He would make that point to imply that less *human* resources are spent fixing bugs *after* production phase has rolled out. Linux, in fact any Open Source peer review, makes for very few lasting bugs (relatively speaking; compared to Windows of course) -- and these bugs are more easily fixed than in a Closed Source OS -- but it's about an order of magnitude more expensive in time and labor to fix something after it has been committed than before. (Thanks to my old CS 312 professor for that tidbit, AFAIK its true)
Not saying Linux has bugs, but that it's more costly to repair than prevention through stricter commit process. OTOH Linux has more hands and eyes than FreeBSD to fix it...
Multiple choices are good. Nothing should be mutually exclusive in the end-user market. I can use Linux; if I don't like it, I can use FreeBSD; if I don't like it, I can use HURD. If I don't like Windows, I can use Be or MacOS or anything...
More than one integrated desktop is good; more than 10 is even better, because it allows us to make all the choices we want...
Re:Software Suite for Ancient PCs:Non-Profit
on
High Tech Junk
·
· Score: 1
Uh.... if it's all in asm you can run DEBUG.EXE to disassemble it.
Opcode-to-instruction translations are literal. It's the instruction-to-C (or whatever) that poses a problem, because different optimizing algorithms and compilers might make the same high-level construct into different assembly-level instructions.
In FreeBSD you can do something to the effects of 'options NO_KLM' (afaik) in the kernel configration and disable loadable modules anyway, after a requisite kernel recompile.
I have a feeling that MEEPT!! is really CmdrTaco's alter ego. Someone's got to prod the whole system from behind , now and then, to get things moving! Nothing does that like unpopular views and crypto-philosophic wisdom. Religion can never survive without a Devil to make it interesting. Slashdot without MEEPT!! is like... (warning: only kevin is going to know what I'm talking about).. Bynum without Scott...
Space versus Time. Optimization versus Portability.
Integrated chips:
1. Take more time to develop and make. 2. Are suited specifically to the task at hand. 3. Are not easily upgradeable. 4. Are not as powerful as they could be due to the limits imposed by small size and close proximity.
Distributed solutions (within the same chipset OR clustering solutions)
1. Take less time to make. 2. Are more flexible. 3. Are less optimized. 4. Have communications overhead between components. (Backpropagation? Crosspropagation? Whuzzat?) 5. Generally are more powerful. 6. Take up more space.
Trends have moved between integrating (wow, less overhead than those wacky multi-piece solutions!) and multi-piece solutions (wow, more powerful than that weak and non-upgradeable integrated solutions!) So the fanfare here won't last long I promise.
Although a beowulf of these things would have the best of both worlds... right?
However, trust can be built in someone who makes appealing assertions, but not necessarily the "best" (by the standards of those who believe him/her) assertions.
I agree with you wholeheartedly about the fact that follow ups get better and better. Presumably, the better postings shine through and the follow up distills itself to more "factual" quality.
I also lament that there is such high article turnover, because it doesnt allow your hypothetical follow up session to take effect.
Oh yeah, and uh (tongue-in-cheek) s/geodesic/curve:-P
Finally, that double major in Econ and Computer Science is worth the money!:)
Amazon's "generalization" of goods and services provided make sense from an economic perspective.
Consider if you will the perfect utopia of "Perfect Competition." In this nonexistent utopia all the firms would sell everything, and an equal amount of the population would go to each of these stores. In other words, if there were 5 firms, then 1/5 of the population would go to each. There would be no differences in price or quality, because each firm would be operating in perfect competition. There would be no premium for location, because travel overhead would be nil. And the products would be the same because information on said product would be free to all (think free speech and free beer.)
Because production overhead would be matched by marginal profits, and because various other costs (information, travel, etc) do not exist in this frictionless world, it would benefit each store to sell everything that can be sold, so that the other stores would not usurp the customer base. If I sold only Linux, FreeBSD and Windows NT CD's, my competitor can sell Linux, FreeBSD, Windows NT, *and* BeOS CD's, and thus usurp my market base (while keeping his). The increased customers from my store will make up for the cost of stocking extra copies of Linux, Windows and FreeBSD.
Of course this is never going to happen in the real world, because we do have travel expenses, and information is not perfect and free. Generalization has usually been restricted to areas which have a need of it, e.g, inner city corner stores where mobility within the block is free (walking) but mobility outside is not, or spread-out midwestern towns with a Wal-mart, but the next town or mall is 20 miles away.
On the Internet, however, the travel restrictions go down exponentially (although bandwidth is a concern, you can theoretically roar through on a text-only connection), and information becomes cheaper to produce (although quality assurance is another story entirely). Slopping up a web page takes no paper, you don't have to pay for a print ad, etc etc.
By this logic (which has gaps, because I don't think I have time to quote my books nor does anyone have time to read them, myself included =P ), it is befitting for Amazon or other on-line firms to generalize.
What would happen if Amazon only sold books and E-bay decided to sell books too? What would stop E-bay from stealing Amazon's customers? E-bay could as easily stock the books that Amazon stocks from the increased ex-Amazon revenue alone. And the cost of developing a front-end is negligible: same on-line engine, it's a sunk cost and it's nearly free to run a search-and-replace function to replace "Auctioning Stupid Computer Parts" to "Selling Stupid Books".
The Internet is the closest thing to Perfect Competition that anyone has seen, and the only thing holding it back is a Quality Assurance system (it's easier to lie over the net).
Perhaps we get the Internet ISO 9000 certified? heh.
The one thing about this post which redeems itself as being not a flame is the following line:
/* That's the way it is. Don't like that? Go work towards changing it. Change is good. But till things change, what I say will hold true. */
So this isn't FUD because it's open to the idea that Linux will improve over time.
The fact is Linux is, for lower end machines, excellent. Linux on an Intel against Solaris on a Sparc of comparable speed, for lower end stations wins hands down in ease of administration and (with the porting of Oracle, etc.) gaining rapidly in number of applications.
But until Linux develops a journaling file system (a real one: think AIX, not NT) or more scalability in SMP, clustering and HA (think VMS, not Janus), it will not take over the datacenter, which is where the real money and durability is.
That having been said, I disagree with the poster as to the "inevitablity" of it happening -- I think Linux, *BSD, etc will improve because you can't stop the desire to make free software better. Plus Free Software never will go out of business, by definition.
Give Linux 2-5 years to develop good HA and clustering. Give it 5-7 years more to get a reputation to compete with AIX, UNICOS, and OpenVMS in the datacenter.
The entire Dell "Shopping Cart" idea *IS* dynamic. Change an option, click a button, new price on same.asp.
I've done some ASP work and it's pretty memory intensive. Kudos to Dell for making it work -- it's slow as molasses sometimes, but it's never been down in my experience.
Again, I am not an NT expert, but as of IIS 4 Microsoft was stating its intention to move IIS in-kernel (cf. the Halloween Documents). IIS 5 which comes with Windows 2000 Beta 3 Advanced Server is a kernel-level module.
Microsoft likes to stick things in kernel space to get better performance; GDI (graphics drivers), IIS (services), etc.
Linux and most Unix like to stick them in user space for increased reliability; if something crashes then it takes down the App, not the box.
These are, IMO, mutually exclusive goals (mutexes, even!). If something takes down GDI, the kernel very often follows, in Windows. But Windows graphics operations are substantially faster and interactive than X operations.
How many times has your Windows box frozen but you could still move the mouse? That's because GDI still has highest priority scheduling.
For an extra $19 I believe you can get Advanced Server as well.
I've been playing with many aspects of Advanced Server, namely Terminal Server and Clustering. Interestingly enough -- *both* can't be installed at the same time on a system. Wouldn't a terminal server be something you want to have redundancy and failover on?
*sigh* microshit....
I have to say, and bear in mind I am a rabid FreeBSD/Linux/NeXTstep(Mac OS X) freak... I'm very happy with Win2000. At least it gives you the option to turn off all the eye candy and tweak performance in ways that 95/98/NT 4 didn't do...
Active Directory is about as confusing a rat's maze as you'd think it is.
DNS service is blah. IIS is *okay* performance wise.
Apparently the system manages its idle cycles OK, cause the Win2k machine cracks RC5 OK.
Zurk: Bear in mind that which is *doable* and that which is *comfortable* may be different.
I've run NT 4.0 *server* on a p100 with 32 megs of RAM. It RAN. It didn't do much. It wasn't fast. It didn't crash, but it didn't perform well.
I've run Win98 Beta on a p75 with 16 megs of RAM. It RAN. It didn't do much. Quake II was comical to play. (I've since added RAM for a total of 72 megs, and it runs fine).
Win2000 will RUN on a p133. Comfortably, maybe not. Professional(Workstation) and Advanced Server chugged along nicely on my pII 350/128M RAM. But I do believe it will RUN, albeit slowly, on sub-200 machines.
IIS integrates into the kernel. This is to overcome the high overhead of spawning new processes in user-space when the web browser receives many hits in, say, a CGI. This is touted as a "Feature" by Microsoft -- because IIS gets kernel-level scheduling, it's going to perform better.
Here's the catch.
AFAIK, and I am no NT expert, ISAPI DLL's are loaded into the kernel space. This as opposed to perl, DBI, CGI and other *user-space* modules in Apache. A bad module would bring down apache in a user-space environment, but a bad *kernel* module can cause a kernel panic (read: Blue Screen)
A Linux analogue would be an out-of-bounds exception for, say, the NFS kernel module (I am no expert, but that *should* cause problems)
As of Beta 3 *Advanced Server*, the requirements were a Pentium 133 with 64 megs of RAM (128 Recommended!) 2 GB of Hard drive space. This is what the manual/CD tell me.
I'd probably say it will run on a p100 with a lot more RAM (as evidence, I have Win98 Beta running on a p75 with 72 megs RAM at a decent clip, even though retail pads it and says it should run on a p90 or more)
Yes, you can buy your own copy of Workstation, Server and Advanced Server on Microsoft's web site, so you don't have to take my word for it. Yes, Win2K will run on your machine.
Slashdot Mirror
He would make that point to imply that less *human* resources are spent fixing bugs *after* production phase has rolled out. Linux, in fact any Open Source peer review, makes for very few lasting bugs (relatively speaking; compared to Windows of course) -- and these bugs are more easily fixed than in a Closed Source OS -- but it's about an order of magnitude more expensive in time and labor to fix something after it has been committed than before. (Thanks to my old CS 312 professor for that tidbit, AFAIK its true)
Not saying Linux has bugs, but that it's more costly to repair than prevention through stricter commit process. OTOH Linux has more hands and eyes than FreeBSD to fix it...
I think that sexual molesters have, by their actions, waived their rights to being offended.
as long as it will be a threat to my kids, 'it' is what it shall be.
Because he or she doesn't like it?
Multiple choices are good. Nothing should be mutually exclusive in the end-user market. I can use Linux; if I don't like it, I can use FreeBSD; if I don't like it, I can use HURD. If I don't like Windows, I can use Be or MacOS or anything...
More than one integrated desktop is good; more than 10 is even better, because it allows us to make all the choices we want...
Uh....
if it's all in asm you can run DEBUG.EXE to disassemble it.
Opcode-to-instruction translations are literal. It's the instruction-to-C (or whatever) that poses a problem, because different optimizing algorithms and compilers might make the same high-level construct into different assembly-level instructions.
In FreeBSD you can do something to the effects of 'options NO_KLM' (afaik) in the kernel configration and disable loadable modules anyway, after a requisite kernel recompile.
/* I'd rather see a sql than a remake. */
CREATE OR REPLACE VIEW new_tron AS
SELECT bm.old_tron,
bm.the_matrix,
bm.lawnmower_man,
bt.reboot,
bt.power_rangers,
bt.beast_wars,
FROM bad_movies bm, bad_tv_shows bt;
:-)
This really should be continued in e-mail, but... nah, got another major to work on.
mike
mxfara@wm.edu
I have a feeling that MEEPT!! is really CmdrTaco's alter ego. Someone's got to prod the whole system from behind , now and then, to get things moving! Nothing does that like unpopular views and crypto-philosophic wisdom. Religion can never survive without a Devil to make it interesting. Slashdot without MEEPT!! is like... (warning: only kevin is going to know what I'm talking about) .. Bynum without Scott...
Mike
mxfara@wm.edu
Space versus Time.
Optimization versus Portability.
Integrated chips:
1. Take more time to develop and make.
2. Are suited specifically to the task at hand.
3. Are not easily upgradeable.
4. Are not as powerful as they could be due to the limits imposed by small size and close proximity.
Distributed solutions (within the same chipset OR clustering solutions)
1. Take less time to make.
2. Are more flexible.
3. Are less optimized.
4. Have communications overhead between components. (Backpropagation? Crosspropagation? Whuzzat?)
5. Generally are more powerful.
6. Take up more space.
Trends have moved between integrating (wow, less overhead than those wacky multi-piece solutions!) and multi-piece solutions (wow, more powerful than that weak and non-upgradeable integrated solutions!) So the fanfare here won't last long I promise.
Although a beowulf of these things would have the best of both worlds... right?
Good point, I agree.
However, trust can be built in someone who makes appealing assertions, but not necessarily the "best" (by the standards of those who believe him/her) assertions.
Remember the hype behind Cold Fusion?
What about the Jonestown Massacre?
I agree with you wholeheartedly about the fact that follow ups get better and better. Presumably, the better postings shine through and the follow up distills itself to more "factual" quality.
:-P
I also lament that there is such high article turnover, because it doesnt allow your hypothetical follow up session to take effect.
Oh yeah, and uh (tongue-in-cheek) s/geodesic/curve
Finally, that double major in Econ and Computer Science is worth the money! :)
Amazon's "generalization" of goods and services provided make sense from an economic perspective.
Consider if you will the perfect utopia of "Perfect Competition." In this nonexistent utopia all the firms would sell everything, and an equal amount of the population would go to each of these stores. In other words, if there were 5 firms, then 1/5 of the population would go to each. There would be no differences in price or quality, because each firm would be operating in perfect competition. There would be no premium for location, because travel overhead would be nil. And the products would be the same because information on said product would be free to all (think free speech and free beer.)
Because production overhead would be matched by marginal profits, and because various other costs (information, travel, etc) do not exist in this frictionless world, it would benefit each store to sell everything that can be sold, so that the other stores would not usurp the customer base. If I sold only Linux, FreeBSD and Windows NT CD's, my competitor can sell Linux, FreeBSD, Windows NT, *and* BeOS CD's, and thus usurp my market base (while keeping his). The increased customers from my store will make up for the cost of stocking extra copies of Linux, Windows and FreeBSD.
Of course this is never going to happen in the real world, because we do have travel expenses, and information is not perfect and free. Generalization has usually been restricted to areas which have a need of it, e.g, inner city corner stores where mobility within the block is free (walking) but mobility outside is not, or spread-out midwestern towns with a Wal-mart, but the next town or mall is 20 miles away.
On the Internet, however, the travel restrictions go down exponentially (although bandwidth is a concern, you can theoretically roar through on a text-only connection), and information becomes cheaper to produce (although quality assurance is another story entirely). Slopping up a web page takes no paper, you don't have to pay for a print ad, etc etc.
By this logic (which has gaps, because I don't think I have time to quote my books nor does anyone have time to read them, myself included =P ), it is befitting for Amazon or other on-line firms to generalize.
What would happen if Amazon only sold books and E-bay decided to sell books too? What would stop E-bay from stealing Amazon's customers? E-bay could as easily stock the books that Amazon stocks from the increased ex-Amazon revenue alone. And the cost of developing a front-end is negligible: same on-line engine, it's a sunk cost and it's nearly free to run a search-and-replace function to replace "Auctioning Stupid Computer Parts" to "Selling Stupid Books".
The Internet is the closest thing to Perfect Competition that anyone has seen, and the only thing holding it back is a Quality Assurance system (it's easier to lie over the net).
Perhaps we get the Internet ISO 9000 certified? heh.
me
mxfara@not^a^chance^spam^breath.wm.edu
The one thing about this post which redeems itself as being not a flame is the following line:
/* That's the way it is. Don't like that? Go work towards changing it. Change is good. But till things change, what I say will hold true. */
So this isn't FUD because it's open to the idea that Linux will improve over time.
The fact is Linux is, for lower end machines, excellent. Linux on an Intel against Solaris on a Sparc of comparable speed, for lower end stations wins hands down in ease of administration and (with the porting of Oracle, etc.) gaining rapidly in number of applications.
But until Linux develops a journaling file system (a real one: think AIX, not NT) or more scalability in SMP, clustering and HA (think VMS, not Janus), it will not take over the datacenter, which is where the real money and durability is.
That having been said, I disagree with the poster as to the "inevitablity" of it happening -- I think Linux, *BSD, etc will improve because you can't stop the desire to make free software better. Plus Free Software never will go out of business, by definition.
Give Linux 2-5 years to develop good HA and clustering. Give it 5-7 years more to get a reputation to compete with AIX, UNICOS, and OpenVMS in the datacenter.
The inevitable reference to self in the third person is MEEPT!! esque,
however MEEPT!! is funny. And off-topic he isn't -- (s)he just puts a new spin on things.
You're not MEEPT!! so don't even try.
The entire Dell "Shopping Cart" idea *IS* dynamic. Change an option, click a button, new price on same .asp.
I've done some ASP work and it's pretty memory intensive. Kudos to Dell for making it work -- it's slow as molasses sometimes, but it's never been down in my experience.
In addition: If you don't believe me try taking a Dr. Watson snapshot of the system.
Again, I am not an NT expert, but as of IIS 4 Microsoft was stating its intention to move IIS in-kernel (cf. the Halloween Documents). IIS 5 which comes with Windows 2000 Beta 3 Advanced Server is a kernel-level module.
Your KDE woes can be summed up in two words:
"user-space."
Microsoft likes to stick things in kernel space to get better performance; GDI (graphics drivers), IIS (services), etc.
Linux and most Unix like to stick them in user space for increased reliability; if something crashes then it takes down the App, not the box.
These are, IMO, mutually exclusive goals (mutexes, even!). If something takes down GDI, the kernel very often follows, in Windows. But Windows graphics operations are substantially faster and interactive than X operations.
How many times has your Windows box frozen but you could still move the mouse? That's because GDI still has highest priority scheduling.
For an extra $19 I believe you can get Advanced Server as well.
I've been playing with many aspects of Advanced Server, namely Terminal Server and Clustering. Interestingly enough -- *both* can't be installed at the same time on a system. Wouldn't a terminal server be something you want to have redundancy and failover on?
*sigh* microshit....
I have to say, and bear in mind I am a rabid FreeBSD/Linux/NeXTstep(Mac OS X) freak... I'm very happy with Win2000. At least it gives you the option to turn off all the eye candy and tweak performance in ways that 95/98/NT 4 didn't do...
Active Directory is about as confusing a rat's maze as you'd think it is.
DNS service is blah. IIS is *okay* performance wise.
Apparently the system manages its idle cycles OK, cause the Win2k machine cracks RC5 OK.
Zurk: Bear in mind that which is *doable* and that which is *comfortable* may be different.
I've run NT 4.0 *server* on a p100 with 32 megs of RAM. It RAN. It didn't do much. It wasn't fast. It didn't crash, but it didn't perform well.
I've run Win98 Beta on a p75 with 16 megs of RAM. It RAN. It didn't do much. Quake II was comical to play. (I've since added RAM for a total of 72 megs, and it runs fine).
Win2000 will RUN on a p133. Comfortably, maybe not. Professional(Workstation) and Advanced Server chugged along nicely on my pII 350/128M RAM. But I do believe it will RUN, albeit slowly, on sub-200 machines.
IIS integrates into the kernel. This is to overcome the high overhead of spawning new processes in user-space when the web browser receives many hits in, say, a CGI. This is touted as a "Feature" by Microsoft -- because IIS gets kernel-level scheduling, it's going to perform better.
Here's the catch.
AFAIK, and I am no NT expert, ISAPI DLL's are loaded into the kernel space. This as opposed to perl, DBI, CGI and other *user-space* modules in Apache. A bad module would bring down apache in a user-space environment, but a bad *kernel* module can cause a kernel panic (read: Blue Screen)
A Linux analogue would be an out-of-bounds exception for, say, the NFS kernel module (I am no expert, but that *should* cause problems)
Haven't tried. Should assume so. Win2k has ISAPI which means yes, the server is tied into the kernel.
There are some 35 million lines of code in Win2k. Approximately 85% (Microsoft Figure: Let's go ahead and say 75%) is NEW code.
As of Beta 3 *Advanced Server*, the requirements were a Pentium 133 with 64 megs of RAM (128 Recommended!) 2 GB of Hard drive space. This is what the manual/CD tell me.
I'd probably say it will run on a p100 with a lot more RAM (as evidence, I have Win98 Beta running on a p75 with 72 megs RAM at a decent clip, even though retail pads it and says it should run on a p90 or more)
Yes, you can buy your own copy of Workstation, Server and Advanced Server on Microsoft's web site, so you don't have to take my word for it. Yes, Win2K will run on your machine.