Um, the free software people have very philosophical attitudes, but open-source software people are usually more practical. I appreciate the work of RMS, as I assume most people do. I'm not quite as ideological, but I've become more so over time. And I wasn't talking about the community as a whole. I was talking about the particular action of offering such software to someone simply as an alternative, not as salvation. Besides, most of these people won't have heard of RMS or any of this--they'll just see a slightly less buggy alternative to a troublesome, buggy piece of software, and they'll consider if it fits their needs.
I really don't like the idea of trying to convert people.
Why not? All they're trying to do is share something that they think is useful and good. It's entirely different from trying to convert someone religiously--it's not like we're saying "If you don't use this software instead, you're going to HELL!!!". It's just saying "Hey, in case you're interested, there is software that's free, useful, and developed cooperatively for the fun of it, instead of by companies." There's nothing wrong with it.
I know it can be awkward sometimes, and I'm not really comfortable just shoving it into people's faces, but there's nothing wrong with offering it.
I find it interesting that people readily accept the notion that Pentium M is a derivative of the PIII while Core 2 is somehow distinct from Core or Pentium M... To claim that Core 2 is the first "Core architecture" product is arbitrary and more absurd than claiming that Pentium M is a new version of PIII, yet that seems to be what's floated around here.
I've seen this information in dozens of articles, and the writers of those articles usually get it from Intel. The Pentium M is a derivative of the Pentium III. It's a very large improvement, but the core is still much more heavily based on the Pentium III than the Pentium 4. And the "Core microarchitecture" is a particular microarchitecture that Intel has talked about. Just like going from the Pentium III/4 to the Pentium M, it's got a bunch of improvements, and, yes, it is loosely based on what worked in the previous Pentium M, just as the Pentium M was based on the good parts of the Pentium III/4.
I think you misread it. but that's easy to do. Decide for yourself:
Website: So how does the mobile version of Core 2 Duo (Merom) actually differ from the desktop version (Conroe)? Actually, the differences are relatively minor - though as it's essentially the same chip that's not really surprising.
You: So the Intel Core 2 Duo chip, also known as Merom was internally known as Conroe...
Website: (well, the website doesn't explicitly say this, but Dothan came after Banias)
You: and is based on the Pentium M 'Banias' mobile chip based out of Haifa.
Website: Rather confusingly though, Core Duo, which is Yonah, is not actually Core architecture - it's was essentially a dual-core version of Pentium M.
You: Additionally, the original Core Duo (Yonah) isn't really a Core product, it's just a P4M with two cores.
Website: Core architecture, with its various improvements and enhancements, actually begins with the Core 2 Duo
You: Core structure actually starts with Core 2 Duo, thus the Core 2 identification?
So the Intel Core 2 Duo chip, also known as Merom was internally known as Conroe...
No. Merom and Conroe are the notebook and desktop versions, respectively, of the same chip. Merom and Conroe are both the internal names of the chips. Officially, they're both called "Core 2 Duo", and the model numbers distinguish the two series. There are physical differences, including FSB speed and (IIRC) cache architecture.
and is based on the Pentium M 'Banias' mobile chip based out of Haifa.
No. Core Duo is based on the Pentium M Dothan, which was an improved (more cache and higher FSB) version of Banias.
Additionally, the original Core Duo (Yonah) isn't really a Core product,
True. Props to Intel for the dumb naming.
it's just a P4M with two cores.
No! It's a Pentium M with two cores! Big difference! The Pentium 4-M is a pathetic, hot, power-hogging, slow version of the Pentium 4. The Pentium M is based mostly on the Pentium III, and was designed from the ground up to be more efficient per watt.
Core structure actually starts with Core 2 Duo, thus the Core 2 identification?
Yes. structure = microarchitecture.
While I agree, Intel's naming and branding sucks... try not to make it worse!
There is no license. So the users are entirely responsible themselves for the use of the software. It also means they can benefit the most.
It doesn't need to be public domain, just free. The patent laws only prevent someone from offering for sale, selling, or importing an invention. Potentially, this means that open-source software, released under any license, would work, as long as nobody charges money for it. Does this also mean that someone could create a nonprofit organization that accepted donations and used the money to manufacture generic versions of patented prescription medicines and distribute them for free to people who can't afford them? That's an interesting idea.
Even hotels charge you for internet access (especially business-oriented ones).
Not true. I just travelled around California looking at colleges, and all of the hotels--from the Country Inns & Suites down to the crappy Days Inn that the Residence Inn by Marriott sent us to after they were "overbooked by central reservations" (read: three guests decided not to leave, and instead of kicking them out, they denied a few customers their reservations) had free wireless internet. (The first one, in Claremont, had wired access, using what appeared to be an internal DSL network.) I would consider a hotel without wireless internet to be lacking.
The Core 1 CPUs were basically Pentium IIIs with extra instructions and much-revamped layout and FSB.
Actually, the Pentium M was basically a Pentium III with extra instructions and a revamped layout and FSB. The Core 1 CPU (I think Yonah, the laptop version, was the only Core 1 CPU; they have started adding the desktop and server versions with Core 2) is basically a dual-core Pentium M, with some more features. Merom is basically a beefed up version of Yonah, the major features being: it's supposedly 64-bit, it's got twice the L2 cache on some models, it can do 128-bit SSE instructions in 1 cycle instead of 2, and it's got some minor improvements in in performance (larger reorder buffer, more issue ports, etc...).
Didn't Intel have a partial 64-bit system where the system could access more than 4GB of memory but the registers were still 32-bit? (It added layers to the page tables.) Is Merom really fully 64-bit, just like an Athlon 64 X2, or can it just access more memory?
I'm curious... in the articles I have read about Core Duo and Core 2 Duo (Yonah and Merom, at least on the notebook end of things), I never saw anyone mention 64-bit support as some amazing new feature. I heard a few mentions of it in the early stages of speculation, but now that it's the eve of the release, nobody's making a fuss about it. Usually, when a new 64-bit processor is coming out, it's a big deal. So is Merom actually 64-bit, or did that part get scrapped, or what?
I'd like 64-bit, since x86_64/amd64/whatever_it_is_called gives me double the registers each with double the bits, which is cool. (Of course, then I have to deal with all the issues that a 64-bit arch has, like not being able to interact directly with 32-bit codec DLL's from Windows and such...)
First, although they do announce that you shouldn't use WiFi, they have no way of knowing. And second, most Windows users don't know how to disable, or don't bother disabling, their WiFi. Every time I've been on a plane, if I fire up Kismet or tcpdump (with the card in rfmon mode, so it doesn't transmit anything itself), I see a handful of Windows machines chattering, either probing for the last AP they were associated with or beaconing for the ad-hoc network they created because they couldn't find the AP. (That feature, by the way, is really stupid. If I wanted an ad-hoc network, I'd make one.)
I reviewed the hp LaserJet 1320 on my web site. In short, it's cheap (maybe not under $300, but definitely under $400, and often discounted on Newegg), has awesome text quality and very good graphics quality, prints relatively quickly, duplexes (an uncommon feature in such a cheap printer!), and conserves toner (I haven't replaced the cartridge yet, in several years of use.
Do you know where to buy ancient phones and accompanying service? I'd be amused to have a bag phone or an OKI (they're really cool old hackable analog phones--remote controllable with DTMF, reprogrammable to display a list of other calls on the cell and let you select one to listen to, computer controllable, etc...), but I don't know where to buy them, and I don't know how to get service for them. (The cheapest plans most companies have now are around $40-$50.) Thanks for any pointers you can give.
Imagine the Plinko board on The Price Is Right. the last 25% of the board (the EBD) could be equipped with gates, that according to current core / threads bings incoming jobs to a less heavily burdened core?
(/me doesn't get the The Price Is Right reference, but...) This is already done by the OS scheduler. It tries to keep an equal load on each core. But my issue is that the application doesn't know how many threads to spawn in the first place, and shouldn't need to know. It should only need to know how many tasks it has to do, and the OS/processor/libraries should work together to decide how many parallel threads should be running to get those tasks done.
BTW, I've posted a slightly bigger version of my original comment on my website: Thoughts on Multithreading.
I think one of the issues in multithreading is knowing when and how much to enable it. If, for example, I'm writing a mail client, and I want to have t worker threads processing messages as they're retrieved. How do I find an optimum value for t? I could use the CPU count, but the user might be running other programs. I could just pick a number, say 4, and let any extras get scheduled, but imagine if every program took this approach--we'd have 2-4 times too many threads (given a single or dual core CPU), and that would seriously hurt throughput (constant rescheduling) not to mention memory usage. So we're stuck with manually configuring every app to use multiple threads. I can rattle off a bunch of apps that do this: apache, mysql, spamd (spamassassin), blender, and more. Even when the processes are suitable for easy multithreading (handling email is inherently threadable, because messages are independent of each other), we still don't automatically multithread, because there's no way to tell if we should and how many threads to make. Perhaps there needs to be an OS implementation of workqueues (not the kernel kind), where you can queue up n jobs, and the kernel will spawn as many threads as is "ideal", and run the jobs in those threads, one by one.
People with millions generally invest their money, not just leave it all in the bank. Because if you're not using it right now, might as well have it doing other useful things for you. Kind of like memory;)
I'm investing it in a disk cache fund.;-) It does benefit performance!
I've never understood why people with 1-2GB of RAM freak out when applications actually use some of that available memory. What good is a ton of memory if it's not being used?
I have only a measly 512MB of memory, but I usually freak out if even half of it is being used by applications. I run a relatively slim desktop, and it uses around 128MB with nothing open. The reason I like to conserve memory is that Linux uses all unused memory as disk cache, so if I have 512MB of memory and only use 128MB, I have 384MB of disk cache, which speeds things up a lot. I do, of course, like having the memory when I'm running Azureus or the GIMP or something else that uses a lot of RAM.
I look at this the other way. Why waste RAM just because you have a lot? If you thought about it in financial terms, what you said is like saying "I've never understood why people with millions of dollars just leave it in the bank!" I'd rather have free RAM speeding up disk access than have it full of bloated applications.
Why can't I reply to this comment anonymously? I get an error "You are not allowed to post to this page", unlike anything I've seen on Slashdot before!
This really pisses me off. I'm so sick of reading newspaper articles that read something like this:
YoYoDyne, Inc. has created a new revolutionary product, a so-called "widget", which "frobs" and "fiddles" with so-called "gizmos".
...where all of the quoted terms are legitimate technical terms. If I turned the tables, and wrote a letter to the editor, saying:
I found the "article" published in the so-called "News" section of your "newspaper" to be quite interesting.
...you know that they would be annoyed, because the quotes and the "so-called" make it sound like the term is not really what it's called, and that it's not really true. If writers are concerned that a reader doesn't know a term, there's no point in putting it in quotes to reassure the dumb reader that they're not dumb. It's much more helpful to write something like this:
YoYoDyne, Inc. has created a new revolutionary product, a widget (a small gadget used to modify gizmos) which frobs (gently adjusts) and fiddles (adjusts more aggressively) with gizmos (common elements of world-domination machines).
Sure, it's a little choppier, but good writers can weave things together better (I could if I weren't lazy and I wasn't posting on Slashdot), and this form provides much more knowledge. Frankly, reporters shouldn't be writing about stuff they really have no clue about. I think if someone's going to be writing about internet addresses, it isn't much to ask that someone explain the rudiments of bits and bytes and binary numbers to them before they run off and misinform the public.
I'm using VMware Server on Gentoo. I don't care about this.
But, assuming some people do, I'm not surprised about this. Everyone seems to be giving away a low-end virtualization or emulation program, because it's a great way to drum up business for the higher-end products, and it adds value to a platform.
No, it's kinda like discussing economics or physics during calculus to make it more practical and show people the real-world applications. Robots are an application of computer science; Bronte is not an application of math, but physics and economics are.
It's a matter of giving people more practical work, which is both more interesting and easier to learn for some people. I usually find that I learn a language better when I can play with it, and doubly so if I can write something real with it. Having a real piece of hardware that responds to your program is more exciting than just printing messages on a console.
I think they did it because someone had already used / as the option prefix (like in dir/w--the Linux equivalent is -, as in ls -l). So they had to use a different character.
No. The whole point of the article is that Skype purposefully intends to be invisible and sneaky. The reason is that it makes it easier to run Skype on firewalled and/or NATted networks, either at home or at work. Many home users have convoluted NAT setups, and most don't have the expertise (or reason) to poke holes in the firewall. Skype likes to advertise that it offers Internet phone service that "just works", so they need to make it work on every network. That may mean using random ports, using ports intended for other protocols, tunneling to remote servers or through peers, or other things that can be interpreted as resourceful or sneaky, depending on your point of view.
This is totally stupid. First, Google is not required to list any website, especially not those of competitors. That would be like suing the yellow pages company for not including a (free!) listing for a competitor's directory. (At least in MA, there are multiple companies selling directories. The phone companies distribute them for free, and then another company, Yellow Book, has started giving them out.) Second, this isn't free speech. Google is free to quash any company it feels like. The Webmaster Guidelines are very clear that anything "evil" won't be tolerated, and could result in delisting of a website. If this site is so great, why don't they buy AdSense ads? Then they can pick the keywords themselves and get whatever rank they want. By the way, 50% of the first two pages of results for "kinderstart" are about the lawsuit.:-b
Does anyone think it's worth Google bombing these guys with some clever phrase, like "sore losers" or "search scammers"?
Slashdot Mirror
Um, the free software people have very philosophical attitudes, but open-source software people are usually more practical. I appreciate the work of RMS, as I assume most people do. I'm not quite as ideological, but I've become more so over time. And I wasn't talking about the community as a whole. I was talking about the particular action of offering such software to someone simply as an alternative, not as salvation. Besides, most of these people won't have heard of RMS or any of this--they'll just see a slightly less buggy alternative to a troublesome, buggy piece of software, and they'll consider if it fits their needs.
Why not? All they're trying to do is share something that they think is useful and good. It's entirely different from trying to convert someone religiously--it's not like we're saying "If you don't use this software instead, you're going to HELL!!!". It's just saying "Hey, in case you're interested, there is software that's free, useful, and developed cooperatively for the fun of it, instead of by companies." There's nothing wrong with it.
I know it can be awkward sometimes, and I'm not really comfortable just shoving it into people's faces, but there's nothing wrong with offering it.
I've seen this information in dozens of articles, and the writers of those articles usually get it from Intel. The Pentium M is a derivative of the Pentium III. It's a very large improvement, but the core is still much more heavily based on the Pentium III than the Pentium 4. And the "Core microarchitecture" is a particular microarchitecture that Intel has talked about. Just like going from the Pentium III/4 to the Pentium M, it's got a bunch of improvements, and, yes, it is loosely based on what worked in the previous Pentium M, just as the Pentium M was based on the good parts of the Pentium III/4.
I think you misread it. but that's easy to do. Decide for yourself:
Website: So how does the mobile version of Core 2 Duo (Merom) actually differ from the desktop version (Conroe)? Actually, the differences are relatively minor - though as it's essentially the same chip that's not really surprising.
You: So the Intel Core 2 Duo chip, also known as Merom was internally known as Conroe...
Website: (well, the website doesn't explicitly say this, but Dothan came after Banias)
You: and is based on the Pentium M 'Banias' mobile chip based out of Haifa.
Website: Rather confusingly though, Core Duo, which is Yonah, is not actually Core architecture - it's was essentially a dual-core version of Pentium M.
You: Additionally, the original Core Duo (Yonah) isn't really a Core product, it's just a P4M with two cores.
Website: Core architecture, with its various improvements and enhancements, actually begins with the Core 2 Duo
You: Core structure actually starts with Core 2 Duo, thus the Core 2 identification?
You are entirely wrong.
No. Merom and Conroe are the notebook and desktop versions, respectively, of the same chip. Merom and Conroe are both the internal names of the chips. Officially, they're both called "Core 2 Duo", and the model numbers distinguish the two series. There are physical differences, including FSB speed and (IIRC) cache architecture.
No. Core Duo is based on the Pentium M Dothan, which was an improved (more cache and higher FSB) version of Banias.
True. Props to Intel for the dumb naming.
No! It's a Pentium M with two cores! Big difference! The Pentium 4-M is a pathetic, hot, power-hogging, slow version of the Pentium 4. The Pentium M is based mostly on the Pentium III, and was designed from the ground up to be more efficient per watt.
Yes. structure = microarchitecture.
While I agree, Intel's naming and branding sucks... try not to make it worse!
It doesn't need to be public domain, just free. The patent laws only prevent someone from offering for sale, selling, or importing an invention. Potentially, this means that open-source software, released under any license, would work, as long as nobody charges money for it. Does this also mean that someone could create a nonprofit organization that accepted donations and used the money to manufacture generic versions of patented prescription medicines and distribute them for free to people who can't afford them? That's an interesting idea.
Not true. I just travelled around California looking at colleges, and all of the hotels--from the Country Inns & Suites down to the crappy Days Inn that the Residence Inn by Marriott sent us to after they were "overbooked by central reservations" (read: three guests decided not to leave, and instead of kicking them out, they denied a few customers their reservations) had free wireless internet. (The first one, in Claremont, had wired access, using what appeared to be an internal DSL network.) I would consider a hotel without wireless internet to be lacking.
Actually, the Pentium M was basically a Pentium III with extra instructions and a revamped layout and FSB. The Core 1 CPU (I think Yonah, the laptop version, was the only Core 1 CPU; they have started adding the desktop and server versions with Core 2) is basically a dual-core Pentium M, with some more features. Merom is basically a beefed up version of Yonah, the major features being: it's supposedly 64-bit, it's got twice the L2 cache on some models, it can do 128-bit SSE instructions in 1 cycle instead of 2, and it's got some minor improvements in in performance (larger reorder buffer, more issue ports, etc...).
Didn't Intel have a partial 64-bit system where the system could access more than 4GB of memory but the registers were still 32-bit? (It added layers to the page tables.) Is Merom really fully 64-bit, just like an Athlon 64 X2, or can it just access more memory?
I'm curious... in the articles I have read about Core Duo and Core 2 Duo (Yonah and Merom, at least on the notebook end of things), I never saw anyone mention 64-bit support as some amazing new feature. I heard a few mentions of it in the early stages of speculation, but now that it's the eve of the release, nobody's making a fuss about it. Usually, when a new 64-bit processor is coming out, it's a big deal. So is Merom actually 64-bit, or did that part get scrapped, or what?
I'd like 64-bit, since x86_64/amd64/whatever_it_is_called gives me double the registers each with double the bits, which is cool. (Of course, then I have to deal with all the issues that a 64-bit arch has, like not being able to interact directly with 32-bit codec DLL's from Windows and such...)
First, although they do announce that you shouldn't use WiFi, they have no way of knowing. And second, most Windows users don't know how to disable, or don't bother disabling, their WiFi. Every time I've been on a plane, if I fire up Kismet or tcpdump (with the card in rfmon mode, so it doesn't transmit anything itself), I see a handful of Windows machines chattering, either probing for the last AP they were associated with or beaconing for the ad-hoc network they created because they couldn't find the AP. (That feature, by the way, is really stupid. If I wanted an ad-hoc network, I'd make one.)
I reviewed the hp LaserJet 1320 on my web site. In short, it's cheap (maybe not under $300, but definitely under $400, and often discounted on Newegg), has awesome text quality and very good graphics quality, prints relatively quickly, duplexes (an uncommon feature in such a cheap printer!), and conserves toner (I haven't replaced the cartridge yet, in several years of use.
Do you know where to buy ancient phones and accompanying service? I'd be amused to have a bag phone or an OKI (they're really cool old hackable analog phones--remote controllable with DTMF, reprogrammable to display a list of other calls on the cell and let you select one to listen to, computer controllable, etc...), but I don't know where to buy them, and I don't know how to get service for them. (The cheapest plans most companies have now are around $40-$50.) Thanks for any pointers you can give.
(/me doesn't get the The Price Is Right reference, but...) This is already done by the OS scheduler. It tries to keep an equal load on each core. But my issue is that the application doesn't know how many threads to spawn in the first place, and shouldn't need to know. It should only need to know how many tasks it has to do, and the OS/processor/libraries should work together to decide how many parallel threads should be running to get those tasks done.
BTW, I've posted a slightly bigger version of my original comment on my website: Thoughts on Multithreading.
I think one of the issues in multithreading is knowing when and how much to enable it. If, for example, I'm writing a mail client, and I want to have t worker threads processing messages as they're retrieved. How do I find an optimum value for t? I could use the CPU count, but the user might be running other programs. I could just pick a number, say 4, and let any extras get scheduled, but imagine if every program took this approach--we'd have 2-4 times too many threads (given a single or dual core CPU), and that would seriously hurt throughput (constant rescheduling) not to mention memory usage. So we're stuck with manually configuring every app to use multiple threads. I can rattle off a bunch of apps that do this: apache, mysql, spamd (spamassassin), blender, and more. Even when the processes are suitable for easy multithreading (handling email is inherently threadable, because messages are independent of each other), we still don't automatically multithread, because there's no way to tell if we should and how many threads to make. Perhaps there needs to be an OS implementation of workqueues (not the kernel kind), where you can queue up n jobs, and the kernel will spawn as many threads as is "ideal", and run the jobs in those threads, one by one.
I'm investing it in a disk cache fund. ;-) It does benefit performance!
I have only a measly 512MB of memory, but I usually freak out if even half of it is being used by applications. I run a relatively slim desktop, and it uses around 128MB with nothing open. The reason I like to conserve memory is that Linux uses all unused memory as disk cache, so if I have 512MB of memory and only use 128MB, I have 384MB of disk cache, which speeds things up a lot. I do, of course, like having the memory when I'm running Azureus or the GIMP or something else that uses a lot of RAM.
I look at this the other way. Why waste RAM just because you have a lot? If you thought about it in financial terms, what you said is like saying "I've never understood why people with millions of dollars just leave it in the bank!" I'd rather have free RAM speeding up disk access than have it full of bloated applications.
Why can't I reply to this comment anonymously? I get an error "You are not allowed to post to this page", unlike anything I've seen on Slashdot before!
This really pisses me off. I'm so sick of reading newspaper articles that read something like this:
...where all of the quoted terms are legitimate technical terms. If I turned the tables, and wrote a letter to the editor, saying:
...you know that they would be annoyed, because the quotes and the "so-called" make it sound like the term is not really what it's called, and that it's not really true. If writers are concerned that a reader doesn't know a term, there's no point in putting it in quotes to reassure the dumb reader that they're not dumb. It's much more helpful to write something like this:
Sure, it's a little choppier, but good writers can weave things together better (I could if I weren't lazy and I wasn't posting on Slashdot), and this form provides much more knowledge. Frankly, reporters shouldn't be writing about stuff they really have no clue about. I think if someone's going to be writing about internet addresses, it isn't much to ask that someone explain the rudiments of bits and bytes and binary numbers to them before they run off and misinform the public.
I'm using VMware Server on Gentoo. I don't care about this.
But, assuming some people do, I'm not surprised about this. Everyone seems to be giving away a low-end virtualization or emulation program, because it's a great way to drum up business for the higher-end products, and it adds value to a platform.
No, it's kinda like discussing economics or physics during calculus to make it more practical and show people the real-world applications. Robots are an application of computer science; Bronte is not an application of math, but physics and economics are.
It's a matter of giving people more practical work, which is both more interesting and easier to learn for some people. I usually find that I learn a language better when I can play with it, and doubly so if I can write something real with it. Having a real piece of hardware that responds to your program is more exciting than just printing messages on a console.
Even though gnucash.org is being crushed by Slashdot, the release is still available at SourceForge.
I think they did it because someone had already used / as the option prefix (like in dir /w--the Linux equivalent is -, as in ls -l). So they had to use a different character.
No. The whole point of the article is that Skype purposefully intends to be invisible and sneaky. The reason is that it makes it easier to run Skype on firewalled and/or NATted networks, either at home or at work. Many home users have convoluted NAT setups, and most don't have the expertise (or reason) to poke holes in the firewall. Skype likes to advertise that it offers Internet phone service that "just works", so they need to make it work on every network. That may mean using random ports, using ports intended for other protocols, tunneling to remote servers or through peers, or other things that can be interpreted as resourceful or sneaky, depending on your point of view.
This is totally stupid. First, Google is not required to list any website, especially not those of competitors. That would be like suing the yellow pages company for not including a (free!) listing for a competitor's directory. (At least in MA, there are multiple companies selling directories. The phone companies distribute them for free, and then another company, Yellow Book, has started giving them out.) Second, this isn't free speech. Google is free to quash any company it feels like. The Webmaster Guidelines are very clear that anything "evil" won't be tolerated, and could result in delisting of a website. If this site is so great, why don't they buy AdSense ads? Then they can pick the keywords themselves and get whatever rank they want. By the way, 50% of the first two pages of results for "kinderstart" are about the lawsuit. :-b
Does anyone think it's worth Google bombing these guys with some clever phrase, like "sore losers" or "search scammers"?