, half or more of them working on competitive browsers and thus not allowed to donate their time, even if they wanted to. What is the target, the goal, for developing a browser? The Mozilla folks dont even know, the existance of Firefox being proof enough that they were wrong for a while. (which is not to say that it is bad, but humanity simply does not have the experience to say that something significantly better will come along)
I wonder why that is... Is MS intentionally breaking OSS software? How would they know what to break?
As a thought, Firefox uses its own installer for Windows (or possibly a third party tool). Not .MSI format. On the Linux side, I wouldnt expect that, say, Anaconda (the RH/Fedora install/upgrate system) to properly deal with packages installed using .deb's, or especiall with things installed from source. How good is the MSI format... Is it good enough that it is reasonable for MS to not spend too much time on SP installs so that they can deal with unpackaged things? Or, has MS officialy or semi-officialy told app developers that MSI should be used, and if they dont, and a SP breaks their app, sucks to be you? Pragmaticly speeking those two things are the same.
To draw an analogy, 95 and NT had different dirver systems. 98 introduced WDM (windows driver model), though still supported 95 drivers as well. At the time, the plan was that the next relase of Windows would combine NT and 9x, and that that line would use WDM exclusvly, so harware manufactures should put out wdm's now. Since this new line was a way off (and took an additional cycle for the merge to happen), many hardware manufactures diddnt put in the effort. So when the drivers diddnt work for the new release (whenever it was that they dropped support for the old system, I happily left the Windows world before then), whose fault was that? MS breaking the old drivers, or the hardware companies for not making the new drivers? As much as it pains me to admit it, I think MS was right.
Maby all the tools that produce msi files cost money. Maby OSS developers are philosophically opposed to using msi. Maby they simply dont want to play the MS game. Maby this is not the problem at all.
Ok fair enough. How about: ... if you want to do that exactly once, $GRAPHICALTOOL is a good idea. But if you possibly would ever do the same thing again, use $REALSYSTEM...
Using linux and using PivotTables are not exclusive. OOo has pivot tables. I dont know about charts.
But you dont understand what I said. I diddnt say: data analysis is useless. What I said was that if you are doing data analysis that could be solved with PivotTables, then you should use a different tool entirely. If you are stuck on a MS platform, that tool might be Access. I would probabaly build some web based system, with a db backend, plotting with GNUPlot, but thats just me.
How else do you explain that Camryn Manheim is a mother?
After I posted that I fired up OOo, and saw what it did... A long time ago I did some Access work, so see it is very different. But they both do visual data manipulation. Pivot tables may be cool, but if you are doing that level of work in a spreadsheet, regardless of how good it is, you are using the wrong tool. (Not that Access would be the tool to use....). That said, Im the kinda guy who thinks a word processor should be used for word processing, and a desktop publishing system should be used for desktop publishing... Anything more complex then a letter should be done not in Word but Pagemaker.
I just read over the Wikipedia article. Is this not what Access is? Conceptually that is...
SGI is taken very seriously as a provider of high end graphic systems.
Lets say that 90% of all modern UNIXes, and Linux, at the kernel level provide the same features. That is, only a small portion of a given system is unique. Some of that uniqueness is advanced features, and some is just unique, period. So 90% of the effort of maintaining an OS is duplication of what everyone else is doing... And no one cares about that 90% because it is the same everywhere.
Well, SGI has all but dropped their version of UNIX in favour of Linux. They took the 10% of Irix that was unique and cool, and ported it to linux. This allowed them to easily take their unique (and commercial) apps and port them to Linux. The now OSS SGI goodies are free for anyone, but they definitly still make money on them.. Some of them are paticularly usefull on very high end systems, like those SGI sells. Some of them allow Linux to better run high end graphics apps, like the ones SGI sells. Now SGI only has to wory about maintaining that 10% that is unique... And to a lesser extent since they are almost definitly getting outside contributions. The model works.
Back to Sun... As for having to compleatly redo all the work that Sun has done on that 10%, this may be the reality today, but it wasen't necessary. Had Sun embraced linux in 1998, they could have produced all their new and unique stuff in such a way to work with both Solaris and Linux. More work up front, but doable. Every day that goes by that Sun doesnt do this kind of thing makes the change harder. They have no one to blame but themselves for this.
I still have a vacuum tube TV. For that matter, I have a vacuum tube computer monitor.
Are you talking about sound PROduction, or sound REproduction?
If you are talking about production, what you say is true. Electric guitars sound "better" with tube amps, because thats how they sound. The player is not "distorting" the sound, the guitar+amp is the sound. A harmon mute "distorts" a trumpet sound, but when you are trying to make that sound, kick ass.
If you are talking about sound REproduction, bullshit. Discrete transistors distort the signal less, and you are trying to play back the recording AS CLOSE AS POSSIBLE TO HOW IT WAS RECORDED. Transistors will do that better then tubes, and have done so for decades. Tubes will fuck up the signal.
Either those distros mangle the path ignoring the spirit of the Filesystem Hierarchy Standard, or they ignore the standard all together.
http://www.pathname.com/fhs/pub/fhs-2.3.html#SBINS YSTEMBINARIES
They have always been in /sbin; "System binaries", required before anything besides / is mounted. It is likely that either you had previously always logged in with root (as opposeed to su), did su -, or otherwise had your normal users parth including /sbin (which it shouldnt.
Indeed.
But thinking about it more, it is unlikely that you could emulate a mainframe at 10x the "speed", at least in reference to a payroll system. Sure you could do the math faster, but what about the IO? I cant find any benchmarks, but I suspect that it wasent untill at least SBus (on Sparcs) or MCA (on IBM PC servers) that microcomputer IO reached 1970 era mainframes.
You cant run a mainframe emulator that will have as good uptime as the real thing on anything less then a mainframe.
Please tell me where you purchased a book that was so good it went back in time 20 years and changed you perception of that topic.
That was a reflection on the difference between software and hardware.
But the unstated point is true: history has shown that sucuess and failures of both hardware and software systems have little to do with their capabilities, and their capabilities in relation to other things available. Its all about business and marketing. DEC in the 80s, Microsoft, the Super Foonly, OS/2, the Amiga, etc, etc, etc. All projects whose fate was unrelated to their quality.
Maby what they should do is have each computer sold have a "region" code burnt into it... shit; nevermind
Uh, because LDAP would be accessable anywhere, whereas network-mounted home directories would only be accessable from a smaller number of places.
First: Mozilla and Firefox are released under the Mozilla Public License. Second: the code is copyright Mozilla Foundation. The FSF has absolutly no standing concerning Mozilla.
It is quite acceptable to distribute a customized browser based on the Mozilla code. It is acceptable to include, as distinct components on the source level, proprietary components, and keep those proprietary components closed. The canonical example being the Netscape browser, which comes with the AOL IM component, and other proprietary extenstions. Since providing diffs from the base mozilla.org code qualify as releasing the changes to the MPLd components, you could release a compleatly customized mozilla based browser and only have to distribute a thousand or so line patch. There are other minor details, providing pointers to where the base code is, documenting the changes (which a diff does by itself), but not very much.
Thus it would be compleatly possible for Microsoft to use Mozilla as the core of the next version of IE. All that they would have to do is post some links to mozilla.org/src (whatever), and distribute a 1000 line patch. Microsoft gives away far more complex things on msdn.microsoft.com all the time.
That may be true, but that 20% is all profit.
Thats the media kit. IE, CDs and manuals.
Yep, using pam_auth_ncpfs to get a Windows desktop to login to a Novell server would be crazy amounts of work. Fortunatly, they have this thing called the "Novell Client" which is for windows. It takes all of about 5 minutes to install. Windows 3.11, Windows 95, and all later versions, are specificlly built to accept network client drivers, like say, those from Novell.
Or do you mean it is crazy amounts of work for Novell to write a widows client? I doubt it. For Windows, Novell currently has two clients, one for 95/98 and one for NT/2000/XP. The "log into the server" part is a minor component, trivial in comparision to every thing else it does, ZENWorks integration for one.
And this is not Windows NT or Netware 3.12 days. You login to the network not into a server.