If you're saying the problem is too complex and should not be attempted, then I might agree.
I'm not suggesting anything quite so definite.
I'm simply saying that if you need the source code to evaluate a model then you're either not qualified to evaluate the model or the model is so complex that source code access is not going to tell you if it's valid or not.
The attached model has been under continuous development for about as long as Windows.
Funny you should say that. In my opinion Windows should not be used for quite a wide variety of applications simply because it's to complex to determine whether it meets the requirements.
I have repeatedly seen glorious pictures of beautiful ocean models, only to discover, after a few months of working with the scientist in question, that they have artfully underemphasized the failings in their models, while, at the same time, being perfectly accurate in their description of the algorithms. It's only when you use their tools that you understand their errors.
If you can't reproduce the model based on the algorithms, then they didn't disclose enough details of the algorithms. If the model is too large and complex to reproduce in that way, then I would question whether it's science at all... regardless of whether they released the source code or not.
If you really need to ship the source code around, then you really have no better understanding... and probably less... of the models and algorithms you're working on than Metzger understood the biology and biochemistry of his fruit flies. The software isn't a model, any more, it's an experimental subject.
All had multiple other responsibilities and only three understood most of the code (I am not one of them and none of us realized we weren't diffusing)
If the people working on the code didn't understand it well enough, then it wasn't science, it was... I don't know, philosophy, or art.
I doubt Richard Feynman would have been impressed.
Apparently some DEC guys presented him with an Alphaserver running a DECsystem emulator that had Bill Gates' old TOPS account from when he was developing MASM on the DECsystem... Under the emulator, it's running Digital UNIX/Tru64.
Why, exactly, do I need to automate something on my Windows box?
Because you want to do it more than once.
I can much more easily tell my desktop PC to send the data to a Linux box, which I can script like there's no tomorrow. The results can then be sent back to me, and I can look at them right here on my Windows desktop. So what's the problem?
OK, tell me how to automatically send my timecard data from SAP to Linux (or OS X, or Interix running on the same computer... you don't need a Linux box in the loop to see the problem), and send a summary through Lotus Notes to my boss, and do it automatically every Friday whether I'm logged in or not.
Re:And there is much of my quarrel with BillG
on
The PC Is Not Dead
·
· Score: 1
If you want my recommendation for your software product, ask yourself, "would there be any point in having this run automatically when nobody is around?" And if the answer is "yes", *make it easy to do so*.
Eg, Airport Express should come up without me having to log in on the console of my Macintosh.
Science functions only on open review. If you can't duplicate someone's results, they are useless (c.f. Ponds and Fleischman [sp?]). A scientific result is only of value if it describes a consistent replicatable process. This is why I consider the closed source work to be completely meaningless. It may be perfect, it may be bug-ridden garbage, we'll never know!
If they claim some algorithm can be used to produce their results, then you can test their claims by implementing their algorithm. That's all the repeatability you need. Simply claiming that you need the code to do that is either politics or intellectual laziness. In this case I suspect politics.
Can't happen. If you turned ActiveX off all the "enhanced" Windows Explorer features would stop working, half the control panel applets would stop working, Outlook would stop working, and Windows Update wouldn't work.
ActiveX isn't off: it's disabled for the "Internet Zone", but it's still vulnerable to cross-zone attacks.
Unless there's something truly amazing happening in some version more recent than this past February they're OK. Not great, just OK. If you want "great" try an IBM Thinkpad, and if you want "terriffic" try an old Toshiba from back when they were putting full-size mechanisms in their laptops (486, early Pentium).
To a certain part of the criminal justice system, there are no innocent people... merely people for whom it has not yet been established what they are guilty of.
But they don't, always, in fact they've varied significantly from 100dpi in the past. And with Quartz Extreme there's really no reason to stick to any specific pixel density... their scaling technology is better than just about anyone's.
I suspect that they stick to 100dpi for the same reason iPods have such powerful output (even my shuffle can drive unpowered external speakers)... Steve Jobs likes it that way, they're well adapted to *his* hearing and visual acuity. For someone with significantly better than average acuity Apple's screens are painfully jaggy.
They're targeting consumers who have little motivation to understand much more than "point-and-click". That being the case, I don't think the article is necessarily an instance of FUD-spreading.
Sure it is, because those "point and click" users are not being helped by installing antivirus software when there's no viruses to catch. Until there are, all Norton or any other AV will do will be to slightly increase the risk of data loss by making their system a little less stable.
The only exploit they point to is a rootkit... which is something you install *after* you've exploited the box... there are no active threats that any antivirus software will work aaginst.
This is like their attempt to talk up a manually-installed program that deleted all your files on the Palm as an exploit, to push their useless PalmOS antivirus. And then their Pocket PC antivirus actually caused people data loss from false alarms.
Until there's an active threat in the wild, AND it's been analysed and an identifying signature discovered, antivirus software's only result is to make your computer less stable and less reliable because of its deep hooks in the OS.
This is not to say that the OS is magically perfectly secure, but anything any AV company tells you about ANY platform but Windows, at the moment, should be taken with a sackful of salt.
You do realize that you can setup an account to automatically login don't you?
I can't believe anyone would actually suggest this as a reasonable workaround. It's at best a necessary evil to be bypassed as soon as possible.
HOWEVER...
remember, everything the SystemStarter and login runs can be run from a script, so you should be able to bring up enough to do wireless without logging in with some poking about...
Mediocre (but improving, they used to be horrid) keyboards, one-button touchpad, and low resolution screens.
Anything above the base 12" iBook should have at least 1280x1024, and the 17" model should be more like 1920x1200. My 14" Thinkpad has more pixels than the 17" Powerbook.
Yeah, um, which is why you only send them half of the key.
If you don't send them the whole key, then there's no point in sending the DRM-protected self-destructing AI. And, for that matter, there's no point in sending the encrypted data at all, since the latency of the link is so high (years) that latency far dominates bandwidth that you're not saving anything.
If the bandwidth is so low that latency doesn't dominate bandwidth, then you can't afford the overhead of the self-destructing AI in the first place.
Mostly it's for the digitized signatures, but the paper signatures are scanned and digitized as well.
My logic is that if someone ganks a file of signatures off a poorly secured server somewhere, they'll see mine's tagged and figure that I'm an anal-retentive asshole and go rip off someone else instead.
Ender's War (the short story) was a much better story than the novel Card expanded it into. Plus, it'd be easier to fit into 2 hours on the screen. Pity they didn't pick it instead.
The ambuiguity at the end over just who the Enemy was is wonderful - see, there's no aliens in there, and the one reference in the short story to the planet Ender's living on implies that it's noth Earth, so it COULD be a rebellious colony... which would make the Enemy planet Earth.
Slashdot Mirror
If you're saying the problem is too complex and should not be attempted, then I might agree.
I'm not suggesting anything quite so definite.
I'm simply saying that if you need the source code to evaluate a model then you're either not qualified to evaluate the model or the model is so complex that source code access is not going to tell you if it's valid or not.
The attached model has been under continuous development for about as long as Windows.
Funny you should say that. In my opinion Windows should not be used for quite a wide variety of applications simply because it's to complex to determine whether it meets the requirements.
Arse.
I shoulda previewed...
you're working on than Metzger understood the biology and biochemistry of his fruit flies
I meant to say...
you're working on than Metzger understood the biology and biochemistry of his fruit flies when he started
I have repeatedly seen glorious pictures of beautiful ocean models, only to discover, after a few months of working with the scientist in question, that they have artfully underemphasized the failings in their models, while, at the same time, being perfectly accurate in their description of the algorithms. It's only when you use their tools that you understand their errors.
If you can't reproduce the model based on the algorithms, then they didn't disclose enough details of the algorithms. If the model is too large and complex to reproduce in that way, then I would question whether it's science at all... regardless of whether they released the source code or not.
If you really need to ship the source code around, then you really have no better understanding... and probably less... of the models and algorithms you're working on than Metzger understood the biology and biochemistry of his fruit flies. The software isn't a model, any more, it's an experimental subject.
All had multiple other responsibilities and only three understood most of the code (I am not one of them and none of us realized we weren't diffusing)
If the people working on the code didn't understand it well enough, then it wasn't science, it was... I don't know, philosophy, or art.
I doubt Richard Feynman would have been impressed.
Wake me when Bill Gates runs Linux on his Mac.
Bill Gates runs Tru64 on his Alphaserver.
Apparently some DEC guys presented him with an Alphaserver running a DECsystem emulator that had Bill Gates' old TOPS account from when he was developing MASM on the DECsystem... Under the emulator, it's running Digital UNIX/Tru64.
If he says positive things about it, where's your pile of cash that qualifies you to argue about it?
You mean this money I spent on college was wasted? I should have just sat on it 'cos it's the money that matters... not what you do with it?
Come now, most of the PCs out there have a thriving ecosystem of viruses and spyware.
Why, exactly, do I need to automate something on my Windows box?
Because you want to do it more than once.
I can much more easily tell my desktop PC to send the data to a Linux box, which I can script like there's no tomorrow. The results can then be sent back to me, and I can look at them right here on my Windows desktop. So what's the problem?
OK, tell me how to automatically send my timecard data from SAP to Linux (or OS X, or Interix running on the same computer... you don't need a Linux box in the loop to see the problem), and send a summary through Lotus Notes to my boss, and do it automatically every Friday whether I'm logged in or not.
If you want my recommendation for your software product, ask yourself, "would there be any point in having this run automatically when nobody is around?" And if the answer is "yes", *make it easy to do so*.
Eg, Airport Express should come up without me having to log in on the console of my Macintosh.
Science functions only on open review. If you can't duplicate someone's results, they are useless (c.f. Ponds and Fleischman [sp?]). A scientific result is only of value if it describes a consistent replicatable process. This is why I consider the closed source work to be completely meaningless. It may be perfect, it may be bug-ridden garbage, we'll never know!
If they claim some algorithm can be used to produce their results, then you can test their claims by implementing their algorithm. That's all the repeatability you need. Simply claiming that you need the code to do that is either politics or intellectual laziness. In this case I suspect politics.
ActiveX was off by default
Can't happen. If you turned ActiveX off all the "enhanced" Windows Explorer features would stop working, half the control panel applets would stop working, Outlook would stop working, and Windows Update wouldn't work.
ActiveX isn't off: it's disabled for the "Internet Zone", but it's still vulnerable to cross-zone attacks.
Powerbook keyboards are terrific.
Unless there's something truly amazing happening in some version more recent than this past February they're OK. Not great, just OK. If you want "great" try an IBM Thinkpad, and if you want "terriffic" try an old Toshiba from back when they were putting full-size mechanisms in their laptops (486, early Pentium).
some limited do-nothing account
If someone can log in to a "limited account" one can attempt local exploits to escalate privileges. Without that one is limited to remote exploits.
To a certain part of the criminal justice system, there are no innocent people... merely people for whom it has not yet been established what they are guilty of.
I wish I was joking.
I think it prefectly reasonable, given the shortcomings of the AirPort drivers. You set up a very limited account and set it to auto-login.
That wouldn't pass even the most cursory security review. I rate that "not acceptable".
The other option is don't use AirPort.
It appears that's the only option.
But they don't, always, in fact they've varied significantly from 100dpi in the past. And with Quartz Extreme there's really no reason to stick to any specific pixel density... their scaling technology is better than just about anyone's.
I suspect that they stick to 100dpi for the same reason iPods have such powerful output (even my shuffle can drive unpowered external speakers)... Steve Jobs likes it that way, they're well adapted to *his* hearing and visual acuity. For someone with significantly better than average acuity Apple's screens are painfully jaggy.
They're targeting consumers who have little motivation to understand much more than "point-and-click". That being the case, I don't think the article is necessarily an instance of FUD-spreading.
Sure it is, because those "point and click" users are not being helped by installing antivirus software when there's no viruses to catch. Until there are, all Norton or any other AV will do will be to slightly increase the risk of data loss by making their system a little less stable.
It's not quite snake oil, but it's close.
The only exploit they point to is a rootkit... which is something you install *after* you've exploited the box... there are no active threats that any antivirus software will work aaginst.
This is like their attempt to talk up a manually-installed program that deleted all your files on the Palm as an exploit, to push their useless PalmOS antivirus. And then their Pocket PC antivirus actually caused people data loss from false alarms.
Until there's an active threat in the wild, AND it's been analysed and an identifying signature discovered, antivirus software's only result is to make your computer less stable and less reliable because of its deep hooks in the OS.
This is not to say that the OS is magically perfectly secure, but anything any AV company tells you about ANY platform but Windows, at the moment, should be taken with a sackful of salt.
You do realize that you can setup an account to automatically login don't you?
I can't believe anyone would actually suggest this as a reasonable workaround. It's at best a necessary evil to be bypassed as soon as possible.
HOWEVER...
remember, everything the SystemStarter and login runs can be run from a script, so you should be able to bring up enough to do wireless without logging in with some poking about...
I love my mini.
I'm really holding off getting a laptop.
Why?
Mediocre (but improving, they used to be horrid) keyboards, one-button touchpad, and low resolution screens.
Anything above the base 12" iBook should have at least 1280x1024, and the 17" model should be more like 1920x1200. My 14" Thinkpad has more pixels than the 17" Powerbook.
It's not worth $500 unless you need the small form factor.
Or unless you want a computer to use instead of work on.
And how many of them swear off Windows?
Swear off Windows, or swear at Windows. There is no third choice.
Yeah, um, which is why you only send them half of the key.
If you don't send them the whole key, then there's no point in sending the DRM-protected self-destructing AI. And, for that matter, there's no point in sending the encrypted data at all, since the latency of the link is so high (years) that latency far dominates bandwidth that you're not saving anything.
If the bandwidth is so low that latency doesn't dominate bandwidth, then you can't afford the overhead of the self-destructing AI in the first place.
Thanks for the link, that puppy's going in my Palm.
Mostly it's for the digitized signatures, but the paper signatures are scanned and digitized as well.
My logic is that if someone ganks a file of signatures off a poorly secured server somewhere, they'll see mine's tagged and figure that I'm an anal-retentive asshole and go rip off someone else instead.
Ender's War (the short story) was a much better story than the novel Card expanded it into. Plus, it'd be easier to fit into 2 hours on the screen. Pity they didn't pick it instead.
The ambuiguity at the end over just who the Enemy was is wonderful - see, there's no aliens in there, and the one reference in the short story to the planet Ender's living on implies that it's noth Earth, so it COULD be a rebellious colony... which would make the Enemy planet Earth.
Whoops.