Being able to hear the TV is just a matter of having good (not alien, just good) hearing.
The scanline transducer operates 525 times per refresh, 30 times a second: 15.750khz. This is within the range of human hearing, so if a transducer is of poor quality, you can hear it.
My guess is that many people just don't pay attention to it, even though they can hear it.
However lots of people who grew up listening to Walkmans and other devices can't hear it, because they've destroyed their high-range hearing.
Something to consider: in California right now, electricity runs $0.10 to $0.25 per kw/hour. That means the cost per 100 watts of 24/7 computer equipment is between $7 and $23 per month. Easy ouch.
Next, don't be a cooling idiot. If it's cold outside and your server room is hot, use the server room to warm the rest of the house. Air circulation. Central placement of server room in basement.
Actually there is no significant difference between JDK118 and J2ME. Minor security classes.
The major addition to J2SE was Swing, which nobody in their right mind would use on a PDA. Unfortunately most RADs only emit Swing code. However for PDA you're talking hand-written small code (AWT).
The speed does not scale, it's slower than a P3/200. The SH chip does not have all the fancy pipelining and superscalar architecture of the P3, and also the compiler sucks. It is a simple CPI with a simple compiler. Figure P3 ~100-150 equivalent.
You don't need to deal with NSI. You can transfer your domain name to another registrar and they can move it from NSI themselves. Expect them to honor the remaining time you have with NSI as well.
This is hardly something new. Interpreted typeless languages work very well for small programs, especially those which address system integration issues, i.e. mostly string processing, not much complexity, short execution times, few error conditions, and much calling of external programs for specific actions. Statically typed languages, which are generally compiled, are what you want for very large self-contained or homogenous systems.
The idea of buying a tiny computer and sticking it in the audio cabinet (or whatnot) is tempting. But one has to consider the electrical costs of doing so:
I have one box which handles my dsl, nat, samba service, etc. As soon as I get a long high quality audio cable, it will run my mp3 on the stereo as well, and I won't have to pay $10/month to have a second computer dedicated to the job.
Crappy material on napster was a problem before people starting intentionally putting up bogus material. The problem is that there is no rating system: either for users or material. Napster lacks a reputation manager.
A reputation manager solves the problem: material can be rated by people after they've downloaded it. The system could track exact duplicates, so the they would inherit (and share) the original's rating - i.e. it would be a distributed rating system, just as the content is distributed. And the rating would be influenced by the repution of the person doing the rating. This catches bogus files, truncated files, mis-named files, low-quality files, etc.
People should be rated as well. People who's material is rated highly would themselves get a high rating, and future material posted by them would have a rating influenced by their poster. Low-rated people would include newcomers, and trouble-makers (intentional or technologically incompetent).
The problem with Iridium was that their satellites only had a 5+ year lifetime, of which 2+ is now used. So a buyer was looking at a lifetime of 3 years before they had to start replacing the entire constellation (88 or so?). Not very smart.
Any anyway, as someone else has pointed out, Iridium will have to deorbit them at the end of their lifetime or before, presumably at company shutdown time.
Just because Sun isn't obviously likely to make revenue, one shouldn't conclude that an OSS license would be Ok for them. With an OSS license they would lose long-term control of the software.
Sun has agendas to push besides a denial-of-income attack on Microsoft. They are very keen on server-based computing, and even say that future versions of StarOffice will feature "portal" features. In order to get people to use these features, there needs to be a single upgrade path - the Sun path.
The SCSL thing really has more to do with superdistribution. Making the source available enables anyone to port it, add minor improvements, bug fixes, but not to distribute their version widely. Sun can incorporate any changes into their sourcebase, or not. The license is similar to Netscape's in that regard, in that all source code changes are subject to their oversight, in that unapproved changes won't go into the main distribution.
Sun has no incentive at all to go with an OSS license. If that was the case then anyone could bypass their future portal "features" (server tracking of users, push media/advertising/pages) by using a different distribution. Sun would much rather than people used this version, liked it, and upgraded to a future version with a mixture of client-based and server-based features: anything that increases server dependencies is good for Sun. The ability of the user base to do bug fixes and add features is just cake.
>What matters is that Netscape attempted to make >the web proprietary with their Netscape only tags
Not entirely. Realize that the existing browser at the time was Mosaic, which was crap (we'd written it and were in a position to know). Marca saw the W3 as a very slow, worthless board, and we decided to "make" the standard. The reasons were mostly to improve the abilities of HTML to the point where it could compare to what AOL could do (ironic isn't it!). For instance before tables, we had "columns" which could automatically wrap text into tables! Great feature, although nixed before release. But all the new tags were there to create a browser which could be sold to ISPs who wanted to make an AOL-competitive content site. The plan really was to superdistribute the browser and make money off the server, so the browser needed to be the best (par none), and also have support for stuff in the server (cookies, SSL). Marca *did* want to dominate the world, and often commented that if we were under investigation by the DOJ then that'd be a sign of success:) But none of that early innovation(!) stuff was proprietary.
I don't really know about "smart browsing" but again I'd say it was an extension of searching-in-the-url-field, a friendly feature you might expect AOL to have, rather than a proprierty protocol push.
I found it amusing that MS was winning the browser war partly through it's bundling and all that, but partly by writing a better browser - NOT a typical MS approach. So Netscape acted somewhat Microsoftian and Microsoft acted, well, more normal. An IE3 download was better than an NS4 download. Of course when they started bundling with Windows w/ IE3&4 they went full tilt to monopoly abuse, but that's different.
My credentials, btw: employee #14 at Netscape, even before jamie. NCSA before that. And long since departed.
Slashdot Mirror
Being able to hear the TV is just a matter of having good (not alien, just good) hearing.
The scanline transducer operates 525 times per refresh, 30 times a second: 15.750khz. This is within the range of human hearing, so if a transducer is of poor quality, you can hear it.
My guess is that many people just don't pay attention to it, even though they can hear it.
However lots of people who grew up listening to Walkmans and other devices can't hear it, because they've destroyed their high-range hearing.
Something to consider: in California right now, electricity runs $0.10 to $0.25 per kw/hour. That means the cost per 100 watts of 24/7 computer equipment is between $7 and $23 per month. Easy ouch.
Next, don't be a cooling idiot. If it's cold outside and your server room is hot, use the server room to warm the rest of the house. Air circulation. Central placement of server room in basement.
Actually there is no significant difference between JDK118 and J2ME. Minor security classes.
The major addition to J2SE was Swing, which nobody in their right mind would use on a PDA. Unfortunately most RADs only emit Swing code. However for PDA you're talking hand-written small code (AWT).
The speed does not scale, it's slower than a P3/200. The SH chip does not have all the fancy pipelining and superscalar architecture of the P3, and also the compiler sucks. It is a simple CPI with a simple compiler. Figure P3 ~100-150 equivalent.
You don't need to deal with NSI. You can transfer your domain name to another registrar and they can move it from NSI themselves. Expect them to honor the remaining time you have with NSI as well.
With IP aliasing it's not a problem; one card could service multiple servers, each with their own (set of) IPs.
This is hardly something new. Interpreted typeless languages work very well for small programs, especially those which address system integration issues, i.e. mostly string processing, not much complexity, short execution times, few error conditions, and much calling of external programs for specific actions. Statically typed languages, which are generally compiled, are what you want for very large self-contained or homogenous systems.
The idea of buying a tiny computer and sticking it in the audio cabinet (or whatnot) is tempting. But one has to consider the electrical costs of doing so:
145W * 24hrs = 3.4kwhours/day
3.4kwhrs * $0.10 = $0.34/day
$0.34/day * 365days = $127/year
I have one box which handles my dsl, nat, samba service, etc. As soon as I get a long high quality audio cable, it will run my mp3 on the stereo as well, and I won't have to pay $10/month to have a second computer dedicated to the job.
A reputation manager solves the problem: material can be rated by people after they've downloaded it. The system could track exact duplicates, so the they would inherit (and share) the original's rating - i.e. it would be a distributed rating system, just as the content is distributed. And the rating would be influenced by the repution of the person doing the rating. This catches bogus files, truncated files, mis-named files, low-quality files, etc.
People should be rated as well. People who's material is rated highly would themselves get a high rating, and future material posted by them would have a rating influenced by their poster. Low-rated people would include newcomers, and trouble-makers (intentional or technologically incompetent).
Jacob Nielsen has written about reputation managers. See his article at http://www.useit.com/alertbox/990905.html .
The problem with Iridium was that their satellites only had a 5+ year lifetime, of which 2+ is now used. So a buyer was looking at a lifetime of 3 years before they had to start replacing the entire constellation (88 or so?). Not very smart.
Any anyway, as someone else has pointed out, Iridium will have to deorbit them at the end of their lifetime or before, presumably at company shutdown time.
Just because Sun isn't obviously likely to make revenue, one shouldn't conclude that an OSS license would be Ok for them. With an OSS license they would lose long-term control of the software.
Sun has agendas to push besides a denial-of-income attack on Microsoft. They are very keen on server-based computing, and even say that future versions of StarOffice will feature "portal" features. In order to get people to use these features, there needs to be a single upgrade path - the Sun path.
The SCSL thing really has more to do with superdistribution. Making the source available enables anyone to port it, add minor improvements, bug fixes, but not to distribute their version widely. Sun can incorporate any changes into their sourcebase, or not. The license is similar to Netscape's in that regard, in that all source code changes are subject to their oversight, in that unapproved changes won't go into the main distribution.
Sun has no incentive at all to go with an OSS license. If that was the case then anyone could bypass their future portal "features" (server tracking of users, push media/advertising/pages) by using a different distribution. Sun would much rather than people used this version, liked it, and upgraded to a future version with a mixture of client-based and server-based features: anything that increases server dependencies is good for Sun. The ability of the user base to do bug fixes and add features is just cake.
-mark
>What matters is that Netscape attempted to make
:) But none of that early innovation(!) stuff was proprietary.
>the web proprietary with their Netscape only tags
Not entirely. Realize that the existing browser at the time was Mosaic, which was crap (we'd written it and were in a position to know). Marca saw the W3 as a very slow, worthless board, and we decided to "make" the standard. The reasons were mostly to improve the abilities of HTML to the point where it could compare to what AOL could do (ironic isn't it!). For instance before tables, we had "columns" which could automatically wrap text into tables! Great feature, although nixed before release. But all the new tags were there to create a browser which could be sold to ISPs who wanted to make an AOL-competitive content site. The plan really was to superdistribute the browser and make money off the server, so the browser needed to be the best (par none), and also have support for stuff in the server (cookies, SSL). Marca *did* want to dominate the world, and often commented that if we were under investigation by the DOJ then that'd be a sign of success
I don't really know about "smart browsing" but again I'd say it was an extension of searching-in-the-url-field, a friendly feature you might expect AOL to have, rather than a proprierty protocol push.
I found it amusing that MS was winning the browser war partly through it's bundling and all that, but partly by writing a better browser - NOT a typical MS approach. So Netscape acted somewhat Microsoftian and Microsoft acted, well, more normal. An IE3 download was better than an NS4 download. Of course when they started bundling with Windows w/ IE3&4 they went full tilt to monopoly abuse, but that's different.
My credentials, btw: employee #14 at Netscape, even before jamie. NCSA before that. And long since departed.
~mark