If Sun makes this move, it could be the worst mistake they've ever made.
Not sure if it would outrank Sun's purchase of Cobalt.
If Sun (actually Schwartz) really thinks that Novell/SuSe will be a growth market, they could still earn a scheissload of money by buying a stake in Novell (say 20% or so). That way, they won't be putting all of their cash reserves in one basket. One consideration in favor of an outright buyout is that Sun does have a close to absolute right to the UNIX IP (as one poster put it, SCO wouldn't have a legal leg to stand on).
IIRC, Sun's Java Desktop System is based on SuSe, so Sun may be trying to protect the basis for JDS (there are some considerations for JDS being Linux based rather than Solaris x86 based).
The first 64 bit version of Solaris was Solaris 7 (SunOS 5.7), with FCS (first customer ship) late 1998. Solaris 2.6, which was 32 bit with large file support, was the default install in 3Q98.
To get Solaris up and running on AMD64 first required porting their C compiler to AMD64 (don't think Solaris would compile with gcc). The next step was dealing with ifdef's that assumed 64bits=Sparc and writing AMD64 versions of the device drivers. The experience with Solaris for IA-32 should minimize the endian problem.
FCS for Solaris 10 has been rumored to be January 2005 - and AMD64 support is expected.
It would be really nice if Sun ports the new drivers written for AMD64 to Sparc.
Re:where are the IPv6 native ISPs?
on
IPv6 is Here
·
· Score: 2, Informative
Speakeasy has been making noises about rolling out IPv6 support and apparently have done experimenting with it. Right now it still looks to be in the "near future".
A possible compromise would be for the ISP's to offer IPv6 tunneling hosts.
In fact the shell account is included as part of their "OSDN" package - and for the Sysadmin package. Haven't played with it too much. One nice thing about a shell account is being able to test ports on my own machine from the outside - especially nice in the days of dial-up (i.e. before home networking).
Had a lot of experience with the shell account on CTSnet (now parting of Hosting.com - sigh) - allowed for very fast access to their newsserver, mail (my mail account was one the same host as the shell account), etc. During their heyday, CTS was one of the best ISP's around.
Yes if it is an 'unrembursed' business expense broadband, cell phone and pager you can write them off on your taxes.
Remember that 'unreimbursed' business expenses cannot be deducted from the AMT. If you're not close to the AMT threshold, then no worries.
Having said that, I would gues that the cost difference between say ISDN and DSL could be deducted since a major reason fo going ISDN is higher reliability.
I'd be curious to know if Microsoft (or Novell -- PC-DOS) owns copyrights or patents on any of the DOS APIs. They're all implemented as interrupt 21h calls I beleive (its been a while), so nothing really has a name (as far as trademarks or patents go). Curious none-the-less, though.
Considering that the original API's were based on CP/M, I don't think M$ would have much of a leg to stand on. In addition, Seattle Computer still retained some rights to DOS after it was sold to M$. Thirdly, back in the 80's, M$ allowed a couple of the big guys to sell their own versions - most notably Compaq with version 3.31 (first DOS to support more than 32 MB per logical partition).
ILEC: Incumbent Local Exchange Carrier, i.e. "The Phone Company", be it SBC, Verizon, etc.
CLEC: Competitive Local Exchange Carrier would be upstarts such as Covad, or long distance companies such as AT&T.
UNE-P: Unbundled Network Element - Platform, which is the twisted pair from the Central Office (CO) to the premises plus the switch, DSLAM or whatever to make it a live connection rpovided by the ILEC. Fiber To The Premises (FTTP) and Fiber To The Node (FTTN) would be covered under UNE-P rules - thus pretty much locking out outfits such as Covad from the very high speed DSL market.
UNE-L: Unbundled Network Element - Line, this is just the twisted pair from the CO to the premises, with the switch, DSLAM or whatever provided by the CLEC. Northpoint (RIP) typically used UNE-L's to avoid problems with line filters.
That same FCC decision, IIRC, also allowed the local baby-Bells to charge whatever they want for access to their networks by other carriers. That effectively *destroys* competition for last-mile service.
A little clarification - the FCC decision affects UNE-P access, but not UNE-L access. The difference is that UNE-L is just the twisted pair from the CO to the premises, UNE-P is where the CLEC would be using the ILEC's DSLAM or switch. Unfortunately the FTTP and FTTN would be covered by the UNE-P rules.
What if racists became the majority? I would hope the laws in place would prevent them from having the government censor ME. And that _can_ happen. See WW2 for more information.
WW1 was even worse (in the US). It was illegal to criticize the war effort and there were 100,000 people signed up to report anyone making negative comments about the war. Before any public event started, you had to listen to a spiel about how evil the Germans were (basically government mandated hate speech).
After the war, much of the propoganda was shown to be outright lies. This had two consequences, one is that the American public wanted no part in an European war (and people would have asked for FDR's impeachment if they knew how he was violating the Neutrality Act), second, when reports of the atrocities of the concentration camps first came out, they were thought to be the same kind of fabrication that went on in WW1.
I brought this up with someone who got her PhD on the chemistry of moon rocks (i.e. she knows what she's talking about) and she thought that it would be a non-starter.
The other way of getting He-3 is to make lots of tritium and let it decay into He-3. Main problem is keeping the tritium out of the hands of the bad guys while it decays.
With the ways things are going, I think we'll see solar cells at $1/watt way before fusion becomes practical.
Hmmm, the vented plasma will contain significant quanties of tritium. The exception would be using a D-He3 reaction - most scenarios for that assume that the He3 will come from mining the moon.
In addition to the tritium, you may be venting a fair amount of first wall material which will be radioactive due to neutron absorption (although a D-He3 reactor will have a couple orders of magnitude less neutron production).
Nothing horrendously bad, but not something orders of magnitude better than a fission reactor. FWIW, I think the dangers of properly designed fission reactors are overstated (RBMK's were not properly designed).
Watson and Crick did not discover DNA and the comment that this is the 50th anniversary of DNA's discovery is BS. Their Nobel prize came from discovering the structure of DNA.
DNA's role in passing along genetic information was discovered ten years earlier by Osgood Avery - who should have received a Nobel prize, but the committee was to timid to award him one.
(And a pound is currently defined as 0.45359237 kg.)
At what value of assumed gravitational acceleration?? (Pounds are a unit of force, not mass - make more sense in defing slugs in terms of kilograms).
This is my one nitpick with your otherwise excellent post. I especially liked your rant on temperature (bullet point #5) - have yet to see anyone come up with a valid reason why Celsius is inherently better than Fahrenheit other than being part of the SI standard.
If you're going to taut the advantages of your measurement system, shouldn't you maybe know a thing or two about it before you try? I can get much better accuracy in my native system by remembering that pure liquid H2O at around 75 degrees Farenheit and about 14.7 psia has a density of about 62.4 lb/(ft^3), and I'll avoid some of the potentially dangerous mistakes you're making.
Reminds me of another rule of thumb - at 70F and 100% RH, there's 1 pound (lbm not lbf) of water per 1,000 cubic feet of air.
One final rant - the US letter size is defined as 8.5 by 11 inches exactly - which works out to 215.9 by 279.4 mm (exactly) and good quality paper will be +/- 0.1 mm. The author of the website rounded the letter size to the nearest integral number of mm, which is just as sleazy as rounding the metric size to the nearest 1/32nd of an inch. While the root 2 aspect ratio of metric paper is nice for folding, there is no way that you can maintain an exact aspect ratio and have the size in an integral number of mm.
110V dates back to Edison's Pearl Street station in 1882.
60HZ vs 50Hz
Edison's Pearl Street station was DC - AC started to show up in the mid 1880's at quite a variety of frequencies (25Hz to 133Hz). There have been reports 60Hz was chosen as an American standard to be incompatible with the European 50Hz standard - not sure what diff that would make with the 110/220 differences. L.A. was on 50 Hz till the 1930's and the surrounding SCE territory was on 50Hz until 1948.
60Hz does have the advantage that transformers and motors can be made slightly smaller and lighter.
What's so awkward about centigrades? You start counting when water freezes and when the water boils you've reached 100 degrees. Simple, eh?
Water boils at 100C?? That's actually a very misleading statement. A more precise rendition is that the vapor pressure of water is equal to standard atmospheric pressure at 100C (actually just slightly less than 100C). In fact, water will boil at temperatures less than 0C, provided that the partial pressure of water is less than the vapor pressure (i.e. relative humidity is less than 100%).
For thermodynamic calculations, you would normally use absolute temperature, which is expressed in units of degrees Kelvin in SI or degrees Rankine in english engineering (R = 1.8 * K, exactly). Either way, you have to make use of lookup tables (e.g. steam tables) or some ugly conversion constant (e.g. Boltzmann's) to make use of temperature given in Celcius, Fahrenheit, Kelvin or Rankine.
Anyway, degrees Celsius is a big hairy wart on the metric system's butt.
Solaris ported to AMD64 will be a kick-ass combo - as it should be relatively easy to port 64 bit SPARC aps to the Opteron.
It can work the other way as well - 64 bits aps running on Solaris/AMD64 should be easy to port to Solaris/SPARC.
I'm figuring that things will pick up for Sun about the end of the year.
While I'm always glad to see progress on every front, gnuplot has been sitting on the 3.* level for a
long time.
I was using 3.7 back in early 1999 - so it took 5+ years to past 3.7.x. Nice to see an upgrade to a very useful program.
What I like best about gnuplot is that interactive mode and batch mode use the same commands - makes it really easy to write scripts and easy to write scripts that write scripts. One favorite was writing an awk script that would spit out a gnuplot script to plot 50+ files worth of raw and fitted data - and spit the output as a PostScript file.
Don't forget packet filtering. Solaris has no native firewall
Ummm - I think it is called "Sunscreen" or something to that effect. IIRC, it allows you the option of passing traffic to firewalled hosts without address translation.
If you just want to protect the host, then you can define secuirity associations for each port - including requiring IPSEC connections. Solaris has had native IPSEC support for several years (albeit with rather sucky documentation).
If a virus has a short incubation period and is very virulent (you die quickly) the less likely it will affect a large proportion of people.
The more successfull virus are the one will long incubation period, take the virus that case AIDS for example.
If you read the book The Great Influenza by John Barry, you would find that people are usually contagious several days before symptoms appear. In some cases death occurred very rapidly after the onset of the symptoms (people dying in the trolley), but they had been infected several days before.
Barry did a very good job researching the book, very well worth reading. Not sure of what scared me more - the virus or the political climate in the US during Wilson's second term as president.
I Am Not An OS/2 1.x expert, but my impression was that OS/2 1.x did make use of the rigid segmentation model imposed by the 286. Instead of malloc'ing out of a heap - OS/2 would give you another data segment whose length was requested in the malloc (N.B. this is a guess on my part). Any attempt to access memory outside of the permitted range in that segment would generate a protection fault.
I do recall reading that M$ programmers preferred developing under OS/2 1.X because it was so picky about memory protection.
Following the logic of the RIAA and MPAA, the entertainment companies should be responsible for whenever someone commits a crime based on what they saw/heard on a movie/recording/video game.
See recent cases involving GTA and Natural Born Killers...
Slashdot Mirror
In my experience, GTK still has a way to go to catch with Motif in respect to working smoothly.
Not sure if it would outrank Sun's purchase of Cobalt.
If Sun (actually Schwartz) really thinks that Novell/SuSe will be a growth market, they could still earn a scheissload of money by buying a stake in Novell (say 20% or so). That way, they won't be putting all of their cash reserves in one basket. One consideration in favor of an outright buyout is that Sun does have a close to absolute right to the UNIX IP (as one poster put it, SCO wouldn't have a legal leg to stand on).
IIRC, Sun's Java Desktop System is based on SuSe, so Sun may be trying to protect the basis for JDS (there are some considerations for JDS being Linux based rather than Solaris x86 based).
You obviously weren't watching TV in the 1960's.
While there were many shows on with zero violence then, there were others with considerable violence - Gunsmoke, Wild Wild West, Bonanza, Combat, etc.
To get Solaris up and running on AMD64 first required porting their C compiler to AMD64 (don't think Solaris would compile with gcc). The next step was dealing with ifdef's that assumed 64bits=Sparc and writing AMD64 versions of the device drivers. The experience with Solaris for IA-32 should minimize the endian problem.
FCS for Solaris 10 has been rumored to be January 2005 - and AMD64 support is expected. It would be really nice if Sun ports the new drivers written for AMD64 to Sparc.
A possible compromise would be for the ISP's to offer IPv6 tunneling hosts.
Had a lot of experience with the shell account on CTSnet (now parting of Hosting.com - sigh) - allowed for very fast access to their newsserver, mail (my mail account was one the same host as the shell account), etc. During their heyday, CTS was one of the best ISP's around.
Remember that 'unreimbursed' business expenses cannot be deducted from the AMT. If you're not close to the AMT threshold, then no worries.
Having said that, I would gues that the cost difference between say ISDN and DSL could be deducted since a major reason fo going ISDN is higher reliability.
Considering that the original API's were based on CP/M, I don't think M$ would have much of a leg to stand on. In addition, Seattle Computer still retained some rights to DOS after it was sold to M$. Thirdly, back in the 80's, M$ allowed a couple of the big guys to sell their own versions - most notably Compaq with version 3.31 (first DOS to support more than 32 MB per logical partition).
You're giving away Free Software's secret weapon.
My experience has also been that portable software is generally higher quality than non-portable software.
ILEC: Incumbent Local Exchange Carrier, i.e. "The Phone Company", be it SBC, Verizon, etc.
CLEC: Competitive Local Exchange Carrier would be upstarts such as Covad, or long distance companies such as AT&T.
UNE-P: Unbundled Network Element - Platform, which is the twisted pair from the Central Office (CO) to the premises plus the switch, DSLAM or whatever to make it a live connection rpovided by the ILEC. Fiber To The Premises (FTTP) and Fiber To The Node (FTTN) would be covered under UNE-P rules - thus pretty much locking out outfits such as Covad from the very high speed DSL market.
UNE-L: Unbundled Network Element - Line, this is just the twisted pair from the CO to the premises, with the switch, DSLAM or whatever provided by the CLEC. Northpoint (RIP) typically used UNE-L's to avoid problems with line filters.
A little clarification - the FCC decision affects UNE-P access, but not UNE-L access. The difference is that UNE-L is just the twisted pair from the CO to the premises, UNE-P is where the CLEC would be using the ILEC's DSLAM or switch. Unfortunately the FTTP and FTTN would be covered by the UNE-P rules.
WW1 was even worse (in the US). It was illegal to criticize the war effort and there were 100,000 people signed up to report anyone making negative comments about the war. Before any public event started, you had to listen to a spiel about how evil the Germans were (basically government mandated hate speech).
After the war, much of the propoganda was shown to be outright lies. This had two consequences, one is that the American public wanted no part in an European war (and people would have asked for FDR's impeachment if they knew how he was violating the Neutrality Act), second, when reports of the atrocities of the concentration camps first came out, they were thought to be the same kind of fabrication that went on in WW1.
I brought this up with someone who got her PhD on the chemistry of moon rocks (i.e. she knows what she's talking about) and she thought that it would be a non-starter.
The other way of getting He-3 is to make lots of tritium and let it decay into He-3. Main problem is keeping the tritium out of the hands of the bad guys while it decays.
With the ways things are going, I think we'll see solar cells at $1/watt way before fusion becomes practical.
Hmmm, the vented plasma will contain significant quanties of tritium. The exception would be using a D-He3 reaction - most scenarios for that assume that the He3 will come from mining the moon.
In addition to the tritium, you may be venting a fair amount of first wall material which will be radioactive due to neutron absorption (although a D-He3 reactor will have a couple orders of magnitude less neutron production).
Nothing horrendously bad, but not something orders of magnitude better than a fission reactor. FWIW, I think the dangers of properly designed fission reactors are overstated (RBMK's were not properly designed).
DNA's role in passing along genetic information was discovered ten years earlier by Osgood Avery - who should have received a Nobel prize, but the committee was to timid to award him one.
At what value of assumed gravitational acceleration?? (Pounds are a unit of force, not mass - make more sense in defing slugs in terms of kilograms).
This is my one nitpick with your otherwise excellent post. I especially liked your rant on temperature (bullet point #5) - have yet to see anyone come up with a valid reason why Celsius is inherently better than Fahrenheit other than being part of the SI standard.
If you're going to taut the advantages of your measurement system, shouldn't you maybe know a thing or two about it before you try? I can get much better accuracy in my native system by remembering that pure liquid H2O at around 75 degrees Farenheit and about 14.7 psia has a density of about 62.4 lb/(ft^3), and I'll avoid some of the potentially dangerous mistakes you're making.
Reminds me of another rule of thumb - at 70F and 100% RH, there's 1 pound (lbm not lbf) of water per 1,000 cubic feet of air.
One final rant - the US letter size is defined as 8.5 by 11 inches exactly - which works out to 215.9 by 279.4 mm (exactly) and good quality paper will be +/- 0.1 mm. The author of the website rounded the letter size to the nearest integral number of mm, which is just as sleazy as rounding the metric size to the nearest 1/32nd of an inch. While the root 2 aspect ratio of metric paper is nice for folding, there is no way that you can maintain an exact aspect ratio and have the size in an integral number of mm.
110V dates back to Edison's Pearl Street station in 1882.
60HZ vs 50Hz
Edison's Pearl Street station was DC - AC started to show up in the mid 1880's at quite a variety of frequencies (25Hz to 133Hz). There have been reports 60Hz was chosen as an American standard to be incompatible with the European 50Hz standard - not sure what diff that would make with the 110/220 differences. L.A. was on 50 Hz till the 1930's and the surrounding SCE territory was on 50Hz until 1948.
60Hz does have the advantage that transformers and motors can be made slightly smaller and lighter.
Water boils at 100C?? That's actually a very misleading statement. A more precise rendition is that the vapor pressure of water is equal to standard atmospheric pressure at 100C (actually just slightly less than 100C). In fact, water will boil at temperatures less than 0C, provided that the partial pressure of water is less than the vapor pressure (i.e. relative humidity is less than 100%).
For thermodynamic calculations, you would normally use absolute temperature, which is expressed in units of degrees Kelvin in SI or degrees Rankine in english engineering (R = 1.8 * K, exactly). Either way, you have to make use of lookup tables (e.g. steam tables) or some ugly conversion constant (e.g. Boltzmann's) to make use of temperature given in Celcius, Fahrenheit, Kelvin or Rankine.
Anyway, degrees Celsius is a big hairy wart on the metric system's butt.
I'm figuring that things will pick up for Sun about the end of the year.
I was using 3.7 back in early 1999 - so it took 5+ years to past 3.7.x. Nice to see an upgrade to a very useful program.
What I like best about gnuplot is that interactive mode and batch mode use the same commands - makes it really easy to write scripts and easy to write scripts that write scripts. One favorite was writing an awk script that would spit out a gnuplot script to plot 50+ files worth of raw and fitted data - and spit the output as a PostScript file.
Ummm - I think it is called "Sunscreen" or something to that effect. IIRC, it allows you the option of passing traffic to firewalled hosts without address translation.
If you just want to protect the host, then you can define secuirity associations for each port - including requiring IPSEC connections. Solaris has had native IPSEC support for several years (albeit with rather sucky documentation).
The more successfull virus are the one will long incubation period, take the virus that case AIDS for example.
If you read the book The Great Influenza by John Barry, you would find that people are usually contagious several days before symptoms appear. In some cases death occurred very rapidly after the onset of the symptoms (people dying in the trolley), but they had been infected several days before.
Barry did a very good job researching the book, very well worth reading. Not sure of what scared me more - the virus or the political climate in the US during Wilson's second term as president.
I do recall reading that M$ programmers preferred developing under OS/2 1.X because it was so picky about memory protection.
See recent cases involving GTA and Natural Born Killers...
Not to mention "Ivan" and "Joe" (?) from Clifford Simak's Cemetery World.
I've been to GD's plant in Westminster - and was impressed by the people that I met there.