The biggest problem I've had over the past few months with Redhat is this:
Heaps of apps targeted and marketed towards the Redhat "platform".
Redhat has dropped support for old products... this means no security updates, which upsets corporate IT departments and PHBs. They all say "thou shall upgrade to a supported version"
However, the RH 8 platform is only two years old and unsupported.
Now if your apps are "supported only under Redhat 8", but your servers "must upgrade to a supported version", short of tedious technical assessments, Redhat has effectively killed the products you're using.
PHB's say "I knew we shouldn't have trusted this Linux crap! Now our apps are unsupported!"
Now... does the application developer target RH AS? Which has already crept to 2.1, or do they consider SuSE, Debian, or ditch Linux altogether?
I'm certain IBM has encountered this problem. There are apps on the IBM website which used to show Linux support for Redhat 8 and under, but now that RH8 is unsupported, the apps show no support for Linux at all.
IBM, a company which still provides support for OS/2, and will probably service your ball typewriter if you paid them enough, has to tell its customers "our supplier has dropped support."
Will IBM stick with Redhat?
Something is going to blow very, very soon. Redhat would be wise to offer support for prior versions... and NOW, and drop the BS version creep.
Their stellar stock performance over the past few months has come at the price of customer loyalty.
I'm pulling down that patch mentioned in a parallel thread... bug 76831 looks like it. I don't think Bugzilla allows Slashdot links, so those interested will need to cut and paste http://bugzilla.mozilla.org/show_bug.cgi?id=76831
I should run a few Linux tests... I think this would appear, but under different circumstances.
I have this problem under NT4, Win2k and WinXP. Win98 didn't have the problem, and Linux doesn't appear to have it either.
If you open up the Task Manager and watch the virtual memory size and memory usage, Firefox (or any variant of Mozilla to date), will show a HUGE amount of VM space. Right now Firefox shows a VM size of ~70MB, current "memory usage" of ~43MB. If I start working with a different app, Windows will spot those inactive memory pages and swap them out. The "memory usage" will slowly fall on Firefox... down to maybe 2MB.
It is classic, predictable and boring application behaviour. Mozilla is a pig for RAM. This isn't a problem if it doesn't get paged out, but when it does, your machine will have to bring up tens of megabytes of swapped out and fragmented pages to re-render the page you've kept in the background.
So when I click on Firefox, it can take over 60 seconds (real stopwatch seconds) for it to pull back from my notebook's 4800rpm drive and reach the foreground...
Note that that is a lot longer than it takes to launch the app.
If there are memory leaks in Mozilla, they're too small for me to notice or worry about, but this swap behaviour is horrible... so much so, that I include it as a caveat when mentioning Mozilla to people, and I've even considered going to IE (I've can honestly say I've never been a regular IE user)... but when I'm on the phone with somebody at work, and I need to look something up, explaining to them that my machine is pulling the browser out of swap is unproductive and embarassing... sometimes I launch IE just to get to the site more quickly.
The hibernation doesn't really click with what I'm describing though. Hibernation brings back the full hibernation file to active RAM... if Mozilla wasn't swapped out before, it won't be swapped out afterwards.
Why NT/2k/XP? If you watch them, they all use similar memory management patterns. Even if you're not using all the available system RAM, they'll swap out inactive pages to increase space for caching drive activity.
For example: right now I have about 7MB available system RAM... a 70MB disk cache, and 564MB allocated on a system with 256MB of RAM. There is 70MB worth of RAM paged out in order to keep a big fat disk cache.
Anyways, watch those memory numbers as your HDD thrashes bringing back Mozilla... it is painfully obvious what is happening. Mozilla is a pig for RAM, and NT & kin are swapping it out.
I could "fix" this by putting ~768MB of RAM in my machine and turning off swap/paging space... but that doesn't bode well for the supieriority of Mozilla/Firefox/whatever.
Does anyone have a trick to tweak the memory management characteristics of Windows?
Saying that the OS doesn't need a media player to function is almost saying that the fact that an OS runs applications is just a side feature, like a stereo in a car. The real purpose for an OS is to... um... yeah.
By giving away their media player with their OS, they're using their OS to control the demand for their servers... which in turn secures the demand for their media player... which is only available with their OS.
Just like giving away gas-powered cars would control the demand for gasoline.
Is bundling a media play with windows somehow unfair to the consumer?
O.B. Automotive analogy:
Imagine if, fearing automakers' adopion of alternative fuels, the big oil companies collude, offering everyone free gas-powered cars.
The price of fuel everywhere is increased. Alternative fuel distribution depends on the pumps picking it up, and the pumps are locked out by the oil industry.
Consumers are still free to buy their own gas-powered car, but they'll be paying extra at the pumps anyways. If they went to alternative fuels, they'd have to carefully plot their trips for vehicle range etc.
. . .
Now imagine if, fearing the adoption of alternative operating systems, Microsoft increases the price of its operating system and gives everyone a free media player.
Music distribution depends on the servers picking it up, and only MS servers support the most widely supported client format. The price of music everywhere is increased to hide the cost.
Consumers are still free to use an alternative operating system. They won't be able to play any music from MS-only sites, but this is a free market right? If the demand was really there, they'd put up servers and support the format...
More precisely, OS is the tertiary product. Their primary product is a solid, supported, consistent API to attract and retain developers. The secondary product is a slick user interface for their desktop API.
In all practial aspects, for most people Linux is a Unix-like environment first, an OS second, and any semblence of a desktop API or slick desktop environment is not really all that important.
Microsoft could sell Win32 on Linux without too much pain... it would not be the first time they changed OSes for their environment.
Projecting 2004 to have 70% of all households with Internet connectivity (doubtful), there are about 70 million Internet connected households in the U.S... let's assume 100% of them read their email (I barely read my email with all the SPAM in it)
I don't know anyone who purchased anything via bulk email... or bulk mail for that matter (except taxis, and ordering fast food...), but it seems that the average person with Internet connectivity in the U.S. is buying about $430 worth of stuff... by email!
To add to this they indicate that the email must be non-fraudulant to count... I can't remember the last potentially non-fraudulent bulk unsolicited email I've seen.
I did the same stuff in highschool, very cool. It should be standard highschool stuff, it will serve you well. It's surprising how few schools have qualified teachers who can create the programs and teach the material.
My school didn't specialize in technology though... it was just big enough to keep a strange instructor and support the program. After the instructor retired, the course disappeared.
Wow... that's a sharp argument. I never thought of "come on" before.
The fact that he could write the software for himself to solve a problem for his parents and license it out as a solution for money is a fluke. It wouldn't have worked if he wasn't working for his parents. The code would be owned by his employer and he wouldn't see a cent... and if he wasn't working for an employer, he wouldn't have understood the problem to be able to develop the software.
...and this is another story from a fellow who is bismirching people for open sourcing software to build resumes which he counters that is insignificant compared to the experience that he earned while working for mommy and daddy
Ugh... Some people don't realize the advantages they have.
Better to write GPL'd code and have a small name than to beg -- with a blank resume -- to be able to code software for dollars.
You can get a lot done in Windows with VBA and calling on routines using OLE. It's evil evil stuff, but it works and doesn't take the brute-force sendkeys kind of route... that stuff is a hack.
I think that would keep Sun pretty firmly in the driver's seat.
It's odd...
Sun perhaps knows this best from OpenOffice/StarOffice. They've made a huge commitment to Open Source.
But... when dealing with companies like Microsoft, Microsoft could embrace the GPL'd Java. Sun would have have a hard time preventing Microsoft from repeating what they did with J++, they could market.net as an environment which is 100% forward compatible and simultaneously extends the capabilities of Java, a nice marketing perk to steal Sun's customers and lock them in to 0% backward compatability...
MS would have to backport and back-patch the new GPL'd releases of Sun's GPL'd Java to their extended environment, and maybe Sun could try to intentionally break MS GPL'd Java extensions, but I think that's easier said than done.
Slashdot Mirror
The biggest problem I've had over the past few months with Redhat is this:
Heaps of apps targeted and marketed towards the Redhat "platform".
Redhat has dropped support for old products... this means no security updates, which upsets corporate IT departments and PHBs. They all say "thou shall upgrade to a supported version"
However, the RH 8 platform is only two years old and unsupported.
Now if your apps are "supported only under Redhat 8", but your servers "must upgrade to a supported version", short of tedious technical assessments, Redhat has effectively killed the products you're using.
PHB's say "I knew we shouldn't have trusted this Linux crap! Now our apps are unsupported!"
Now... does the application developer target RH AS? Which has already crept to 2.1, or do they consider SuSE, Debian, or ditch Linux altogether?
I'm certain IBM has encountered this problem. There are apps on the IBM website which used to show Linux support for Redhat 8 and under, but now that RH8 is unsupported, the apps show no support for Linux at all.
IBM, a company which still provides support for OS/2, and will probably service your ball typewriter if you paid them enough, has to tell its customers "our supplier has dropped support."
Will IBM stick with Redhat?
Something is going to blow very, very soon. Redhat would be wise to offer support for prior versions... and NOW, and drop the BS version creep.
Their stellar stock performance over the past few months has come at the price of customer loyalty.
True, a bit more difficult to control is when the voter wants to be able to prove who they voted for.
Would a jaded population be driven to vote if they got paid $5?
Wouldn't you have to pay duty on the laptop? $2k in purchases over a four day stay probably exceeds any allowance the U.K. has for computers.
I'm pulling down that patch mentioned in a parallel thread... bug 76831 looks like it. I don't think Bugzilla allows Slashdot links, so those interested will need to cut and paste http://bugzilla.mozilla.org/show_bug.cgi?id=76831
I should run a few Linux tests... I think this would appear, but under different circumstances.
I have this problem under NT4, Win2k and WinXP. Win98 didn't have the problem, and Linux doesn't appear to have it either.
If you open up the Task Manager and watch the virtual memory size and memory usage, Firefox (or any variant of Mozilla to date), will show a HUGE amount of VM space. Right now Firefox shows a VM size of ~70MB, current "memory usage" of ~43MB. If I start working with a different app, Windows will spot those inactive memory pages and swap them out. The "memory usage" will slowly fall on Firefox... down to maybe 2MB.
It is classic, predictable and boring application behaviour. Mozilla is a pig for RAM. This isn't a problem if it doesn't get paged out, but when it does, your machine will have to bring up tens of megabytes of swapped out and fragmented pages to re-render the page you've kept in the background.
So when I click on Firefox, it can take over 60 seconds (real stopwatch seconds) for it to pull back from my notebook's 4800rpm drive and reach the foreground...
Note that that is a lot longer than it takes to launch the app.
If there are memory leaks in Mozilla, they're too small for me to notice or worry about, but this swap behaviour is horrible... so much so, that I include it as a caveat when mentioning Mozilla to people, and I've even considered going to IE (I've can honestly say I've never been a regular IE user)... but when I'm on the phone with somebody at work, and I need to look something up, explaining to them that my machine is pulling the browser out of swap is unproductive and embarassing... sometimes I launch IE just to get to the site more quickly.
The hibernation doesn't really click with what I'm describing though. Hibernation brings back the full hibernation file to active RAM... if Mozilla wasn't swapped out before, it won't be swapped out afterwards.
Why NT/2k/XP? If you watch them, they all use similar memory management patterns. Even if you're not using all the available system RAM, they'll swap out inactive pages to increase space for caching drive activity.
For example: right now I have about 7MB available system RAM... a 70MB disk cache, and 564MB allocated on a system with 256MB of RAM. There is 70MB worth of RAM paged out in order to keep a big fat disk cache.
Anyways, watch those memory numbers as your HDD thrashes bringing back Mozilla... it is painfully obvious what is happening. Mozilla is a pig for RAM, and NT & kin are swapping it out.
I could "fix" this by putting ~768MB of RAM in my machine and turning off swap/paging space... but that doesn't bode well for the supieriority of Mozilla/Firefox/whatever.
Does anyone have a trick to tweak the memory management characteristics of Windows?
Saying that the OS doesn't need a media player to function is almost saying that the fact that an OS runs applications is just a side feature, like a stereo in a car. The real purpose for an OS is to... um... yeah.
By giving away their media player with their OS, they're using their OS to control the demand for their servers... which in turn secures the demand for their media player... which is only available with their OS.
Just like giving away gas-powered cars would control the demand for gasoline.
Is bundling a media play with windows somehow unfair to the consumer?
O.B. Automotive analogy:
Imagine if, fearing automakers' adopion of alternative fuels, the big oil companies collude, offering everyone free gas-powered cars.
The price of fuel everywhere is increased. Alternative fuel distribution depends on the pumps picking it up, and the pumps are locked out by the oil industry.
Consumers are still free to buy their own gas-powered car, but they'll be paying extra at the pumps anyways. If they went to alternative fuels, they'd have to carefully plot their trips for vehicle range etc.
. . .
Now imagine if, fearing the adoption of alternative operating systems, Microsoft increases the price of its operating system and gives everyone a free media player.
Music distribution depends on the servers picking it up, and only MS servers support the most widely supported client format. The price of music everywhere is increased to hide the cost.
Consumers are still free to use an alternative operating system. They won't be able to play any music from MS-only sites, but this is a free market right? If the demand was really there, they'd put up servers and support the format...
Right?
If those servers were MX51's, you wouldn't need the paging system, you could blast it at 180db!
These things have been featured on Tom's Hardware!
http://www.tomshardware.com/column/20040317/index. html
Intramuscular shot in a tense muscle? Interesting observation though.
Yeah, and there are no nearsighted squirrels either.
GNU formally started in '83. DOS came out in '81.
And I never implied they were dawdling, free software is focused on something very different than closed proprietary software.
Worse, Windows has become pretty solid in the time it has taken Linux to put together a tolerable desktop environment.
Or read it ten years ago.
More precisely, OS is the tertiary product. Their primary product is a solid, supported, consistent API to attract and retain developers. The secondary product is a slick user interface for their desktop API.
In all practial aspects, for most people Linux is a Unix-like environment first, an OS second, and any semblence of a desktop API or slick desktop environment is not really all that important.
Microsoft could sell Win32 on Linux without too much pain... it would not be the first time they changed OSes for their environment.
Their superior minds are no match for Robocop III's puny weapons.
There's a clause in NAFTA saying that if Canada ever starts selling bulk water, it cannot stop.
So there's a good incentive to never sell bulk water to the U.S.
You should hold on until the 10th anniversary post-Hobbit LOTR ultra-high-density HDTV version. That will probably be out in 2016 or so.
http://www.infoplease.com/ipa/A0902841.html
http://www.infoplease.com/ipa/A0005055.html
Projecting 2004 to have 70% of all households with Internet connectivity (doubtful), there are about 70 million Internet connected households in the U.S... let's assume 100% of them read their email (I barely read my email with all the SPAM in it)
I don't know anyone who purchased anything via bulk email... or bulk mail for that matter (except taxis, and ordering fast food...), but it seems that the average person with Internet connectivity in the U.S. is buying about $430 worth of stuff... by email!
To add to this they indicate that the email must be non-fraudulant to count... I can't remember the last potentially non-fraudulent bulk unsolicited email I've seen.
From a survey of 1000 respondants... $32.5 billion on solicited and unsolicited combined.
What's the U.S. population these days?
250,000,000?
$130 for every man, woman and child in the U.S.?
How much per household with a computer and an internet connection?
By email?
Based on a survey?
Of people who responded?
Of people who knew what email was?
Of people who knew what it meant to respond to an email?
Of people who knew the difference between a solicited and an unsolicited email?
Sponsored by the Direct Marketing Association?
I call BS.
I did the same stuff in highschool, very cool. It should be standard highschool stuff, it will serve you well. It's surprising how few schools have qualified teachers who can create the programs and teach the material.
My school didn't specialize in technology though... it was just big enough to keep a strange instructor and support the program. After the instructor retired, the course disappeared.
Wow... that's a sharp argument. I never thought of "come on" before.
The fact that he could write the software for himself to solve a problem for his parents and license it out as a solution for money is a fluke. It wouldn't have worked if he wasn't working for his parents. The code would be owned by his employer and he wouldn't see a cent... and if he wasn't working for an employer, he wouldn't have understood the problem to be able to develop the software.
...and this is another story from a fellow who is bismirching people for open sourcing software to build resumes which he counters that is insignificant compared to the experience that he earned while working for mommy and daddy
Ugh... Some people don't realize the advantages they have.
Better to write GPL'd code and have a small name than to beg -- with a blank resume -- to be able to code software for dollars.
You can get a lot done in Windows with VBA and calling on routines using OLE. It's evil evil stuff, but it works and doesn't take the brute-force sendkeys kind of route... that stuff is a hack.
Check out what the sick Perl bastards have been up to : http://www.xav.com/perl/faq/Windows/ActivePerl-Win faq12.html#use_ole
I've always had trouble finding good docs on this stuff when I try to take advantage of it.
Please stop about the McDonald's coffee lady.
http://lawandhelp.com/q298-2.htm
Answer:
Don't speak to anyone at Comdex, and for the love of humanity do not swipe your card at any vendor's booths much less buy anything.
For that matter, don't go to Comdex!
I think that would keep Sun pretty firmly in the driver's seat.
It's odd...
Sun perhaps knows this best from OpenOffice/StarOffice. They've made a huge commitment to Open Source.
But... when dealing with companies like Microsoft, Microsoft could embrace the GPL'd Java. Sun would have have a hard time preventing Microsoft from repeating what they did with J++, they could market .net as an environment which is 100% forward compatible and simultaneously extends the capabilities of Java, a nice marketing perk to steal Sun's customers and lock them in to 0% backward compatability...
MS would have to backport and back-patch the new GPL'd releases of Sun's GPL'd Java to their extended environment, and maybe Sun could try to intentionally break MS GPL'd Java extensions, but I think that's easier said than done.