Why does QT need to access databases, instead of programs doing it themselves?
Perhaps for the embedded people.
X does multi-monitor. Tell me why QT needs to
For the embedded people.
No, you're definitely right on this one. Add support for.0001 % of the linux population so they can run X a little quicker. Good usage of time
Actually I think it's called future-proofing. You might want to look into it. Cleaning up the code and ensuring it works on 64 bit platforms is a good thing to do.
I can see this more then the others, but I still believe it can be implemented outside of QT.
As another poster already stated, the toolkit needs to know about this for obvious reasons. Every international-ready toolkit I can think of has this.
Anyway I can see exactly why they're doing SQL and further distancing themselves from X. As an embedded programmer I'd love for QT to be able to give me a backend-independent SQL interface. DBI rox muh sox but Perl is a pig. That's important when your processor gets a whole 4 bogoMIPS.
If he bought it in '84 or '85 he must have worked at Intel. As I recall Compaq was the first with a 386 box and they brought it out in 1986.
I'll have to reboot it tonight to see what the BIOS says. I'm pretty certain it's 1985 (I used to have a printscreen cap of the bios settings showing June 1995 I believe) but you may be right here... probably are.:-)
Mhh, would that box be worth anything at an auction? or would it get treated like the rest of the landfill x86 clones?
Probably landfill.:-) I've still got the original 120M (I think) HDD, but it doesn't work.
Computers don't really wear out, my master browser at home is a 486/33 running FreeBSD, it must be 7 year old at least. But I'm getting ready to rotate it out and replace it was a P90.
I'm in a similar boat. I bought one of the first commercially available 80386DX/33s from my dad about 4 years ago. He'd bought it new in '84 or '85 and spent a fortune on it. It's my firewall but may have to finally get laid to rest since it doesn't seem to like the PCMCIA bridge I put in it for my wireless network.
The funny thing is that this computer is one of the original 80386DX motherboards. The cache logic (64k!) is done in discrete GALs and PALs. 16M is the total amount of memory you can put on it (4M in DIP, 4M in SIPP and 8M in 30-pin SIMM on a special riser card). I've never replaced the power supply, supply fan nor the 3.5" floppy drive (although I think that drive is a year or two newer than the system) and while the BIOS thinks that the year is 1901, the system works perfectly. Theoretically it should have given up the ghost long ago: DIP chips tend to become unseated and a motherboard that BIG and having that many solder connections should have had something go by now. But no... I have to replace P2 systems after a couple years of work but this old DTK system keeps on going and going.
Maybe those old systems were just engineered "better". I say better because as a consumer it's never given out on me but, as a systems designer myself, I know that truly engineering something includes End of Life calculations and having a design survive past that is an indication of underengineering.:-)
It is a very popular misconseption that milk is good for you. This is just bulshit and dairy farmers proapaganda. Go and ask a biochemist how hard it is to digest caseine for a human that is older then 16-18 months
Um... that's why Vitamin D is added? Hello? Get out in the Sun and get some D that way too.
If you don't believe me, you go ask a biochemist. Milk is very good for you and not just for the calcium. And No, I'm not a dairy farmer or a dairy farmer's propagandist.
Perhaps it's losing my DSL line, perhaps I don't get out enough, but until this story I'd never even heard of these guys.
Exactly. And I don't even have high speed access.:-(
I don't mind people asking for money for something they do. I've never heard of them so I really doubt I'll be affected.
Now if Slackware were to start charging for distribution I would consider that alright as well. I grab an.iso once every 18 to 24 months or so so it's not like it's a major thing. And since I know what I'm getting with Slackware, I'll likely pay for it, which is something I don't think I'd do with these guys. Pay $15 for a distro I've never heard of? No thanks.
with biometrics you can't reissue a retina. If someone gets an face scan of an employee they've got a login forever.
That's why most sane security measures take a combination of Things You Have and Things You Know. So a sane biometric-based security measure would be a face/eye/voice/fingerprint scan and a short (5-6 digit?) passcode. In the unlikely event that you lose the face/eye/voice/finger the system can train a new one through adminstrative or "group of knowns" means.
In the end, it looks like electromagnetism has the ability to produce more instantaneous force than ye olde combustion engine.
Well DUH... Have you fallen off the stupid bus or were you just trying to troll?
The AC induction motor has an extraordinarly steep (almost vertical) speed-torque curve when it's near full speed. Now take into account that variable frequency drives (VFDs) keep the Volts/Hertz ratio in check at all times and you find that the AC induction motor, when being driven by a VFD, is "always" at full speed and is therefore always capable of delivering rated torque. An internal combustion engine can't even come close to this kind of performance.
The speed/torque curve of most internal combustion engines looks like an upside-down 'u' -- i.e. your maximum torque is somewhere in the middle of its speed range. You need gearing to keep the engine in this maximum torque range or you're just wasting fuel and heating things up. They are also mechanically complex and require more maintenance compared to an AC induction motor. Internal combustion engines burn gasoline which has one of the highest energy densities of any available consumer fuel around which is why they're everywhere.
Inverters (variable frequency drives) are solid-state, capable of regenerating the energy from the motor slowing down to brake and are relatively efficient so long as you don't mind a bit of a whine from the IGBTs or FETs switching. An AC induction motor is extraordinarily efficient (the shitty little 3HP one (which could probably fling my little Jeep around at a good clip) sitting beside me here is 87% efficient. What's the most efficient readily-available internal combustion engine? 50%? 60%?
The AC induction motor is simply amazing from almost every perspective. No brushes, almost zero maintenance (greasing only really), great efficiency and decent size/power ratios. Electromagnetism has always been way more capable than the internal combustion engine when it comes to producing instantaneous force. It's only recently that battery and hybrid technology has been developed to take advantage of them in automotive use.
Supports everything but DirectX from what I've heard.
You forgot a few things:
NETWORKING
Scrollwheel mouse
Cut'n'paste between host and guest
Win95/98 only (no NT)
USB, DVD, PCMCIA (must be done through host)
Don't get me wrong, I use it every day as I believe it is faster than VMWare and I can customize my screen size (960x768 gives me my wharf on the right side). The networking hack is very poor and its biggest flaw, IMO. It would have been MUCH better just to emulate a network device and require a seperate IP for the guest like VMWare. I get no Network Neighbourhood, no ICMP and flaky TCP/UDP. DreamWeaver's FTP client crashes under Win4Lin.
Don't tell me to use Samba and mount under Linux. That's hokey as hell and doesn't solve some of my Windows Networking problems.
Win4Lin also uses the host filesystem (the windows filesystem is just a subdirectory). This is great and terrible at the same time.
I keep a seperate card specifically for online transactions. It has a woefully small credit limit so I'll never be out by more than I can afford.
Almost every credit card I know of does not hold you responsible for transactions on the card if stolen. When my card was stolen I paid $50 (Maximum liability) even though there was over $8k put on the card.
For God's sake, if you are concerned about security, don't broadcast your messages over the air! If you're f'n paranoid, use fibre, as it has no magnetic field around it to be intercepted.
Remember that you can bend a piece of fiber enough to pick up a tiny bit of signal which can't make the bend and read that. Without being detected.
Granted, fiber probably gives you the best protection today if you must have secure data in a network. Strong encryption over copper really isn't all that bad though. And I really don't know what's all that wrong about strong encryption over wireless. Both copper and wireless will give the snoop something to read but if your encryption is good enough (and I don't believe that 128 bit is) the snoop will have your data long after it's usefullness is past.
Plus, let's not forget that 128-bit gear isn't cheap.
Uhhh... compared to what? There's only like $16 difference between the silver (64) and gold (128) bit encrypted versions of the Lucent (Orinoco) wireless cards. Silver is $133.76 and Gold is $149.76 from CDW, not exactly the cheapest place on the planet.
In fact, I'm picking up a half dozen of them and a few PCMCIA-PCI bridge cards this Saturday.
man that is bad, what is up with that big LEAF?????
While criticism is welcome, I need more to go on. And the leaf? Easy: I'm Canadian. Sorry about the link to the Government of Canada, I don't feel like digging up a witty link this morning.:-)
I too prefer Postgres but I've given up on PHP; I do all my scripting in Perl now. I'd been looking for arguments to keep my PHP skills up but haven't found anything besides "just to know another language" so I dropped it. Perl just works for me... Heaven forbid if I were an actual Perl zealot.:-)
Mind you, for any real work I turn to C. Interpreted languages are great for quick hacks but when performance is key you need compiled code. Maybe that's just my embedded background talking though.:-)
Twice in AM did I temporary lock out two different IT Specialists.
Huh?
One was because he going to porn sites, then, deleting the records in the ADC of it, and the other was because he was viewing other people's emails.
Okay, these are reasons for termination, not pros for making people sign non-competetion agreements.
Their contracts allowed me to suspend them upto three months at only 50% advanced pay and that they weren't allowed to work in the same field during that time.
That's all fine and dandy... I really don't know what that last point has to do with getting terminated/suspended for doing something wrong.
Here is my current "general" resume. It fits into two pages but remains fairly general. I have a broad range of interests and abilities and I had the same problem you did... Polished it up and boom... 3 pages.
What I ended up doing was keeping the actual job details kind of skimpy and putting my abilities and experiences up top. I also dropped off a few older jobs which didn't really show too much.
Lastly, I stuck on some keywords at the very bottom to help bots find the right resume. I get about two to three requests a week for more info so I think it seems to be working alright.:-)
When I get a request for more info they usually ask for the resume in Word.doc format. What I do then is adjust the resume to target the job more specifically.
A coil of wire is an RC circuit (actually RLC)... Take some EE courses before you spout your half-baked sys-admin Electrical knowledge.
As someone who designes embedded systems for a living I feel I have every right to tell you to go fuck yourself. The original writer said that the cashier burnt out the circuit (implying that the device doesn't work anymore once it is out the door. This is incorrect. I am not spouting any "half-backed sys-admin Electrical knowledge" as you put it. Yes a coil of wire is an inductor. Two tips of wire between a dielectric is a cap and all these things have parasitic inductances, capacatances and resistances associated with them. My original statement (paraphrased: "I don't think they're burning out anything") hasn't changed. Give me proof that I'm wrong or shut the fuck up.
Chances are, you can do this the same way they handle the anti-theft "stickers" on CDs and such. They work by the same principle (inductive coupling) but have a simple RC circuit at the center of the spiral antenna. In a weak RF field, they couple to the field and give a detectable signal. when the store runs your CD over the eraser, it generates a moderate strength RF field that burns out the RC ciruit.
Not the ones I've seen. They are simply a coil of wire and no circuit at all. The store (de)magnetize the backing and that changes the properties of the sticker. If you take a coil of household electrical wire (buy the small pack) and walk through the sensor it will go off. I've done this and can verify it.
Libraries use something very close to this: They put the coil of wire sticker in the return card pocket and then use a metallized return card. When you put the return card in the pocket it changes the properties of the coil and the sensor doesn't go off. The library I frequent used to use a single strip in the spine and magnetize/demagnetize it repeatedly.
Me, if it can't come from `apt-get -b source foo' then it doesn't get installed.
Exactly why I dislike package tools of any kind.
Why should I wait for debian (or software writer X) to come out with packages? Grab the source tarball and install it. Hell there are even a couple of really nice "install" replacements which log what got put where and at what time for nice and easy removal.
Requiring everything to be an RPM in order to keep the RPM database happy is what turned me off of RedHat (well that and the idiotic system config structure). The exact same thing keeps me off of Debian and its derivatives. Package management must be as low-level as possible or it is useless.
Glibc 2.0 is NOT glibc 2.1 which is NOT glibc 2.2. The glibc guys fucked up, this is true. I would have called these libraries something different (glibc2/3/4?) This is easy to fix though: I can install all three of these libraries and the softare can link up with whatever lib it wants. You can't do that (properly) under the Win* systems. I wouldn't call this DLL hell at all. A nuisance, yes. A potential security risk (having old libs with old holes), yes but I think that can be worked out (proper perms?) The point is this is a rather simple problem, and certainly workable if need be.
My next point is that you are really bitching without reason. When an app asks for glibc2.1, bloody well give it glibc2.1. If you either can't or won't, don't bitch at the software. Are you also the kind of person who demands his car be fixed and then insists on putting Toyota parts into a Honda?
Nothing will protect us... from idiots and those who take advantage of them..... people can't be bothered to set there security settings and everytime they get ANYTHING, they just double-click it, infect the file servers, and infect me even though I'm careful...
Well obviously you aren't being careful enough.:-)
Protection can be helped through gentle enforcement. Make all the default settings secure and pop up a big nasty warning if they go to disable something. Yeah I know they just click-through but it's a start.
Then implement procmail security on incoming mail. I don't have a link handy but it is a beautiful little procmail script which defangs emails, optionally scans, all kinds of nice things. Next, enforce security a little more strongly by refusing to turn off such protection and making sure that the default policy for viewing attachments is through a virus scanner.
Back to the world of calendaring: there is no need whatsoever to have scripting of any kind. Simple objects such as "allocate resource" (meeting room, projector, etc.) and "allocate person" and have the client able to send authorization off to a third party (i.e. a secretary) if you desire. From these two base objects you can span out and do more and fo course... if someone decides to make a whiz-bang object your client can always not install it or the administrator can refuse to allow clients to install support for them.
This remark is kind of flippant; I'm not saying I just designed the perfect groupware system in three or four paragraphs; just that your security negativity is unfounded.
If you can't find the SCSI emulation, that's your problem. But if you need help getting the IDE system to give up a drive so it can be picked up as SCSI, look no further than LILO. "append hdc=ide-scsi" works for me.
Mission critical is not tracking trouble tickets. You could do that in a flat file.
You talk the talk... But your knowledge is next to zero.
Mission critical is anything that will cost your company big bucks if it fails. Doesn't matter if it's your financial systems or your coffeemaker.
Simple trouble ticketting software can run with flat files, sure, but when you're running your whole customer service department off of it and it ties in with your companies engineering and production departments the flat file system tends to fall a bit short. If you can't understand that it's your own problem but don't go spouting off that it's not a mission critical system.
UIDs aren't just required for shell access and mail. Most ISPs provide web space now, and some mechanism to allow the user - and that user alone - to modify the content.
I addressed this in my original post: Proftpd and its own passwd file. Set all the user accounts to some arbitrary UID/GID (I created a web user which apache runs as) and the user's directory set to/www/users/[username]. Proftpd is then configured with "Userdir ~" or some such thing (It's late here) and when the user logs in they are scooted off to their own directory and can't get below that. Not exactly a chroot but then again this isn't shell we're talking about.
Yeah if they compromised proftpd they could get access to and fuck with my apache install but then again, that is why I'm running proftpd and not wu-ftpd.
The retail computer market is extremely cut-throat.
My brother did this for a year. I warned him about this exact thing. And it's truer than you can possibly believe.
He tried going the support angle. It works extraordinarly well for some customers (i.e. the ones you want) but horribly bad for most. Truth be known, most people have someone they can call (usually to that someone's chagrin) when they need help; they don't want you. You will not make money on hardware unless you stock and can therefore be ready for the guy who comes rushing in at 10 to 6 looking for a widget. You can't make money when you stock because prices drop quickly on things like RAM, drives, processors and video cards. Prices tend to stay stable on cases, power supplies, fans, peripherals (CD-ROMs, floppies, keyboards and mice) and consumables (paper, toner and ink). Your support time usually gets eaten up by someone who just...won't...go...away but who doesn't think they really have a problem. Your store usually attracts a couple guys on welfare who just hang around and chatter all day. The list is endless.
My brother managed to make a bit of money because he ended up selling to the office supply store in town. Hell he did alright even though there were two other computer stores like his in the 11k-person city, but he couldn't make enough to live on.
Very bad business to be in. You've got to either have great connections or incredible marketing. Probably both.
Slashdot Mirror
Why does QT need to access databases, instead of programs doing it themselves?
Perhaps for the embedded people.
X does multi-monitor. Tell me why QT needs to
For the embedded people.
No, you're definitely right on this one. Add support for .0001 % of the linux population so they can run X a little quicker. Good usage of time
Actually I think it's called future-proofing. You might want to look into it. Cleaning up the code and ensuring it works on 64 bit platforms is a good thing to do.
I can see this more then the others, but I still believe it can be implemented outside of QT.
As another poster already stated, the toolkit needs to know about this for obvious reasons. Every international-ready toolkit I can think of has this.
Anyway I can see exactly why they're doing SQL and further distancing themselves from X. As an embedded programmer I'd love for QT to be able to give me a backend-independent SQL interface. DBI rox muh sox but Perl is a pig. That's important when your processor gets a whole 4 bogoMIPS.
If he bought it in '84 or '85 he must have worked at Intel. As I recall Compaq was the first with a 386 box and they brought it out in 1986.
I'll have to reboot it tonight to see what the BIOS says. I'm pretty certain it's 1985 (I used to have a printscreen cap of the bios settings showing June 1995 I believe) but you may be right here... probably are. :-)
Mhh, would that box be worth anything at an auction? or would it get treated like the rest of the landfill x86 clones?
Probably landfill. :-) I've still got the original 120M (I think) HDD, but it doesn't work.
Computers don't really wear out, my master browser at home is a 486/33 running FreeBSD, it must be 7 year old at least. But I'm getting ready to rotate it out and replace it was a P90.
I'm in a similar boat. I bought one of the first commercially available 80386DX/33s from my dad about 4 years ago. He'd bought it new in '84 or '85 and spent a fortune on it. It's my firewall but may have to finally get laid to rest since it doesn't seem to like the PCMCIA bridge I put in it for my wireless network.
The funny thing is that this computer is one of the original 80386DX motherboards. The cache logic (64k!) is done in discrete GALs and PALs. 16M is the total amount of memory you can put on it (4M in DIP, 4M in SIPP and 8M in 30-pin SIMM on a special riser card). I've never replaced the power supply, supply fan nor the 3.5" floppy drive (although I think that drive is a year or two newer than the system) and while the BIOS thinks that the year is 1901, the system works perfectly. Theoretically it should have given up the ghost long ago: DIP chips tend to become unseated and a motherboard that BIG and having that many solder connections should have had something go by now. But no... I have to replace P2 systems after a couple years of work but this old DTK system keeps on going and going.
Maybe those old systems were just engineered "better". I say better because as a consumer it's never given out on me but, as a systems designer myself, I know that truly engineering something includes End of Life calculations and having a design survive past that is an indication of underengineering. :-)
It is a very popular misconseption that milk is good for you. This is just bulshit and dairy farmers proapaganda. Go and ask a biochemist how hard it is to digest caseine for a human that is older then 16-18 months
Um... that's why Vitamin D is added? Hello? Get out in the Sun and get some D that way too.
If you don't believe me, you go ask a biochemist. Milk is very good for you and not just for the calcium. And No, I'm not a dairy farmer or a dairy farmer's propagandist.
Perhaps it's losing my DSL line, perhaps I don't get out enough, but until this story I'd never even heard of these guys.
Exactly. And I don't even have high speed access. :-(
I don't mind people asking for money for something they do. I've never heard of them so I really doubt I'll be affected.
Now if Slackware were to start charging for distribution I would consider that alright as well. I grab an .iso once every 18 to 24 months or so so it's not like it's a major thing. And since I know what I'm getting with Slackware, I'll likely pay for it, which is something I don't think I'd do with these guys. Pay $15 for a distro I've never heard of? No thanks.
with biometrics you can't reissue a retina. If someone gets an face scan of an employee they've got a login forever.
That's why most sane security measures take a combination of Things You Have and Things You Know. So a sane biometric-based security measure would be a face/eye/voice/fingerprint scan and a short (5-6 digit?) passcode. In the unlikely event that you lose the face/eye/voice/finger the system can train a new one through adminstrative or "group of knowns" means.
In the end, it looks like electromagnetism has the ability to produce more instantaneous force than ye olde combustion engine.
Well DUH... Have you fallen off the stupid bus or were you just trying to troll?
The AC induction motor has an extraordinarly steep (almost vertical) speed-torque curve when it's near full speed. Now take into account that variable frequency drives (VFDs) keep the Volts/Hertz ratio in check at all times and you find that the AC induction motor, when being driven by a VFD, is "always" at full speed and is therefore always capable of delivering rated torque. An internal combustion engine can't even come close to this kind of performance.
The speed/torque curve of most internal combustion engines looks like an upside-down 'u' -- i.e. your maximum torque is somewhere in the middle of its speed range. You need gearing to keep the engine in this maximum torque range or you're just wasting fuel and heating things up. They are also mechanically complex and require more maintenance compared to an AC induction motor. Internal combustion engines burn gasoline which has one of the highest energy densities of any available consumer fuel around which is why they're everywhere.
Inverters (variable frequency drives) are solid-state, capable of regenerating the energy from the motor slowing down to brake and are relatively efficient so long as you don't mind a bit of a whine from the IGBTs or FETs switching. An AC induction motor is extraordinarily efficient (the shitty little 3HP one (which could probably fling my little Jeep around at a good clip) sitting beside me here is 87% efficient. What's the most efficient readily-available internal combustion engine? 50%? 60%?
The AC induction motor is simply amazing from almost every perspective. No brushes, almost zero maintenance (greasing only really), great efficiency and decent size/power ratios. Electromagnetism has always been way more capable than the internal combustion engine when it comes to producing instantaneous force. It's only recently that battery and hybrid technology has been developed to take advantage of them in automotive use.
Supports everything but DirectX from what I've heard.
You forgot a few things:
Don't get me wrong, I use it every day as I believe it is faster than VMWare and I can customize my screen size (960x768 gives me my wharf on the right side). The networking hack is very poor and its biggest flaw, IMO. It would have been MUCH better just to emulate a network device and require a seperate IP for the guest like VMWare. I get no Network Neighbourhood, no ICMP and flaky TCP/UDP. DreamWeaver's FTP client crashes under Win4Lin.
Don't tell me to use Samba and mount under Linux. That's hokey as hell and doesn't solve some of my Windows Networking problems.
Win4Lin also uses the host filesystem (the windows filesystem is just a subdirectory). This is great and terrible at the same time.
I keep a seperate card specifically for online transactions. It has a woefully small credit limit so I'll never be out by more than I can afford.
Almost every credit card I know of does not hold you responsible for transactions on the card if stolen. When my card was stolen I paid $50 (Maximum liability) even though there was over $8k put on the card.
For God's sake, if you are concerned about security, don't broadcast your messages over the air! If you're f'n paranoid, use fibre, as it has no magnetic field around it to be intercepted.
Remember that you can bend a piece of fiber enough to pick up a tiny bit of signal which can't make the bend and read that. Without being detected.
Granted, fiber probably gives you the best protection today if you must have secure data in a network. Strong encryption over copper really isn't all that bad though. And I really don't know what's all that wrong about strong encryption over wireless. Both copper and wireless will give the snoop something to read but if your encryption is good enough (and I don't believe that 128 bit is) the snoop will have your data long after it's usefullness is past.
Plus, let's not forget that 128-bit gear isn't cheap.
Uhhh... compared to what? There's only like $16 difference between the silver (64) and gold (128) bit encrypted versions of the Lucent (Orinoco) wireless cards. Silver is $133.76 and Gold is $149.76 from CDW, not exactly the cheapest place on the planet.
In fact, I'm picking up a half dozen of them and a few PCMCIA-PCI bridge cards this Saturday.
man that is bad, what is up with that big LEAF?????
While criticism is welcome, I need more to go on. And the leaf? Easy: I'm Canadian. Sorry about the link to the Government of Canada, I don't feel like digging up a witty link this morning. :-)
I too prefer Postgres but I've given up on PHP; I do all my scripting in Perl now. I'd been looking for arguments to keep my PHP skills up but haven't found anything besides "just to know another language" so I dropped it. Perl just works for me... Heaven forbid if I were an actual Perl zealot. :-)
Mind you, for any real work I turn to C. Interpreted languages are great for quick hacks but when performance is key you need compiled code. Maybe that's just my embedded background talking though. :-)
Twice in AM did I temporary lock out two different IT Specialists.
Huh?
One was because he going to porn sites, then, deleting the records in the ADC of it, and the other was because he was viewing other people's emails.
Okay, these are reasons for termination, not pros for making people sign non-competetion agreements.
Their contracts allowed me to suspend them upto three months at only 50% advanced pay and that they weren't allowed to work in the same field during that time.
That's all fine and dandy... I really don't know what that last point has to do with getting terminated/suspended for doing something wrong.
Here is my current "general" resume. It fits into two pages but remains fairly general. I have a broad range of interests and abilities and I had the same problem you did... Polished it up and boom... 3 pages.
What I ended up doing was keeping the actual job details kind of skimpy and putting my abilities and experiences up top. I also dropped off a few older jobs which didn't really show too much.
Lastly, I stuck on some keywords at the very bottom to help bots find the right resume. I get about two to three requests a week for more info so I think it seems to be working alright. :-)
When I get a request for more info they usually ask for the resume in Word .doc format. What I do then is adjust the resume to target the job more specifically.
A coil of wire is an RC circuit (actually RLC)... Take some EE courses before you spout your half-baked sys-admin Electrical knowledge.
As someone who designes embedded systems for a living I feel I have every right to tell you to go fuck yourself. The original writer said that the cashier burnt out the circuit (implying that the device doesn't work anymore once it is out the door. This is incorrect. I am not spouting any "half-backed sys-admin Electrical knowledge" as you put it. Yes a coil of wire is an inductor. Two tips of wire between a dielectric is a cap and all these things have parasitic inductances, capacatances and resistances associated with them. My original statement (paraphrased: "I don't think they're burning out anything") hasn't changed. Give me proof that I'm wrong or shut the fuck up.
Chances are, you can do this the same way they handle the anti-theft "stickers" on CDs and such. They work by the same principle (inductive coupling) but have a simple RC circuit at the center of the spiral antenna. In a weak RF field, they couple to the field and give a detectable signal. when the store runs your CD over the eraser, it generates a moderate strength RF field that burns out the RC ciruit.
Not the ones I've seen. They are simply a coil of wire and no circuit at all. The store (de)magnetize the backing and that changes the properties of the sticker. If you take a coil of household electrical wire (buy the small pack) and walk through the sensor it will go off. I've done this and can verify it.
Libraries use something very close to this: They put the coil of wire sticker in the return card pocket and then use a metallized return card. When you put the return card in the pocket it changes the properties of the coil and the sensor doesn't go off. The library I frequent used to use a single strip in the spine and magnetize/demagnetize it repeatedly.
2^32 generates an overfflow error. Therefore you get spanked.
I guess that's what you get for using an ancient 32 bit system these days.
Me, if it can't come from `apt-get -b source foo' then it doesn't get installed.
Exactly why I dislike package tools of any kind.
Why should I wait for debian (or software writer X) to come out with packages? Grab the source tarball and install it. Hell there are even a couple of really nice "install" replacements which log what got put where and at what time for nice and easy removal.
Requiring everything to be an RPM in order to keep the RPM database happy is what turned me off of RedHat (well that and the idiotic system config structure). The exact same thing keeps me off of Debian and its derivatives. Package management must be as low-level as possible or it is useless.
Glibc 2.0 is NOT glibc 2.1 which is NOT glibc 2.2. The glibc guys fucked up, this is true. I would have called these libraries something different (glibc2/3/4?) This is easy to fix though: I can install all three of these libraries and the softare can link up with whatever lib it wants. You can't do that (properly) under the Win* systems. I wouldn't call this DLL hell at all. A nuisance, yes. A potential security risk (having old libs with old holes), yes but I think that can be worked out (proper perms?) The point is this is a rather simple problem, and certainly workable if need be.
My next point is that you are really bitching without reason. When an app asks for glibc2.1, bloody well give it glibc2.1. If you either can't or won't, don't bitch at the software. Are you also the kind of person who demands his car be fixed and then insists on putting Toyota parts into a Honda?
Nothing will protect us ... from idiots and those who take advantage of them..... people can't be bothered to set there security settings and everytime they get ANYTHING, they just double-click it, infect the file servers, and infect me even though I'm careful...
Well obviously you aren't being careful enough. :-)
Protection can be helped through gentle enforcement. Make all the default settings secure and pop up a big nasty warning if they go to disable something. Yeah I know they just click-through but it's a start.
Then implement procmail security on incoming mail. I don't have a link handy but it is a beautiful little procmail script which defangs emails, optionally scans, all kinds of nice things. Next, enforce security a little more strongly by refusing to turn off such protection and making sure that the default policy for viewing attachments is through a virus scanner.
Back to the world of calendaring: there is no need whatsoever to have scripting of any kind. Simple objects such as "allocate resource" (meeting room, projector, etc.) and "allocate person" and have the client able to send authorization off to a third party (i.e. a secretary) if you desire. From these two base objects you can span out and do more and fo course... if someone decides to make a whiz-bang object your client can always not install it or the administrator can refuse to allow clients to install support for them.
This remark is kind of flippant; I'm not saying I just designed the perfect groupware system in three or four paragraphs; just that your security negativity is unfounded.
If you can't find the SCSI emulation, that's your problem. But if you need help getting the IDE system to give up a drive so it can be picked up as SCSI, look no further than LILO. "append hdc=ide-scsi" works for me.
Mission critical is not tracking trouble tickets. You could do that in a flat file.
You talk the talk... But your knowledge is next to zero.
Mission critical is anything that will cost your company big bucks if it fails. Doesn't matter if it's your financial systems or your coffeemaker.
Simple trouble ticketting software can run with flat files, sure, but when you're running your whole customer service department off of it and it ties in with your companies engineering and production departments the flat file system tends to fall a bit short. If you can't understand that it's your own problem but don't go spouting off that it's not a mission critical system.
UIDs aren't just required for shell access and mail. Most ISPs provide web space now, and some mechanism to allow the user - and that user alone - to modify the content.
I addressed this in my original post: Proftpd and its own passwd file. Set all the user accounts to some arbitrary UID/GID (I created a web user which apache runs as) and the user's directory set to /www/users/[username]. Proftpd is then configured with "Userdir ~" or some such thing (It's late here) and when the user logs in they are scooted off to their own directory and can't get below that. Not exactly a chroot but then again this isn't shell we're talking about.
Yeah if they compromised proftpd they could get access to and fuck with my apache install but then again, that is why I'm running proftpd and not wu-ftpd.
The retail computer market is extremely cut-throat.
My brother did this for a year. I warned him about this exact thing. And it's truer than you can possibly believe.
He tried going the support angle. It works extraordinarly well for some customers (i.e. the ones you want) but horribly bad for most. Truth be known, most people have someone they can call (usually to that someone's chagrin) when they need help; they don't want you. You will not make money on hardware unless you stock and can therefore be ready for the guy who comes rushing in at 10 to 6 looking for a widget. You can't make money when you stock because prices drop quickly on things like RAM, drives, processors and video cards. Prices tend to stay stable on cases, power supplies, fans, peripherals (CD-ROMs, floppies, keyboards and mice) and consumables (paper, toner and ink). Your support time usually gets eaten up by someone who just...won't...go...away but who doesn't think they really have a problem. Your store usually attracts a couple guys on welfare who just hang around and chatter all day. The list is endless.
My brother managed to make a bit of money because he ended up selling to the office supply store in town. Hell he did alright even though there were two other computer stores like his in the 11k-person city, but he couldn't make enough to live on.
Very bad business to be in. You've got to either have great connections or incredible marketing. Probably both.