I should have been more specific - allow me to rephrase.
Microsoft has been playing games with versions and their OS since DOS 5 and the setver days.
Admittedly, their motivations were different then - at least for setver.
I would perhaps have been more accurate to push actual evil intent out to DOS 6. Remember the DR DOS shenanigans? Wikipedia does...
"Though DR-DOS was almost 100% binary compatible with applications written for MS-DOS, Microsoft nevertheless expended considerable effort in attempts to break compatibility. In one example, they inserted code into the beta version of Windows 3.1 to return a non-fatal error message if it detected a non-Microsoft DOS. With the detection code disabled (or if the user canceled the error message), Windows ran perfectly under DR-DOS." - Wikipedia article on DR-DOS
I was doing tech support for Quarterdeck in those days. MS played games in some versions with QEMM as well. It was fairly annoying.
Point being - MS has a history of changing code to break compatability, for no real reason than to push their own products. It is unsurprising that they've continued this with Vista.
This isn't a surprise - the exact same thing is happening right now, and has been, with XP.
I never "upgraded" past 2k - XP didn't have anything I needed or wanted.
When Age of Empires III came out, I bought it, as I did with all of the other Age of Empires games and expansions. No luck - the game "requires XP" to function. Not that it really does - there are undocumented switches to let it install on 2k, and it works fine (the demo was the same way). Ditto Rise of Legends, *another* game that MS bought that now "requires XP", even though it's not doing anything that's beyond what 2k offers.
MS has always had the strategy of selling OS upgrades by artifically requiring them - hell, they've done this since DOS version 5 and "setver". That they would make something Vista-only purely to drive Vista sales is par for the course.
Of course, in my case, it *didn't* drive sales of a new OS - at least not Microsoft's. I still run 2k to this day, alongside what I did upgrade to, my Powermac dual G5. Someday 2k will be useless to me (probably when I upgrade my mac to an intel based one and virtualize windows), and I'll reformat and run the old box as a freebsd one or such. I guess MS actually did me a favor in some sense...
It may be expensive to filter packets at the router level - but it's not the cisco that's doing NTP service (I assume), it's a unix box of some sort.
So - make a whitelist (it's only a few thousand legit servers), hash it (so a yes/no lookup is cheap), and give a bogus response on a miss. On a modern architecture compiled language, the extra processing should be sub-millisecond - hopefully fast enough to avoid messing NTP up for legit users. If you're lucky, a really bad time will cause real problems for Dlink customers, who can then complain to the vendor. If you're *really* lucky, they'll patch to avoid the support burden.
Yes, this would take a patched NTP. Yes, this doesn't deal with traffic, or your expenses. Yes, it only indirectly causes problems for DLink. And Yes, you shouldn't have to do this at all.
Dlink isn't going to do anything unless forced, that's clear. Your open letter may help - I hope it does - but if it doesn't, you do have a way to deny them service without grossly inconveniencing your legitimate users. You still bear the bandwidth costs... but it's something.
Having said all that - simply renumbering your IP shouldn't be as heinous as all that. Any semi-competent server admin should have a fallback NTP in case yours is down - so a renumber, while disruptive, isn't impossible. If you do it - do it sooner rather than later.
Dlink - you can fix this. Apologize, and contract with this man to provide NTP services, covering his expenses and time. He's not the only one you're hammering, and if they all take steps you're screwed. Nip this in the bud - you may well have not realized that NTP is something you have to pay for one way or the other (your alternative would have been running your own server, really) - but you do. Don't let this bad PR situation get worse.
The entire discussion is silly. Too many scientists, not enough engineers.
If cow tipping is possible, it should be easily demonstratable. Anyone have a video?
No video, after all this time and discussion, and you can be pretty sure it's exceedingly difficult, at least. Googling shows that to be the likely case.
Don't agree? Show us the video. One video showing it proves it can be done, no matter what the scientists say.
Except, of course, you're leaving out things (which, to be fair, weren't in the blurb you cut and pasted from the apple store) - the built-in bluetooth, the built-in webcam, the freaking *remote control*... and of course, the OS and bundled software, which is the whole point of this thread.
What's XP run now? Legitimately? Unless you're just stealing the OS...
Once you tack everything else on, they're very comperable in price. Now - look at what you've wrought.
You have a cobbled together PC, with all of the driver woes and incompatabilities that comes with - again, not as bad now as it used to be, but EWWW anyway. Big and Clunky. (You're *NOT* getting a nice case for $50)
Or you have a mac. Go look at it! Seriously - go look at it at apple's online store. Beautiful. Stable. Fast. and unix under the hood.
I left my prior employment in similar circumstances.
Two weeks is customary - 4 weeks was quite generous.
The fact that he responded with 6 weeks plus other "requirements" shows you he's not looking to make a transition easier - he's looking to milk you for as much as he can get. Oh, I've *sooo* been there.
Tell him he has two weeks, and if he has an issue with that you can leave now. Consulting past that is negotiable (check with your new employer first to avoid a conflict of interest) - and be sure you charge *more* than you were making, not less, or you'll have just traded your job for a lower-paid consulting gig.
And be ready to just walk. Seriously. I know you don't want to leave anyone in a lurch, but this isn't a situation of your own making - it was your bosses job to find and hire people within a reasonable time to replace those who had left, and just because he let it go this far is no reason for *you* to pay the consequences...
> Voteing for the lesser of two evils still gives you evil.
Yes, I keep hearing that.
But - *less* evil. *less*. That means not as much! I'm not sure why this is so hard for everyone to understand. Less evil is generally better than more evil, unless evil is your bag, baby.
Nader is pulling for *no* evil - and while that's laughably innocent, it's not gonna happen this time around - and the efforts squandered in going after no evil are efforts removed from getting little evil into the white house.
Making a complete formal system is easy - but making a non-trivial one that's still complete is significantly harder. Turns out simple mathematics is sufficiently complex for the system to be provably non-complete. So you can effectively substitute "interesting" for "sufficiently complex" in Godel's theorem... (I'm *not* going to figure out how to umlaut that "o" in Godel)
It's working fine for me under Mozilla 1.7b (windows version, ack)
Re:PHP OpenGL on SourceForge
on
OpenGL in PHP
·
· Score: 5, Funny
Please, "not slashdot worthy"? I believe the page that details how to make an Enterprise out of a dead floppy disk made it up here, twice. You got that *BEAT*.
It's not FUD, but it's not entirely accurate either.
In a nutshell, the MS JVM by default has a smaller Heap (ie. it takes less memory to use for itself). This makes garbage collection faster, as there's less memory to scan. When you look at something like an animation applet, they tend to run "choppier" on default sun JVM installs, because the sun JVM is scanning 128megs (on my system) of memory, while the MS JVM is scanning like 8 megs. When you run a *big* app on MS JVM, or tune the Sun JVM, that particular issue disappears. There are likely other factors as well, but that's the one that's made a difference for me.
The newer Sun JVMs do Garbage Collection differently now, so this isn't nearly as much of an issue as it used to be.
A thread about this is at the Sun Developer Network - http://forum.java.sun.com/thread.jsp?forum=54&thre ad=439174
So, in summary - the MS JVM is tuned for toy applets, while Sun's is tuned for larger applications. You can presumably tweak both to get the performance you are looking for.
Given that this bug appears to be *unexploitable* in OpenBSD, no, I suspect they will *not* change that claim. The hole is only an issue if you you're running OpenSSH on a slightly less secure OS, apparently. Huh, whooda thunk it?
I don't know about davezilla, but mozilla should be ok. IANAL, But I thought trademarks had to be agressively defended to "stay fresh" - mozilla.org has been around since 1998 (at least), and the netscape browser's been called mozilla for longer than that, in public. I'd say letting it slide for 4+ years is an indication that they have *not* been defending this trademark, at least in this arena...
need a lot more work before prime-time. This is nothing to do with OS X per se, and not all that much to do with Perl - it's the bindings between the two that need work. I've spent some time trying to track down the status of all of the various ways to do Graphics (GUI and just plain old drawing) in Perl under OS X, and none of them are ready for prime time. (My Mac.com homepage has details I won't go into here - http://homepage.mac.com/bortels if you want to get into sad detail.
Having said that, if you're *very* familiar with Cocoa already, and are looking more towards GUI and not so much towards video-games, Camelbones is the leading suspect right now. I personally found it confusing, and you need to do development in the Apple Project Builder for the most part, but it has one key thing going for it - it works, and there's sample source.
(I'm personally on the video-game quest - I want to be able to say "use GLUT", and go nuts, without an IDE getting in the way. But I digress...)
The Sourceforge page for Camelbones is here: http://sourceforge.net/projects/camelbones/
A homepage discussing Camelbones (but seemingly not up-to-date - it talks about 0.1, but 0.2 is available) is here: http://camelbones.sourceforge.net/
They havn't posted a date for the playable demo, of course, but Blizzard has indicated that after the closed 1000 person beta there will be a larger, open beta, prior to release. I guess the concept is to (1) test the servers under load, and (2) get a few hundred thousand hooked, then yank the rug out from under then and make them buy it. Hey, it works for me...
Worth pointing out is that Phil isn't bashing Perl alone - he bashes almost every tool he uses at one point or another. I went ahead and did a search on his site for "Perl", and in almost every instance where he was griping about it, he was griping about other things as well. He's not griping about Perl alone, he's griping about *everything* he's used at one point or another.
*This is a good thing*. He's not so much narrow-minded as opinionated, and opinions from people who actually have done the work are worth their weight in gold. Nothing's perfect, and it's refreshing to hear honest, educated opinions about this stuff, even if I don't necessarily agree.
Case in point - I love Perl. I can safely say that Perl has saved me *thousands* of hours that would have otherwise been spent trying to deal with things in other languages. Is Perl perfect? Not by a longshot. Is it the ideal language? There *is* no ideal language for everything - each has it's strengths and weaknesses.
Phil likes, and recommends, the stuff he uses and is familiar with. I suppose if I had a strong grounding in Lisp, I might find the syntax of Perl abhorrent too. (Personally, my brain can't handle the multiple levels of nesting of parenthesis in Lisp) To each his own - I'm always interested in the advice, even if I choose to do things my way for my own reasons. Specifically, the fact that I know Perl *well* makes it a more suitable language than TCL or Lisp for what I do - and it means that I disagree with Phil here, but *neither one of us are wrong*. It's opinion.
Slashdot Mirror
I should have been more specific - allow me to rephrase.
Microsoft has been playing games with versions and their OS since DOS 5 and the setver days.
Admittedly, their motivations were different then - at least for setver.
I would perhaps have been more accurate to push actual evil intent out to DOS 6. Remember the DR DOS shenanigans? Wikipedia does...
"Though DR-DOS was almost 100% binary compatible with applications written for MS-DOS, Microsoft nevertheless expended considerable effort in attempts to break compatibility. In one example, they inserted code into the beta version of Windows 3.1 to return a non-fatal error message if it detected a non-Microsoft DOS. With the detection code disabled (or if the user canceled the error message), Windows ran perfectly under DR-DOS." - Wikipedia article on DR-DOS
I was doing tech support for Quarterdeck in those days. MS played games in some versions with QEMM as well. It was fairly annoying.
Point being - MS has a history of changing code to break compatability, for no real reason than to push their own products. It is unsurprising that they've continued this with Vista.
This isn't a surprise - the exact same thing is happening right now, and has been, with XP.
I never "upgraded" past 2k - XP didn't have anything I needed or wanted.
When Age of Empires III came out, I bought it, as I did with all of the other Age of Empires games and expansions. No luck - the game "requires XP" to function. Not that it really does - there are undocumented switches to let it install on 2k, and it works fine (the demo was the same way). Ditto Rise of Legends, *another* game that MS bought that now "requires XP", even though it's not doing anything that's beyond what 2k offers.
MS has always had the strategy of selling OS upgrades by artifically requiring them - hell, they've done this since DOS version 5 and "setver". That they would make something Vista-only purely to drive Vista sales is par for the course.
Of course, in my case, it *didn't* drive sales of a new OS - at least not Microsoft's. I still run 2k to this day, alongside what I did upgrade to, my Powermac dual G5. Someday 2k will be useless to me (probably when I upgrade my mac to an intel based one and virtualize windows), and I'll reformat and run the old box as a freebsd one or such. I guess MS actually did me a favor in some sense...
It may be expensive to filter packets at the router level - but it's not the cisco that's doing NTP service (I assume), it's a unix box of some sort.
So - make a whitelist (it's only a few thousand legit servers), hash it (so a yes/no lookup is cheap), and give a bogus response on a miss. On a modern architecture compiled language, the extra processing should be sub-millisecond - hopefully fast enough to avoid messing NTP up for legit users. If you're lucky, a really bad time will cause real problems for Dlink customers, who can then complain to the vendor. If you're *really* lucky, they'll patch to avoid the support burden.
Yes, this would take a patched NTP. Yes, this doesn't deal with traffic, or your expenses. Yes, it only indirectly causes problems for DLink. And Yes, you shouldn't have to do this at all.
Dlink isn't going to do anything unless forced, that's clear. Your open letter may help - I hope it does - but if it doesn't, you do have a way to deny them service without grossly inconveniencing your legitimate users. You still bear the bandwidth costs... but it's something.
Having said all that - simply renumbering your IP shouldn't be as heinous as all that. Any semi-competent server admin should have a fallback NTP in case yours is down - so a renumber, while disruptive, isn't impossible. If you do it - do it sooner rather than later.
Dlink - you can fix this. Apologize, and contract with this man to provide NTP services, covering his expenses and time. He's not the only one you're hammering, and if they all take steps you're screwed. Nip this in the bud - you may well have not realized that NTP is something you have to pay for one way or the other (your alternative would have been running your own server, really) - but you do. Don't let this bad PR situation get worse.
The entire discussion is silly. Too many scientists, not enough engineers.
If cow tipping is possible, it should be easily demonstratable. Anyone have a video?
No video, after all this time and discussion, and you can be pretty sure it's exceedingly difficult, at least. Googling shows that to be the likely case.
Don't agree? Show us the video. One video showing it proves it can be done, no matter what the scientists say.
Except, of course, you're leaving out things (which, to be fair, weren't in the blurb you cut and pasted from the apple store) - the built-in bluetooth, the built-in webcam, the freaking *remote control*... and of course, the OS and bundled software, which is the whole point of this thread.
What's XP run now? Legitimately? Unless you're just stealing the OS...
Once you tack everything else on, they're very comperable in price. Now - look at what you've wrought.
You have a cobbled together PC, with all of the driver woes and incompatabilities that comes with - again, not as bad now as it used to be, but EWWW anyway. Big and Clunky. (You're *NOT* getting a nice case for $50)
Or you have a mac. Go look at it! Seriously - go look at it at apple's online store. Beautiful. Stable. Fast. and unix under the hood.
I left my prior employment in similar circumstances.
Two weeks is customary - 4 weeks was quite generous.
The fact that he responded with 6 weeks plus other "requirements" shows you he's not looking to make a transition easier - he's looking to milk you for as much as he can get. Oh, I've *sooo* been there.
Tell him he has two weeks, and if he has an issue with that you can leave now. Consulting past that is negotiable (check with your new employer first to avoid a conflict of interest) - and be sure you charge *more* than you were making, not less, or you'll have just traded your job for a lower-paid consulting gig.
And be ready to just walk. Seriously. I know you don't want to leave anyone in a lurch, but this isn't a situation of your own making - it was your bosses job to find and hire people within a reasonable time to replace those who had left, and just because he let it go this far is no reason for *you* to pay the consequences...
> Voteing for the lesser of two evils still gives you evil.
Yes, I keep hearing that.
But - *less* evil. *less*. That means not as much! I'm not sure why this is so hard for everyone to understand. Less evil is generally better than more evil, unless evil is your bag, baby.
Nader is pulling for *no* evil - and while that's laughably innocent, it's not gonna happen this time around - and the efforts squandered in going after no evil are efforts removed from getting little evil into the white house.
Making a complete formal system is easy - but making a non-trivial one that's still complete is significantly harder. Turns out simple mathematics is sufficiently complex for the system to be provably non-complete. So you can effectively substitute "interesting" for "sufficiently complex" in Godel's theorem... (I'm *not* going to figure out how to umlaut that "o" in Godel)
It's working fine for me under Mozilla 1.7b (windows version, ack)
Please, "not slashdot worthy"? I believe the page that details how to make an Enterprise out of a dead floppy disk made it up here, twice. You got that *BEAT*.
It's not FUD, but it's not entirely accurate either.
e ad=439174
In a nutshell, the MS JVM by default has a smaller Heap (ie. it takes less memory to use for itself). This makes garbage collection faster, as there's less memory to scan. When you look at something like an animation applet, they tend to run "choppier" on default sun JVM installs, because the sun JVM is scanning 128megs (on my system) of memory, while the MS JVM is scanning like 8 megs. When you run a *big* app on MS JVM, or tune the Sun JVM, that particular issue disappears. There are likely other factors as well, but that's the one that's made a difference for me.
The newer Sun JVMs do Garbage Collection differently now, so this isn't nearly as much of an issue as it used to be.
A thread about this is at the Sun Developer Network - http://forum.java.sun.com/thread.jsp?forum=54&thr
So, in summary - the MS JVM is tuned for toy applets, while Sun's is tuned for larger applications. You can presumably tweak both to get the performance you are looking for.
-- Tom Bortels
Given that this bug appears to be *unexploitable* in OpenBSD, no, I suspect they will *not* change that claim. The hole is only an issue if you you're running OpenSSH on a slightly less secure OS, apparently. Huh, whooda thunk it?
I don't know about davezilla, but mozilla should be ok. IANAL, But I thought trademarks had to be agressively defended to "stay fresh" - mozilla.org has been around since 1998 (at least), and the netscape browser's been called mozilla for longer than that, in public. I'd say letting it slide for 4+ years is an indication that they have *not* been defending this trademark, at least in this arena...
-- the Ogre
Pathetic? yeah, maybe.
But it can be pretty:
http://www.frozen-bubble.org/
That's Perl and perlsdl.
need a lot more work before prime-time. This is nothing to do with OS X per se, and not all that much to do with Perl - it's the bindings between the two that need work. I've spent some time trying to track down the status of all of the various ways to do Graphics (GUI and just plain old drawing) in Perl under OS X, and none of them are ready for prime time. (My Mac.com homepage has details I won't go into here - http://homepage.mac.com/bortels if you want to get into sad detail.
Having said that, if you're *very* familiar with Cocoa already, and are looking more towards GUI and not so much towards video-games, Camelbones is the leading suspect right now. I personally found it confusing, and you need to do development in the Apple Project Builder for the most part, but it has one key thing going for it - it works, and there's sample source.
(I'm personally on the video-game quest - I want to be able to say "use GLUT", and go nuts, without an IDE getting in the way. But I digress...)
The Sourceforge page for Camelbones is here: http://sourceforge.net/projects/camelbones/
A homepage discussing Camelbones (but seemingly not up-to-date - it talks about 0.1, but 0.2 is available) is here: http://camelbones.sourceforge.net/
-- Tom Bortels
They havn't posted a date for the playable demo, of course, but Blizzard has indicated that after the closed 1000 person beta there will be a larger, open beta, prior to release. I guess the concept is to (1) test the servers under load, and (2) get a few hundred thousand hooked, then yank the rug out from under then and make them buy it. Hey, it works for me...
Worth pointing out is that Phil isn't bashing Perl alone - he bashes almost every tool he uses at one point or another. I went ahead and did a search on his site for "Perl", and in almost every instance where he was griping about it, he was griping about other things as well. He's not griping about Perl alone, he's griping about *everything* he's used at one point or another.
*This is a good thing*. He's not so much narrow-minded as opinionated, and opinions from people who actually have done the work are worth their weight in gold. Nothing's perfect, and it's refreshing to hear honest, educated opinions about this stuff, even if I don't necessarily agree.
Case in point - I love Perl. I can safely say that Perl has saved me *thousands* of hours that would have otherwise been spent trying to deal with things in other languages. Is Perl perfect? Not by a longshot. Is it the ideal language? There *is* no ideal language for everything - each has it's strengths and weaknesses.
Phil likes, and recommends, the stuff he uses and is familiar with. I suppose if I had a strong grounding in Lisp, I might find the syntax of Perl abhorrent too. (Personally, my brain can't handle the multiple levels of nesting of parenthesis in Lisp) To each his own - I'm always interested in the advice, even if I choose to do things my way for my own reasons. Specifically, the fact that I know Perl *well* makes it a more suitable language than TCL or Lisp for what I do - and it means that I disagree with Phil here, but *neither one of us are wrong*. It's opinion.