A good friend of mine is a network admin for a small campus. On a campus with 700 users, 90% of the traffic was being used by 2 users via Napster.
Now, this is just rediculous.
As I understand it, they are currently looking into setting up some bandwith throttles. However, in the mean time, they banned Napster. What it flat came down to was people using Napster were bringing their Internet access to a grinding halt, and it was this guy's responsibility to insure that the Internet connection could be used for academic related purposes.
It is worth noting that this is a campus that has never banned any other sites. I also suspect the ban will be lifted once bandwith throttling is in place. However, I firmly support any network admin doing what is necessary to prevent a handful of users from disrupting service for everyone.
Well, in my experience, Suns are simply not workstation machines. We use sun's for a lot of web work, and custom web apps... mainly using java servlets. With those, the Suns appear to outperform the Intel Linux boxes... but it's so hard to make a comparison that is worth while, because the architectures themselves are differant.
Aside from that, my main point was that Solaris is more suited to the server range Suns than SparcLinux. As far as the differnaces between a Sun Server and an x86 Linux Server, it would entirely depend on application, budget, and requirements.
Well, I'm probably going to get shot at for this, but I'll say it anyway...
Solaris is not a bad O.S. In fact, on a high performance Sun, Solaris probably is far better than Linux.
Of course, any blanket statement like that has to be qualified or it's useless. On lower end Suns (sparc 2 through sparc 20, Ultra 5's, etc), linux will probably run faster, and perhaps more stablely. However, on higher end Sun boxes, the Solaris multiproc code, the volume management software (extra package), and the years of experience Sun has with it's own hardware give it a distinct advantage over current day SparcLinux.
SparcLinux is relatively new, and from what I've seen, doesn't have some of the more robust features of Solaris. The day will come when Linux beats Solaris on it's own turf - the Linux kernel development team has already proven themselves as masters of performance tweaking, and the more mature features are in development now. However, for the time being, when I install a Sun Ultra Enterprise box, I install Solaris on it. And until SparcLinux has proven itself (as Linux x86 has, and is doing), I will continue to use Sun Solaris on my Sun machines.
After reading some of the responses, I'm going to slightly modify my stance...
Current wireless solutions are not enough, unless you happen to live in one of the major cities which are offering unlimited wireless access.
Though, I must also admit that things are improving greatly. It looks like my area may be getting such access sometime in the near future (next 6 - 18 months). The spotty coverage is still less than ideal, though.
As far as LAN wireless, those products exist and are pretty good already. My main concern was coverage while not at home or office - which is still an average of 5-6 hours a day, sometimes more.
Anyway, thank you for the pointers... AT&T doesn't offer it here yet, but there is a chance for someone to offer it in the future!
Personally, I find the amazing thing is that they didn't spin off the palm into a new company ages ago.
3Com's big name is in networking... no matter what else they do, they continue to be a network equipment supplier. The Palm, on the other hand, has little to do with 3com's popular image. In fact, some people probably are unaware of anything else 3com does... just work on the Palm.
By splitting the Palm away from the 3com name, the Palm is left to survive, and thrive, on it's own. The Palm doesn't have a huge impact on 3com's stock, because it is a side game. However, a company dedicated to producing the Palm would probably perform wonders in it's IPO... because the Palm is popular, works well, and generally has a name already.
Disclaimer: Hey, I could be wrong - I'm a geek, not a market analyst. Technically speaking, I'm way out of my depth.
The best solution I have seen so far is almost a nonsolution. Basically, a small recharging battery pack internal to the main wearable "box," so that you can swap external battery packs without a problem. The next step is simply to have several battery packs which you strap on according to the current demand... most of us have a "home station," or a backpack, or something that we carry around/can get to throughout the day. If you have 5 x 4 hour battery packs, and you swap them in and out every 3 hours or so, you can dramatically cut down the weight in batteries actually strapped to your person. This allows you to also have a large, daylong battery pack you use in situations where you won't be able to swap battery packs.
Granted, this is all second hand information - I'm building my wearable now, so I haven't had a chance to really figure out if this is more of a pain than what it's worth.
I've been researching my own wearable for a while now, and I see two main problems. The first is that I work in an office - going entirely gargoyle isn't an option. Basically, I need something that is inconspicuous. I can walk around with a box or 3 hooked to my hip... I'm a tech, I already walk around with pounds of hardware strapped to my belt. But covering the face with an eyepiece just isn't going to happen.
Anyway, that's not the worst problem... between a removable eyepiece and a hidden earphone w/ tts software (I have long hair, easy to hide an earphone), I think I have the unobtrusive bit down.
The real problem is internet access. To make full use of a wearable computer, it needs to be connected to the network. I should be able to real time monitor my servers, be alerted of email, read slashdot, send email, fix my servers, post to slashdot... all while riding in the car, on the train, in the restauraunt. Cell modems are simply too expensive.
And the current solutions aren't enough. The Palm network offers "web clipping" - which amounts to them sending you what they want you to see. Telnet? Not a chance. Sprint's wireless web service? It has possibilities, but with the metered usage they currently offer, it's not much better than a cell modem for price.
Anyway, I doubt anyone has a solution (short of start up your own telco/ISP). However, as near as I can tell, the worst barrier to making the most of mobile computing is the communications barrier.
Personally, I don't think that the "feature" they put in their software is a great sin. It collects information which they need to get paid for their efforts (even if/I/ don't think it's worth a penny, obviously they have customers who do). Yes, it can potentially be cross-referenced with legal identities. Of course, there are a lot of ways to do that now (cookies, web logs, etc).
What is deplorable is that they did not release such information to the people who downloaded the software.
If a company wants to produce software that monitors every keystroke I ever type on my computer, fine. If I want to use it, fine. However, I should be told before installing the software that such information will be collected.
If we are going to condem their actions, then let us condem them for their real crime. Collecting this information was not a crime. Collecting this information without the consent of their users is a crime, if not in a legal since, then certainly in a moral since.
I would expect the people here to understand this better than most. Software is never the issue, it's what's done with the software and in what manner that is the issue. The government wants to regulate crypto because it can be used for illegal purposes. The music and vidio industry want software and hardware that can reverse engineer/defeat copy protection to be illegal because it can be used for pirating. Yet, crypto allows private communication, e-commerce, and user identification that is desperately needed in a world that is rapidly becoming dependant on computer communications. And the same software and hardware that can be used to defeat copy protection can be used to help debug programs, burn CD archives of our work, and play DVD's on our linux boxes.
A tool is just that. A tool. However, someone who uses a crowbar to break into people's homes is a far cry from someone who uses a crowbar in the process of construction.
Please. Remember their crime. It's not the software, it's the lack of consent.
My question is, as we improve nanotechnology, eventually we hit a point where we can no longer go any smaller (the use of single molecules as gates, for example). After we reach the point where it is impossible to shrink circuits any smaller, where will we go from there? So far, the only work I've heard of in this direction is the "quantum computer," and that is still in the theoretical stage. Does anyone know of any other technologies that we may switch to when our current methods of chip design simply can not become any smaller?
Slashdot Mirror
A good friend of mine is a network admin for a small campus. On a campus with 700 users, 90% of the traffic was being used by 2 users via Napster.
Now, this is just rediculous.
As I understand it, they are currently looking into setting up some bandwith throttles. However, in the mean time, they banned Napster. What it flat came down to was people using Napster were bringing their Internet access to a grinding halt, and it was this guy's responsibility to insure that the Internet connection could be used for academic related purposes.
It is worth noting that this is a campus that has never banned any other sites. I also suspect the ban will be lifted once bandwith throttling is in place. However, I firmly support any network admin doing what is necessary to prevent a handful of users from disrupting service for everyone.
Well, in my experience, Suns are simply not workstation machines. We use sun's for a lot of web work, and custom web apps... mainly using java servlets. With those, the Suns appear to outperform the Intel Linux boxes... but it's so hard to make a comparison that is worth while, because the architectures themselves are differant.
Aside from that, my main point was that Solaris is more suited to the server range Suns than SparcLinux. As far as the differnaces between a Sun Server and an x86 Linux Server, it would entirely depend on application, budget, and requirements.
Well, I'm probably going to get shot at for this, but I'll say it anyway...
Solaris is not a bad O.S. In fact, on a high performance Sun, Solaris probably is far better than Linux.
Of course, any blanket statement like that has to be qualified or it's useless. On lower end Suns (sparc 2 through sparc 20, Ultra 5's, etc), linux will probably run faster, and perhaps more stablely. However, on higher end Sun boxes, the Solaris multiproc code, the volume management software (extra package), and the years of experience Sun has with it's own hardware give it a distinct advantage over current day SparcLinux.
SparcLinux is relatively new, and from what I've seen, doesn't have some of the more robust features of Solaris. The day will come when Linux beats Solaris on it's own turf - the Linux kernel development team has already proven themselves as masters of performance tweaking, and the more mature features are in development now. However, for the time being, when I install a Sun Ultra Enterprise box, I install Solaris on it. And until SparcLinux has proven itself (as Linux x86 has, and is doing), I will continue to use Sun Solaris on my Sun machines.
After reading some of the responses, I'm going to slightly modify my stance...
Current wireless solutions are not enough, unless you happen to live in one of the major cities which are offering unlimited wireless access.
Though, I must also admit that things are improving greatly. It looks like my area may be getting such access sometime in the near future (next 6 - 18 months). The spotty coverage is still less than ideal, though.
As far as LAN wireless, those products exist and are pretty good already. My main concern was coverage while not at home or office - which is still an average of 5-6 hours a day, sometimes more.
Anyway, thank you for the pointers... AT&T doesn't offer it here yet, but there is a chance for someone to offer it in the future!
Personally, I find the amazing thing is that they didn't spin off the palm into a new company ages ago.
3Com's big name is in networking... no matter what else they do, they continue to be a network equipment supplier. The Palm, on the other hand, has little to do with 3com's popular image. In fact, some people probably are unaware of anything else 3com does... just work on the Palm.
By splitting the Palm away from the 3com name, the Palm is left to survive, and thrive, on it's own. The Palm doesn't have a huge impact on 3com's stock, because it is a side game. However, a company dedicated to producing the Palm would probably perform wonders in it's IPO... because the Palm is popular, works well, and generally has a name already.
Disclaimer: Hey, I could be wrong - I'm a geek, not a market analyst. Technically speaking, I'm way out of my depth.
The best solution I have seen so far is almost a nonsolution. Basically, a small recharging battery pack internal to the main wearable "box," so that you can swap external battery packs without a problem. The next step is simply to have several battery packs which you strap on according to the current demand... most of us have a "home station," or a backpack, or something that we carry around/can get to throughout the day. If you have 5 x 4 hour battery packs, and you swap them in and out every 3 hours or so, you can dramatically cut down the weight in batteries actually strapped to your person. This allows you to also have a large, daylong battery pack you use in situations where you won't be able to swap battery packs.
Granted, this is all second hand information - I'm building my wearable now, so I haven't had a chance to really figure out if this is more of a pain than what it's worth.
I've been researching my own wearable for a while now, and I see two main problems. The first is that I work in an office - going entirely gargoyle isn't an option. Basically, I need something that is inconspicuous. I can walk around with a box or 3 hooked to my hip... I'm a tech, I already walk around with pounds of hardware strapped to my belt. But covering the face with an eyepiece just isn't going to happen.
Anyway, that's not the worst problem... between a removable eyepiece and a hidden earphone w/ tts software (I have long hair, easy to hide an earphone), I think I have the unobtrusive bit down.
The real problem is internet access. To make full use of a wearable computer, it needs to be connected to the network. I should be able to real time monitor my servers, be alerted of email, read slashdot, send email, fix my servers, post to slashdot... all while riding in the car, on the train, in the restauraunt. Cell modems are simply too expensive.
And the current solutions aren't enough. The Palm network offers "web clipping" - which amounts to them sending you what they want you to see. Telnet? Not a chance. Sprint's wireless web service? It has possibilities, but with the metered usage they currently offer, it's not much better than a cell modem for price.
Anyway, I doubt anyone has a solution (short of start up your own telco/ISP). However, as near as I can tell, the worst barrier to making the most of mobile computing is the communications barrier.
Personally, I don't think that the "feature" they put in their software is a great sin. It collects information which they need to get paid for their efforts (even if /I/ don't think it's worth a penny, obviously they have customers who do). Yes, it can potentially be cross-referenced with legal identities. Of course, there are a lot of ways to do that now (cookies, web logs, etc).
What is deplorable is that they did not release such information to the people who downloaded the software.
If a company wants to produce software that monitors every keystroke I ever type on my computer, fine. If I want to use it, fine. However, I should be told before installing the software that such information will be collected.
If we are going to condem their actions, then let us condem them for their real crime. Collecting this information was not a crime. Collecting this information without the consent of their users is a crime, if not in a legal since, then certainly in a moral since.
I would expect the people here to understand this better than most. Software is never the issue, it's what's done with the software and in what manner that is the issue. The government wants to regulate crypto because it can be used for illegal purposes. The music and vidio industry want software and hardware that can reverse engineer/defeat copy protection to be illegal because it can be used for pirating. Yet, crypto allows private communication, e-commerce, and user identification that is desperately needed in a world that is rapidly becoming dependant on computer communications. And the same software and hardware that can be used to defeat copy protection can be used to help debug programs, burn CD archives of our work, and play DVD's on our linux boxes.
A tool is just that. A tool. However, someone who uses a crowbar to break into people's homes is a far cry from someone who uses a crowbar in the process of construction.
Please. Remember their crime. It's not the software, it's the lack of consent.
My question is, as we improve nanotechnology, eventually we hit a point where we can no longer go any smaller (the use of single molecules as gates, for example). After we reach the point where it is impossible to shrink circuits any smaller, where will we go from there? So far, the only work I've heard of in this direction is the "quantum computer," and that is still in the theoretical stage. Does anyone know of any other technologies that we may switch to when our current methods of chip design simply can not become any smaller?