I was rather like you when I was in undergrad (in the late 90's). I started out as a Math major (Operations Research) which required certain CS classes. As I learned more about CS, I found that there is a very rich mathematical basis for Computer Science -- from the theories of computation to graphics to algorithm analysis. Almost any serious PhD in CS involves a heavy dose of mathematics in one form or another. Think of it as applied mathematics, in a geeky twisted way:)
Part of what I'm saying is that you can do CS and not end up as a programmer, per se.
The other half of the equation is that there *are* significant (well-paying) jobs for mathematicians. Now, I doubt that you'd want to (or could) seriously pursue any of them with just a BS, but a PhD need not be a requirement. My S/O's employer has several math/statistics majors on staff who perform marketing analysis, trending, etc... some of it rather high-powered stuff. If you look into the Actuarial or Operations Research fields (if that floats your boat), there are awesome opportunities.
Whatever direction you choose, I strongly encourage you to go past a BS -- at least stay in school through an MS program. For one thing, it opens more doors down the road (I've gotten at least two jobs partially because I have an MS/CS). More importantly though, IMHO, it makes you a better professional; you learn a heck of a lot more in grad school than in undergrad -- at least that was my experience. You study your subject in far more depth and with far more rigor than in undergrad and you're treated more like a colleague than a student. It's an awesome experience, particularly if your're more of the geeky theoretical type:)
Whatever you do, make sure you enjoy it. Of course, you can always go back and get a second degree in underwater basket weaving or Anthropology, but it's a heck of a lot easier to get it right the first time. The sooner you identify a career path (at least vaguely), the better choices you can make in courses, internships, research focus, etc.
This is truly one of the wost ideas I've ever heard and it only highlights prevalent misconceptions with respect to firearms safety.
Modern guns, themselves, are not inherently unsafe. Guns do not spontaneously jump up and shoot people. Guns only discharge when the trigger is pulled; while this can occur unintentionally, e.g., due to a dog stepping on the trigger of a loaded, unsafed gun left laying on the ground (don't laugh, this does happen!), almost 100% of such incidents are due to NEGLIGENCE.
Nearly all incidents of unintended or illegal shootings are due to negligence, lack of training and practice, or intentional criminal activity. Negligence includes such actions as: allowing unauthorized access to a weapon by a minor, pointing a weapon at something other than a target or a safe downrange area, and placing one's finger on the trigger when the weapon is not pointed in a safe direction. Lack of training and practice leads to negligence; there are numerous incidents of police officers, who, in principle, should be some of the best-trained firearms handlers among us, who have shot themselves in the foot or leg while handling their own weapons.
At the end of the day, it is the person, the gun owner, who is responsible for safety. When a gun discharges, it is because of someone's actions; full stop. It's not the gun's fault and it's not the manufacturer's fault.
We also must remember that the purpose of most weapons -- handguns, assault rifles, tactical shotguns, etc. -- is for defensive or offensive use against other humans. Put more simply: they're designed to stop human adversaries, by injury or death. In principle, their use, particularly by civilians, should be very infrequent. I am a relatively highly-trained defensive shooter; I believe that I am capable of defending myself, my family, and my home, should the need arise. But I hope and pray that I never need to do so. I think that most police will tell you that they hope to have to shoot a suspect, but that they are trained and prepared to do so to protect others or themselves.
If and when, however, the time comes that a weapon is needed, one must be supremely confident in the reliability of the weapon. This means that simpler is necessarily better. When you pull the trigger, you want to hear "BANG", not "click" or "beep". You don't want to have to fiddle with magic decoder rings, tiny keys, batteries and secret codes, etc. in the dark, under pressure, with your child screaming in the background. And a cop can't be worrying about passwords and encrypted ammunition in the heat of a pursuit. He must know that his weapon will fire when he pulls the trigger -- he's betting his life on it.
The technology described in this article is just another way to make owning firearms more difficult and more expensive. Criminals, by definition, have no regard for laws. You can make all guns illegal and the bad guys will still find a way to arm themselves. Look at Chicago or Washington, DC for prime examples. Those of you in Austraila and Britian have seen a rise in violent crime, including home invasions and broad-daylight robberies, since you banned guns.
The technological achievement expressed in this article is impressive. I'm happy that people are exploring the uses of modern computer and cryptographic techniques. But be skeptical and wary as well... Your rights are at stake here.
Laptop theft is the least of SF's criminal problems. SF has the largest absolute population of homeless individuals in the whole country. Considering SF's relatively small footprint, this is astounding. With that many hungry, poor, desperate, and possibly substance-addicted people wandering around, crime in general will go up. It is true that laptops are a lucrative target and WiFi hotspots encourage a greater density of targets.
However, all this article says to me is that if you create an environment where vagrancy and homelessness is subsidized (both by public and private money), immigration and other laws are not enforced, and people are not permitted to defend themselves (handguns are now illegal in SF, and forget about getting a CCW in any metro. area in CA), then crime will thrive.
Laptop theft is but an unfortunate symptom of a much larger problem.
I get my internet service from Comcast. I don't use my Comcast email account, but my girlfriend needed an account, so I created her one under mine (you get like 5 aliases). Her first name happens to be 4 letters long and rather uncommon, so I was able to get her first choice, xxxx@comcast.net.
I created the account at maybe 10:00am. She logged in, for the very first time, at around 11:00am. By that time, having never ever used the account, she already had about 10-20 spam messages.
Bloody amazing.
I think that, next time around, I'll try a much longer user name, as several people have already suggested.
OnStar scares the hell out of me. I don't like the idea of paying Big Brother to monitor my car and its travels and travails.
I don't know how accurate the CSI and Law and Order shows are with respect to the ability of the police to use such systems to track vehicles and people, but it seems totally plausable. Regardless of whether they currently do it, they (OnStar, a hacker, the feds, etc.) could, in principle, log your every move. Good if your car gets stolen or you need an alibi. Bad if you, for example, commute through 'sketchy' neighborhoods, are trying to avoid a stalker, or are being persecuted by the Department of Homeland (in)Security.
Personally, I don't have OnStar in my vehicle, and I don't want it. If it becomes "standard equipment" on new cars, I'll want a way to disable it or remove the hardware from the car without disabling the vehicle itself.
The problem with many of these technologies is they allow third parties (GM, the cops, etc.) to collect information about you while you don't have access to said information. For example, try getting a list of the phone numbers that called your cell phone (caller-id blocked or not) -- the phone company won't give it to you. I know, I tried. But if the police get a warrant, they sure as hell can get this information.
I'd rather that certain information not be collected at all. But I'd be happier to accept that collection if I had full and free access to it. I want to know what they know about me. It shouldn't take a court order to access your own phone records. And if your driving is being monitored and possibly logged (or even if this is possible), there should be full disclosure of this fact and the resulting data.
In the mean time, for those of us who value liberty, freedom, and privacy, I'd suggest finding another car company.:)
I was shopping for a new laptop recently, to replace my still-working but very slow IBM Thinkpad A22p (800MHz).
I have tried time and again to get used to touchpads, but I never feel like I have good control with them. The experience is always frustrating.
So, my #1 criterion for a new laptop is a non-touchpad pointing device, a la IBM's trackpoint, i.e., the "nipple"-type pointing stick.
I looked far and wide on the net. Of the major vendors, I found that IBM, Dell, and Toshiba all still make laptops with pointing stick devices. Note that Dell and Toshiba only offer the "stick" on certain models. I have used both vendors' pointing devices and, while not as good as a TrackPoint, they're FAR better than a touch pad.
I ended up buying a new ThinkPad anyway, mostly because I have a bunch of accessories that are thinkpad-specific and also because I know that, like it or not, ThinkPads are some of the most durable, long-lasting notebook computers out there.
Does anyone out there know if it's possible to put an IBM keyboard (with the TrackPoint goodie in it) into another vendor's laptop? Is the keyboard form factor and interface standard?
All of FreeBSD, OpenBSD, and NetBSD are good distributions, IMHO, based around a solid kernel. However, the best bang for your buck (in this case, your time) is FreeBSD, I think.
FreeBSD is pretty easy to install and comes with an extensive collection of "ports" -- packages that you can download and install or build for your system. I think it supports a larger range of applications than the other BSD's which are targeted more towards server and headless situations.
Check our the FreeBSD website for lots of great documentation on installing and configuring your new system. I think you'll be fine with the hardware you're running (I've run it on very wimpy pentium and 486 boxes in the past).
If you're going to grab a distro, get either the last 4.x-STABLE build or one of 5.2 or 5.3-RELEASE. While it's fun to run the bleeding-edge last night's build stuff, it's probably not right for a BSD beginner.
You'll likely feel like a fish out of water for a few days using BSD, coming from a Linux background. Config files are in different places, the boot process is a little different, and so forth. However, I think you'll learn to appreciate its design, just as I'm sure you appreciate the way Linux works. It's a great OS for servers -- I use it for my email server at home.
IMHO, undergraduate CS programs really only teach you what you don't know. Unless you go to a slave-driving engineering school, where you do a ton of programming, it probably doesn't really matter where you go as long as you get a solid grounding in the fundamentals.
I would strongly recommend going to graduate school. From my own experience, the two years I spent in grad school were FAR more valuable in my development as an engineer than the years I spent in undergrad.
I'd view your undergrad education as a chance to become a well-rounded individual, a good citizen, etc. Take some history, philosophy, english, etc. classes. You're going to be staring at code for the next 25+ years, so have some fun now while you're young.
As far as getting a decent job when you get out, a degree from Stanford, MIT, CalTech, etc. may help, but a cheaper, easier, and probably better way is to get some industry experience while you're in college. Look for summer co-op/internship positions at your college's career center. I did several internships during undergrad and grad school. The more industry experience you get, the better idea you'll have as to whether you really want to be a programmer, what kind of work you want do to, and what kind of company you want to work for.
Remember that, in the long run, it doesn't matter what school you went to. 5-10 years down the line, employers will be looking at what jobs you've held, what experience and skill-sets you have, and so forth, not where you got your BS. Your intelligence, drive, ambition, and choices will govern your career, not your choice of university.
Getting an MS is also a good idea. It sets you above the rest of the candidates and, IMHO, gives you a level of professional maturity that takes years to develop in industry. You also get exposed to more advanced concepts and technologies, and you have the opportunity to start specializing and yet still maintain the option to change your mind.
The other factors that employers look at is your GPA, regardless of where you go to school. I'd rather hire a guy with a 3.9 from a state school than someone else with a 3.1 from an Ivy League-class school. We want to see smarts, consistency, and the ability to follow through on committments. The fact that you didn't take classes with some wazoo-famous professor is really irrelevant in the big picture.
Stay where you're at. Look at MS programs at top-25 CS schools. Do some internships. Work at the UNIX lab on campus. Look at undergrad research internships in your CS program. Keep your GPA up. Read, program, and do things on the side to broaden your skill-set. And, most of all, have fun!
1. Bright white text on a black background. Yellow or green also works. Less glare, good contrast, etc. Great for x-terms. Harder in Word. Find what works for you.
2. Dialup your font sizes and/or zoom in. My friends used to tease me about using "fonts for the blind" on my x-terms, but I like not squinting and leaning into the screen to read stuff. Yes, you cannot fit as much stuff on the screen, but what you can see is readable.
3. Bigger is better. Get at least a 21" CRT monitor or a 19" LCD. It helps make (2) easier -- you're less tempted to use small fonts to see everything you need to see.
4. LCD's seem to experience less glare, be clearer, flicker-free, and have a myriad of other benefits. I like them, but I still appreciate a nice, big flat-screen Trinitron CRT monitor. The moral of the story is to spend a few extra bucks for bigger, clearer, and higher-quality. Money comes and goes, but you only lose your eyesight once.
5. Your eye doctor can prescribe "computer glasses". For the older folks among us, this can be a huge help. The problem is that most computer users sit about 3' from the screen. This is too close for your distance/driving glasses but too far away for your reading glasses. Finding a mid-range prescription can make a huge difference and reduce eye-strain.
6. Don't work in the dark. Don't work under really bright lights. I find the best computer lighting to be just a little too dark to comfortable read by.
7. Follow all those ergonomics suggestions that your HR department hands out. For example, get up and walk around every hour or two. This gives you a chance to stretch, focus your eyes at a distance, etc. Drink more water -- it's good for you and you'll have to pee more, which makes you get up more. You'll actually be more productive, despite the breaks.
It almost seems like this should be a national policy rather than state-by-state. I guess that gets into states' rights issuesm, a topic on which one could expound for volumes.
I think the whole problem is compounded by the two-party system. I'm at the point where I think it would be wise to disband political parties all together. Candidates should run on their own platforms and not be forced to shoehorn themselves into a particular party's view of the world. Coalitions (say, in Congress) should be formed around individual issues, not along party lines. This might make the system slower, but it also breeds consensus and allows representatives to actually *represent* their constituencies (sp?).
Your point is well-taken. The system is the system and, according to the rules of that system, Bush was the winner.
Of course, this doesn't address the whole Florida isssue, which is another can of worms entirely. Without descending into an extended argument about hanging chads, the issues experienced in Florida underscore the way that small errors in tallying votes can really affect elections. It's my understanding that *any* voting method, be it punch cards, touch screens, color-in-the-box, or whatever, has some small, but non-zero, margin for error. In most cases, the outcome is not jeapordized by this errorbar, but, in some cases, it is. I am doubtful that introducing computers will solve the problem, but it might have a chance of reducing the errorbar.
Hey, all Nader had to do was get enough petitions signed in each state. He didn't. Regardless of what the Democrats think, say, or do, if enough people were in favor of Nader, he'd be on the ballot in 50 states, hands down.
It's easy to blame things on parties or liberals or conservatives, but sometimes a rose is just a rose. Nader didn't get enough signatures. He loses. That's it.
The reason that the states you mention are "battleground" states is that:
a) They carry a non-trivial number of electoral votes and b) They are potentially "swing" states, i.e., both candidates have a chance of winning them
You don't hear much about Kerry campaigning in, say, Georgia, even though it carries 15 electoral votes, more than twice that of, say Idaho (7 votes). That's because Georgia is going to vote for Bush.
Now, if the election were decided on percentage of the total vote or, say, by obtaining a majority (>50%) of the vote, then it might be worth it for Kerry to campaign in urban Georgia areas, such as Atlanta, where he might be able to pick up thousands of votes.
You also have all the little or low-population states who, perhaps, get more than their fair share of electoral votes. The minimum number of electoral votes allocated to a state is 3. So, states such as Alaska, Montana, Wyoming, DC, North and South Dakota, and Vermont each gets three votes, regardless of population. Is that fair? In a sense, each voter in such small states has control over a larger fraction of an electoral vote than does a voter in a large state, such as California.
But that's how the Founding Fathers designed the system and, at least for the 2004 election, that's what we're stuck with, like it or not. Personally, I'm undecided as to whether, if I had the power, I would move away from the Electoral College system or stick with it. It clearly has its advantages and disadvantages.
Fundamentally, however, I think any system in which one candidate can get 500,000 more votes than his opponent and still lose is somehow, fundamentally, messed up.
Another salient point comes from my own experience as a voter in North Carolina, where I lived during the 2000 election. At that time, there was no way in hell I was going to vote for Bush. However, he was polling at over 60% in NC, so, in essence, my vote was irrelavent. I could vote for Gore, but it wouldn't matter, because the majority of the state would have, metaphorically, drowned me out. How I ended up voting is not germane, but I think you can see the quandry one might be in.
I think this situation may lead some folks to apathy, i.e., not caring about or getting involved in the system, perhaps to the extent of not voting at all. If you live in a state where you are in the minority in re your political opinions, then your vote really doesn't count for presidential elections. That's just how it is.
Now, if we moved away from an Electoral College system, then all those "minority voters" (as defined in the previous paragraph) WOULD have their voices heard, since their votes would "count" towards the candidate's total. Somehow, that seems more democratic to me.
"Cisco Systems Inc., which acknowledged its popular routers were among those vulnerable, distributed software repairs and tips to otherwise protect large corporate customers."
[http://www.msnbc.msn.com/id/4788445/]
Cisco is pretty good about releasing fixes for exploits in their software as fast as possible.
Being big has it's plusses and minuses. On the down side, they're routing packets for 80%+ of the internet. On the plus side, they have an army of engineers and a fire lit under their asses to get these kinds of things fixed.
Windows XP (and ME, I think) have a feature called "System Restore". Basically, what it does is track changes to the registry, driver database, and other parts of the system. It takes snapshots of the system periodically and sometimes during the installation of hardware or certain drivers.
If you break something (as I have been known to do from time to time), you can "roll back" to a previous snapshot. In my experience, this works pretty well for solving certain problems.
I'm not sure if it tracks installed software or whether such software would automagically be removed by rolling back. But the snapshots must have some such information in them, if only we could extract it...:)
Why not? Because laptops aren't designed to be servers. They generally underperform similarly-classed full-size systems, are generally not designed to be always-on, have flakey hardware, etc., etc.
Furthermore, laptop harddrives are much slower than server-class storage and, in my experience, have a much lower life expectancy.
Finally, laptops dump a lot of heat out their bottoms and don't like to be stacked on atop another. If you try this, you're liable to end up with a fire or, at the very least, hardware failures, in short order.
I think "use the right tool for the job" applies here...:)
>>Go with 4U rack mount cases. They are cheap, and easy. They will fit all your standard hardware out there.
This is a good point. Dell, in fact, sells "workstations" and "servers" that can be configured either as tower or 4U rack-mounts. You basically bolt on a couple rack ears and you're good to go.
If you find the right size case, you can buy rack ears, drill a couple holes in the case (be careful about the metal bits!) and bolt them on. Rack-mount servers on a budget.
If you're not much of home machinist, you could consider rack shelves -- I highly recommend that you find shelves that bolt on to the front AND back for stability. You can stack your tower/desktop systems on their sides. This gets a little messy if you ever need to move the rack, and it takes up a hideous amount of space compared to 2U systems (a shelf + tower consumes basically 5-6U of space), but it's an alternative or a way to rack-adapt existing hardware.
Rack mounting your toys is great, IMHO. I don't do it (yet) at my home because I don't have very much stuff, but if I did, I would.
There are a number of companies making "shorty" racks (that's my term, don't Google for "shorty"). They are normal 19" racks, but sit only 30 or 36 inches high. One I seriously looked at for work had a door on the front, too.
If you just want to mount a couple 1-2U servers, a switch/router, etc., this would be a sweet way to go. You can tuck it undera desk or in a (ventilated) closet and forget about it.
Rack equipment (particularly computer) is more expensive, apples to apples, compared to non-rack hardware, but it's compact, often better-cooled (read: noisy), and probably bare-bones enough for your requirements. Many of the rack computers I've worked with also have full BIOS support for a serial/TTY console, so you can remotely access the computer, reboot it, etc. You can gang all the TTY's into a terminal server (Cisco makes a couple models) or use serial-to-USB adapters and run them off of one box (just make sure it stays up:) ).
Make sure you calculate how much rack space you need before buying a rack. Also, from my experience, you'll probably want to allow at least 20-40% extra for expansion, testing, etc. Also think about cooling and the cooling requirements of your hardware.
Another benefit of a small rack is that if you move or relocate, you just unplug the rack from the wall and move the whole bloody thing with you. No recabling your equipment or lost toys. And, if you put a locking door on the unit and bolt it to the floor, it's pretty secure too.
If this all pans out, let us all know and maybe post some pictures. I think this is a slick idea that more individuals and small businesses should seriously consider.
VNC will allow you to access your windows desktop from any routable computer that runs a VNC client. VNC runs on almost every OS, including Palm, I think.
You won't be able to see remote filesystems unless you can mount them on the windows machine -- for UNIX, you could use Samba... for other OS's, well, I don't know.
I actually do the opposite from what you want to do -- I run a headless FreeBSD box and then view the VNC (X) desktop remotely on a Windows machine (full-screen, no less). Performance is about the same as running locally and I can still use all my Office products (necessary for my job, sadly).
Best of all, VNC is free and open-source (does that make us for me using Windows?) >:-)
I disagree. They're great for sniffing packets. If you've got an ethernet-connected device that doesn't have a sniffer onboard, and you want to see what the heck it's doing, a hub is a handy tool to have on your shelf. I use them quite often to intercept traffic while debugging software and hardware at work.
They also allow you to run a trace on a separate machine, so as not to interfere with the unit-under-test.
Yes, you could use a monitor-port, but that assumes that your employer will spring for fancy managed switches AND will allow you log into them and mess around. Mine doesn't, but they're more than happy to buy me a $40 hub and a couple patch cables:)
I'm not a net. admin either, I'm a software engineer.
When you're writing network software, or software that uses the network, you often run into wierd and hard-to-debug problems. The task of finding the cause of these bugs if often expedited by looking at the packets on the wire.
For example, you think you're sending a particular pattern of bits (1's and 0's) -- that's what you think you coded in your program. But for some reason, the other end doesn't understand your packets. You could put a bunch of debugging statements in your program, recompile, and hope you can see the problem, or, you can simply sniff the packets and see what's really going out on the wire.
As another person mentioned, sniffing is also useful for reverse-engineering closed-source software that uses the network. That's how those guys implimented clients for AIM -- they just figured out what messages to send back and forth. (Sadly, AOHell decided to change the protocol every 2 minutes so the open-source clients don't work very well.)
As far as security, sniffing or analyzing traffic is one of best tools available to see what's passing through your network. It's analagous to the security cameras in the local stop-and-rob (gas station) or in a casino -- they let the security guys watch what's going on, review it after the fact, and find/identify the bad guys.
You know, Sun really pissed me off in this particular area. Snoop is basically tcpdump with all the options and syntax changed. The output is a little prettier, but it's essentially the same. Switching between the two used to give me no end of headaches. Why reinvent the darned wheel?
(Note: I'm not griping at you, and I love Solaris... Sun just does some things that really annoy me:) )
I guess I'm oldschool, but I still use tcpdump for most day-to-day things. It's handy, it's fast, and it runs on just about every OS (including Windows (google for windump)). The output is ugly, but once you get used to it, you hardly notice.
When I really need to analyze a stream or set of streams, or I'm going to be staring at packets for more than about 10 minutes, I switch to ethereal. Again, it's free, runs on most OS's (including Windows, again), and the GUI is a little clunky, but quite usable. As several people have mentioned, the capture filter syntax is identical to tcpdump. The display filter syntax is different and I find is a little tricky to get right, so I try to prefilter (or filter with tcpdump beforehand) as much as possible.
One handy feature is the ability to analyze certain types of streams, such as a TCP session (filter out the whole session and see all the data in one window) and SIP (analyze jitter, loss, extract audio session, etc.). It's also open-source, so if it doesn't understand some kind of traffic, you can write your own extension. I haven't had to do this yet, but I know people who have, and it seems easy enough for a compitent programmer.
My employer has a site license for WildPackets Etherpeek (it comes in several versions... I think we have one of the higher-end ones). Frankly, it's prettier than ethereal, but, at least for the debugging I do, provides very little extra functionality. The capture filters are embedded in a GUI which I find makes it hard to see how they're configured.
Etherpeek is pretty and may be easier for novices to use. But I wouldn't waste the money unless it has some quirky feature you just can't live without.
Something to keep in mind: often, the place where you capture packets is not where you'd like to analyze them. For example, I've had situations where I needed to sniff traffic on a remote server -- I had ssh access to the server (and root, of course:) ), but couldn't/didn't want to install all kinds of GUI tools, etc. This is where tcpdump really shines. You can capture to a binary file and read the file with tcpdump, ethereal, Etherpeek, and many other packages. As long as you can get the file off the machine, you can analyze the data.
There are also handy tools for managing and analyzing tcpdump files, such as tcpslice, which breaks up large dumps by time, date, etc.; there is a tool that "anonomizes" (sp?) packets so that you can analyze streams without violating anyone's privacy (this is largely for academic use, but if, for example, you wanted to do some kind of traffic analysis on your uplink, you could do so without ruffling as many feathers).
Finally, note that tcpdump will sniff on pretty much any interface that supports libpcap. Tools like Etherpeek only talk to certain (ethernet) adapters, for example. Caveat emptor.
Your point is well taken and apparently, I didn't make mine clearly enough.
What I was trying to do is to explore the possible sources of and solution to the gentleman's problem. You are correct that his cat might just like eating cables, just as my dog liked eating USB headsets. So, you try correcting the animal or using sour apple or hotsauce, as you suggest. (FYI: my dog likes hot sauce, so that wouldn't work for him!), in addition to providing him with acceptible toys.
However, if these techniques don't work, then it could be a deeper issue.
What I should have said was that, chances are, it's NOT a deeper issue and that the other techniques should work.
One thing to remember -- don't blame the pet, blame the owner. Nothing against you, myself, or the poster; instead, when your pet acts badly, the question you should ask is, "what did I do wrong" or "what can I do to teach him that this is bad." Pet's don't intrinsically know right and wrong, or what is appropriate or inappropriate behavior. You have to teach them (just like children). It's not the dog's fault that he ate your $120 headset -- he didn't know that it wasn't a toy.:)
One final thought, which applies more to dogs than to cats (who are more agile and better climbers) -- consider "puppy-proofing" your house (or "cat-proofing", if you will). This is to say that you should basically crawl around on the floor and look for things that your pet would have access to which you don't want him getting at. We're talking wires, open trash cans, food products, etc. If you've ever raised a kid, you know the term "baby-proofing" or "toddler-proofing" -- this is the same idea. Remove attractive nusiances and you will be less stressed and your pet will be safer.:)
[Sorry -- formatting got fouled up on first post. Here's the full post.]
I had the same problem with my golden retriever when he was young (8mos). Among other things, he chewed through:
o The power adapter for my laptop o An AC computer power cable (luckily, it wasn't live at the time, though that might have broken him of the habit) o Several sets of $100 USB headsets
When animals chew on inappropriate things, it can mean one or more of several things. [Caveat: my experience is largely with dogs, though I expect some of the insights will apply to felines too.]
o They are teething or have a dental problem. This is normal when they are young (and new teeth are coming in). o They are bored or feel abandoned -- inappropriate chewing, marking, etc. can be a compulsive disorder for dogs, kind of like ADHD. Sometimes it is an experssion of anxiety (particularly separation anxiety) or lack of attention. There are resources available for training this out of dogs (and cats, too, I assume) o In the case of both cats and dogs, it can mean they are pissed off at you, either for leaving them alone, not playing with them, or whatever. o Finally, it may indicate that they are hungry (is your cat *eating* the wire or just chewing it to bits?). If they are eating what they are chewing up, you'll find... evidence... routed though their South end in a couple days...:)
Deterrance and correction. The best course of action is to catch them in the act and correct them, usually with a loud "DON'T!" or something similar. In the case of dogs in particular, this is a good opportunity to establish dominance -- e.g., roll the dog on his back, pin him down (primarily by the neck or mouth), and hold him there until he stops struggling. He'll understand that you're the Alpha and that he's done something unappropriate.
[Note: this is a fight that you absolutely have to win -- if he escapes, catch him; if he tries to get away, restrain him. If you don't win the engagement, the dog will conclude that the's the Alpha and behavior problems will persist. Also, I'm not advocating *hurting* or being violent with your pet. The technique I described mimics, at least for dogs, actual dominant dog behavior, but does not inflict injury or physical pain. It sends the right message with a minimum of violence and a maximum of effectiveness and is a well-accepted training technique.]
One thing that is not effective is correcting the animal after the fact -- dogs and cats have short memories. If he chewed up your cables and you yell at him even 5 minutes later, he won't understand -- most animals don't have the sophisticated reasoning system to connect the correction with the past event. You have to catch him in the act and issue the correction post haste.
One way to ensure that you can catch them in the act is to set up a "trap" -- leave a cable on the floor and keep an eye on the animal. If, or when, he goes after it, nail him then and there. If you do this a couple times, then hopefully he'll get the idea that cables are not food/playthings:)
Another technique is to use a product such as "Sour Apple", which you coat on surfaces (such as cables) that the animal tends to eat. The product is non-toxic, but tastes quite nasty to them (very bitter/sour) and will deter them. Your local pet shop should have a selection of such products. I never had to use them with my dog, but I have friends who have used it with great success.
The behavior problem may also be due to the fact that the animal does not have any or enough appropriate toys to chew on. Go to your pet shop and buy a selection of animal- and breed- (read: size) appropriate toys. If the problem is chewing, then get chew-toys. If the problem is scratching, get scratching posts or similar products.
Introduce the toys to your pet and praise or reward him when he chews on them and plays with them. This will build a positive association between playing with the appropriate toys. To
Slashdot Mirror
I was rather like you when I was in undergrad (in the late 90's). I started out as a Math major (Operations Research) which required certain CS classes. As I learned more about CS, I found that there is a very rich mathematical basis for Computer Science -- from the theories of computation to graphics to algorithm analysis. Almost any serious PhD in CS involves a heavy dose of mathematics in one form or another. Think of it as applied mathematics, in a geeky twisted way :)
:)
Part of what I'm saying is that you can do CS and not end up as a programmer, per se.
The other half of the equation is that there *are* significant (well-paying) jobs for mathematicians. Now, I doubt that you'd want to (or could) seriously pursue any of them with just a BS, but a PhD need not be a requirement. My S/O's employer has several math/statistics majors on staff who perform marketing analysis, trending, etc... some of it rather high-powered stuff. If you look into the Actuarial or Operations Research fields (if that floats your boat), there are awesome opportunities.
Whatever direction you choose, I strongly encourage you to go past a BS -- at least stay in school through an MS program. For one thing, it opens more doors down the road (I've gotten at least two jobs partially because I have an MS/CS). More importantly though, IMHO, it makes you a better professional; you learn a heck of a lot more in grad school than in undergrad -- at least that was my experience. You study your subject in far more depth and with far more rigor than in undergrad and you're treated more like a colleague than a student. It's an awesome experience, particularly if your're more of the geeky theoretical type
Whatever you do, make sure you enjoy it. Of course, you can always go back and get a second degree in underwater basket weaving or Anthropology, but it's a heck of a lot easier to get it right the first time. The sooner you identify a career path (at least vaguely), the better choices you can make in courses, internships, research focus, etc.
Good luck to you!
This is truly one of the wost ideas I've ever heard and it only highlights prevalent misconceptions with respect to firearms safety.
Modern guns, themselves, are not inherently unsafe. Guns do not spontaneously jump up and shoot people. Guns only discharge when the trigger is pulled; while this can occur unintentionally, e.g., due to a dog stepping on the trigger of a loaded, unsafed gun left laying on the ground (don't laugh, this does happen!), almost 100% of such incidents are due to NEGLIGENCE.
Nearly all incidents of unintended or illegal shootings are due to negligence, lack of training and practice, or intentional criminal activity. Negligence includes such actions as: allowing unauthorized access to a weapon by a minor, pointing a weapon at something other than a target or a safe downrange area, and placing one's finger on the trigger when the weapon is not pointed in a safe direction. Lack of training and practice leads to negligence; there are numerous incidents of police officers, who, in principle, should be some of the best-trained firearms handlers among us, who have shot themselves in the foot or leg while handling their own weapons.
At the end of the day, it is the person, the gun owner, who is responsible for safety. When a gun discharges, it is because of someone's actions; full stop. It's not the gun's fault and it's not the manufacturer's fault.
We also must remember that the purpose of most weapons -- handguns, assault rifles, tactical shotguns, etc. -- is for defensive or offensive use against other humans. Put more simply: they're designed to stop human adversaries, by injury or death. In principle, their use, particularly by civilians, should be very infrequent. I am a relatively highly-trained defensive shooter; I believe that I am capable of defending myself, my family, and my home, should the need arise. But I hope and pray that I never need to do so. I think that most police will tell you that they hope to have to shoot a suspect, but that they are trained and prepared to do so to protect others or themselves.
If and when, however, the time comes that a weapon is needed, one must be supremely confident in the reliability of the weapon. This means that simpler is necessarily better. When you pull the trigger, you want to hear "BANG", not "click" or "beep". You don't want to have to fiddle with magic decoder rings, tiny keys, batteries and secret codes, etc. in the dark, under pressure, with your child screaming in the background. And a cop can't be worrying about passwords and encrypted ammunition in the heat of a pursuit. He must know that his weapon will fire when he pulls the trigger -- he's betting his life on it.
The technology described in this article is just another way to make owning firearms more difficult and more expensive. Criminals, by definition, have no regard for laws. You can make all guns illegal and the bad guys will still find a way to arm themselves. Look at Chicago or Washington, DC for prime examples. Those of you in Austraila and Britian have seen a rise in violent crime, including home invasions and broad-daylight robberies, since you banned guns.
The technological achievement expressed in this article is impressive. I'm happy that people are exploring the uses of modern computer and cryptographic techniques. But be skeptical and wary as well... Your rights are at stake here.
Laptop theft is the least of SF's criminal problems. SF has the largest absolute population of homeless individuals in the whole country. Considering SF's relatively small footprint, this is astounding. With that many hungry, poor, desperate, and possibly substance-addicted people wandering around, crime in general will go up. It is true that laptops are a lucrative target and WiFi hotspots encourage a greater density of targets.
However, all this article says to me is that if you create an environment where vagrancy and homelessness is subsidized (both by public and private money), immigration and other laws are not enforced, and people are not permitted to defend themselves (handguns are now illegal in SF, and forget about getting a CCW in any metro. area in CA), then crime will thrive.
Laptop theft is but an unfortunate symptom of a much larger problem.
The dictionary attackers are brutal.
I get my internet service from Comcast. I don't use my Comcast email account, but my girlfriend needed an account, so I created her one under mine (you get like 5 aliases). Her first name happens to be 4 letters long and rather uncommon, so I was able to get her first choice, xxxx@comcast.net.
I created the account at maybe 10:00am. She logged in, for the very first time, at around 11:00am. By that time, having never ever used the account, she already had about 10-20 spam messages.
Bloody amazing.
I think that, next time around, I'll try a much longer user name, as several people have already suggested.
OnStar scares the hell out of me. I don't like the idea of paying Big Brother to monitor my car and its travels and travails.
:)
I don't know how accurate the CSI and Law and Order shows are with respect to the ability of the police to use such systems to track vehicles and people, but it seems totally plausable. Regardless of whether they currently do it, they (OnStar, a hacker, the feds, etc.) could, in principle, log your every move. Good if your car gets stolen or you need an alibi. Bad if you, for example, commute through 'sketchy' neighborhoods, are trying to avoid a stalker, or are being persecuted by the Department of Homeland (in)Security.
Personally, I don't have OnStar in my vehicle, and I don't want it. If it becomes "standard equipment" on new cars, I'll want a way to disable it or remove the hardware from the car without disabling the vehicle itself.
The problem with many of these technologies is they allow third parties (GM, the cops, etc.) to collect information about you while you don't have access to said information. For example, try getting a list of the phone numbers that called your cell phone (caller-id blocked or not) -- the phone company won't give it to you. I know, I tried. But if the police get a warrant, they sure as hell can get this information.
I'd rather that certain information not be collected at all. But I'd be happier to accept that collection if I had full and free access to it. I want to know what they know about me. It shouldn't take a court order to access your own phone records. And if your driving is being monitored and possibly logged (or even if this is possible), there should be full disclosure of this fact and the resulting data.
In the mean time, for those of us who value liberty, freedom, and privacy, I'd suggest finding another car company.
I was shopping for a new laptop recently, to replace my still-working but very slow IBM Thinkpad A22p (800MHz).
.
I have tried time and again to get used to touchpads, but I never feel like I have good control with them. The experience is always frustrating
So, my #1 criterion for a new laptop is a non-touchpad pointing device, a la IBM's trackpoint, i.e., the "nipple"-type pointing stick.
I looked far and wide on the net. Of the major vendors, I found that IBM, Dell, and Toshiba all still make laptops with pointing stick devices. Note that Dell and Toshiba only offer the "stick" on certain models. I have used both vendors' pointing devices and, while not as good as a TrackPoint, they're FAR better than a touch pad.
I ended up buying a new ThinkPad anyway, mostly because I have a bunch of accessories that are thinkpad-specific and also because I know that, like it or not, ThinkPads are some of the most durable, long-lasting notebook computers out there.
Does anyone out there know if it's possible to put an IBM keyboard (with the TrackPoint goodie in it) into another vendor's laptop? Is the keyboard form factor and interface standard?
All of FreeBSD, OpenBSD, and NetBSD are good distributions, IMHO, based around a solid kernel. However, the best bang for your buck (in this case, your time) is FreeBSD, I think.
FreeBSD is pretty easy to install and comes with an extensive collection of "ports" -- packages that you can download and install or build for your system. I think it supports a larger range of applications than the other BSD's which are targeted more towards server and headless situations.
Check our the FreeBSD website for lots of great documentation on installing and configuring your new system. I think you'll be fine with the hardware you're running (I've run it on very wimpy pentium and 486 boxes in the past).
If you're going to grab a distro, get either the last 4.x-STABLE build or one of 5.2 or 5.3-RELEASE. While it's fun to run the bleeding-edge last night's build stuff, it's probably not right for a BSD beginner.
You'll likely feel like a fish out of water for a few days using BSD, coming from a Linux background. Config files are in different places, the boot process is a little different, and so forth. However, I think you'll learn to appreciate its design, just as I'm sure you appreciate the way Linux works. It's a great OS for servers -- I use it for my email server at home.
Good luck!
IMHO, undergraduate CS programs really only teach you what you don't know. Unless you go to a slave-driving engineering school, where you do a ton of programming, it probably doesn't really matter where you go as long as you get a solid grounding in the fundamentals.
I would strongly recommend going to graduate school. From my own experience, the two years I spent in grad school were FAR more valuable in my development as an engineer than the years I spent in undergrad.
I'd view your undergrad education as a chance to become a well-rounded individual, a good citizen, etc. Take some history, philosophy, english, etc. classes. You're going to be staring at code for the next 25+ years, so have some fun now while you're young.
As far as getting a decent job when you get out, a degree from Stanford, MIT, CalTech, etc. may help, but a cheaper, easier, and probably better way is to get some industry experience while you're in college. Look for summer co-op/internship positions at your college's career center. I did several internships during undergrad and grad school. The more industry experience you get, the better idea you'll have as to whether you really want to be a programmer, what kind of work you want do to, and what kind of company you want to work for.
Remember that, in the long run, it doesn't matter what school you went to. 5-10 years down the line, employers will be looking at what jobs you've held, what experience and skill-sets you have, and so forth, not where you got your BS. Your intelligence, drive, ambition, and choices will govern your career, not your choice of university.
Getting an MS is also a good idea. It sets you above the rest of the candidates and, IMHO, gives you a level of professional maturity that takes years to develop in industry. You also get exposed to more advanced concepts and technologies, and you have the opportunity to start specializing and yet still maintain the option to change your mind.
The other factors that employers look at is your GPA, regardless of where you go to school. I'd rather hire a guy with a 3.9 from a state school than someone else with a 3.1 from an Ivy League-class school. We want to see smarts, consistency, and the ability to follow through on committments. The fact that you didn't take classes with some wazoo-famous professor is really irrelevant in the big picture.
Stay where you're at. Look at MS programs at top-25 CS schools. Do some internships. Work at the UNIX lab on campus. Look at undergrad research internships in your CS program. Keep your GPA up. Read, program, and do things on the side to broaden your skill-set. And, most of all, have fun!
A couple tips:
1. Bright white text on a black background. Yellow or green also works. Less glare, good contrast, etc. Great for x-terms. Harder in Word. Find what works for you.
2. Dialup your font sizes and/or zoom in. My friends used to tease me about using "fonts for the blind" on my x-terms, but I like not squinting and leaning into the screen to read stuff. Yes, you cannot fit as much stuff on the screen, but what you can see is readable.
3. Bigger is better. Get at least a 21" CRT monitor or a 19" LCD. It helps make (2) easier -- you're less tempted to use small fonts to see everything you need to see.
4. LCD's seem to experience less glare, be clearer, flicker-free, and have a myriad of other benefits. I like them, but I still appreciate a nice, big flat-screen Trinitron CRT monitor. The moral of the story is to spend a few extra bucks for bigger, clearer, and higher-quality. Money comes and goes, but you only lose your eyesight once.
5. Your eye doctor can prescribe "computer glasses". For the older folks among us, this can be a huge help. The problem is that most computer users sit about 3' from the screen. This is too close for your distance/driving glasses but too far away for your reading glasses. Finding a mid-range prescription can make a huge difference and reduce eye-strain.
6. Don't work in the dark. Don't work under really bright lights. I find the best computer lighting to be just a little too dark to comfortable read by.
7. Follow all those ergonomics suggestions that your HR department hands out. For example, get up and walk around every hour or two. This gives you a chance to stretch, focus your eyes at a distance, etc. Drink more water -- it's good for you and you'll have to pee more, which makes you get up more. You'll actually be more productive, despite the breaks.
It almost seems like this should be a national policy rather than state-by-state. I guess that gets into states' rights issuesm, a topic on which one could expound for volumes.
I think the whole problem is compounded by the two-party system. I'm at the point where I think it would be wise to disband political parties all together. Candidates should run on their own platforms and not be forced to shoehorn themselves into a particular party's view of the world. Coalitions (say, in Congress) should be formed around individual issues, not along party lines. This might make the system slower, but it also breeds consensus and allows representatives to actually *represent* their constituencies (sp?).
Your point is well-taken. The system is the system and, according to the rules of that system, Bush was the winner.
Of course, this doesn't address the whole Florida isssue, which is another can of worms entirely. Without descending into an extended argument about hanging chads, the issues experienced in Florida underscore the way that small errors in tallying votes can really affect elections. It's my understanding that *any* voting method, be it punch cards, touch screens, color-in-the-box, or whatever, has some small, but non-zero, margin for error. In most cases, the outcome is not jeapordized by this errorbar, but, in some cases, it is. I am doubtful that introducing computers will solve the problem, but it might have a chance of reducing the errorbar.
Hey, all Nader had to do was get enough petitions signed in each state. He didn't. Regardless of what the Democrats think, say, or do, if enough people were in favor of Nader, he'd be on the ballot in 50 states, hands down.
It's easy to blame things on parties or liberals or conservatives, but sometimes a rose is just a rose. Nader didn't get enough signatures. He loses. That's it.
The reason that the states you mention are "battleground" states is that:
a) They carry a non-trivial number of electoral votes and
b) They are potentially "swing" states, i.e., both candidates have a chance of winning them
You don't hear much about Kerry campaigning in, say, Georgia, even though it carries 15 electoral votes, more than twice that of, say Idaho (7 votes). That's because Georgia is going to vote for Bush.
Now, if the election were decided on percentage of the total vote or, say, by obtaining a majority (>50%) of the vote, then it might be worth it for Kerry to campaign in urban Georgia areas, such as Atlanta, where he might be able to pick up thousands of votes.
You also have all the little or low-population states who, perhaps, get more than their fair share of electoral votes. The minimum number of electoral votes allocated to a state is 3. So, states such as Alaska, Montana, Wyoming, DC, North and South Dakota, and Vermont each gets three votes, regardless of population. Is that fair? In a sense, each voter in such small states has control over a larger fraction of an electoral vote than does a voter in a large state, such as California.
But that's how the Founding Fathers designed the system and, at least for the 2004 election, that's what we're stuck with, like it or not. Personally, I'm undecided as to whether, if I had the power, I would move away from the Electoral College system or stick with it. It clearly has its advantages and disadvantages.
Fundamentally, however, I think any system in which one candidate can get 500,000 more votes than his opponent and still lose is somehow, fundamentally, messed up.
Another salient point comes from my own experience as a voter in North Carolina, where I lived during the 2000 election. At that time, there was no way in hell I was going to vote for Bush. However, he was polling at over 60% in NC, so, in essence, my vote was irrelavent. I could vote for Gore, but it wouldn't matter, because the majority of the state would have, metaphorically, drowned me out. How I ended up voting is not germane, but I think you can see the quandry one might be in.
I think this situation may lead some folks to apathy, i.e., not caring about or getting involved in the system, perhaps to the extent of not voting at all. If you live in a state where you are in the minority in re your political opinions, then your vote really doesn't count for presidential elections. That's just how it is.
Now, if we moved away from an Electoral College system, then all those "minority voters" (as defined in the previous paragraph) WOULD have their voices heard, since their votes would "count" towards the candidate's total. Somehow, that seems more democratic to me.
According to MSNBC,
"Cisco Systems Inc., which acknowledged its popular routers were among those vulnerable, distributed software repairs and tips to otherwise protect large corporate customers."
[http://www.msnbc.msn.com/id/4788445/]
Cisco is pretty good about releasing fixes for exploits in their software as fast as possible.
Being big has it's plusses and minuses. On the down side, they're routing packets for 80%+ of the internet. On the plus side, they have an army of engineers and a fire lit under their asses to get these kinds of things fixed.
Windows XP (and ME, I think) have a feature called "System Restore". Basically, what it does is track changes to the registry, driver database, and other parts of the system. It takes snapshots of the system periodically and sometimes during the installation of hardware or certain drivers.
:)
If you break something (as I have been known to do from time to time), you can "roll back" to a previous snapshot. In my experience, this works pretty well for solving certain problems.
I'm not sure if it tracks installed software or whether such software would automagically be removed by rolling back. But the snapshots must have some such information in them, if only we could extract it...
Why not? Because laptops aren't designed to be servers. They generally underperform similarly-classed full-size systems, are generally not designed to be always-on, have flakey hardware, etc., etc.
:)
Furthermore, laptop harddrives are much slower than server-class storage and, in my experience, have a much lower life expectancy.
Finally, laptops dump a lot of heat out their bottoms and don't like to be stacked on atop another. If you try this, you're liable to end up with a fire or, at the very least, hardware failures, in short order.
I think "use the right tool for the job" applies here...
>>Go with 4U rack mount cases. They are cheap, and easy. They will fit all your standard hardware out there.
This is a good point. Dell, in fact, sells "workstations" and "servers" that can be configured either as tower or 4U rack-mounts. You basically bolt on a couple rack ears and you're good to go.
If you find the right size case, you can buy rack ears, drill a couple holes in the case (be careful about the metal bits!) and bolt them on. Rack-mount servers on a budget.
If you're not much of home machinist, you could consider rack shelves -- I highly recommend that you find shelves that bolt on to the front AND back for stability. You can stack your tower/desktop systems on their sides. This gets a little messy if you ever need to move the rack, and it takes up a hideous amount of space compared to 2U systems (a shelf + tower consumes basically 5-6U of space), but it's an alternative or a way to rack-adapt existing hardware.
Rack mounting your toys is great, IMHO. I don't do it (yet) at my home because I don't have very much stuff, but if I did, I would.
:) ).
There are a number of companies making "shorty" racks (that's my term, don't Google for "shorty"). They are normal 19" racks, but sit only 30 or 36 inches high. One I seriously looked at for work had a door on the front, too.
If you just want to mount a couple 1-2U servers, a switch/router, etc., this would be a sweet way to go. You can tuck it undera desk or in a (ventilated) closet and forget about it.
Rack equipment (particularly computer) is more expensive, apples to apples, compared to non-rack hardware, but it's compact, often better-cooled (read: noisy), and probably bare-bones enough for your requirements. Many of the rack computers I've worked with also have full BIOS support for a serial/TTY console, so you can remotely access the computer, reboot it, etc. You can gang all the TTY's into a terminal server (Cisco makes a couple models) or use serial-to-USB adapters and run them off of one box (just make sure it stays up
Make sure you calculate how much rack space you need before buying a rack. Also, from my experience, you'll probably want to allow at least 20-40% extra for expansion, testing, etc. Also think about cooling and the cooling requirements of your hardware.
Another benefit of a small rack is that if you move or relocate, you just unplug the rack from the wall and move the whole bloody thing with you. No recabling your equipment or lost toys. And, if you put a locking door on the unit and bolt it to the floor, it's pretty secure too.
If this all pans out, let us all know and maybe post some pictures. I think this is a slick idea that more individuals and small businesses should seriously consider.
VNC will allow you to access your windows desktop from any routable computer that runs a VNC client. VNC runs on almost every OS, including Palm, I think.
You won't be able to see remote filesystems unless you can mount them on the windows machine -- for UNIX, you could use Samba... for other OS's, well, I don't know.
I actually do the opposite from what you want to do -- I run a headless FreeBSD box and then view the VNC (X) desktop remotely on a Windows machine (full-screen, no less). Performance is about the same as running locally and I can still use all my Office products (necessary for my job, sadly).
Best of all, VNC is free and open-source (does that make us for me using Windows?) >:-)
>> Hubs are pure, unadulterated evil.
:)
I disagree. They're great for sniffing packets. If you've got an ethernet-connected device that doesn't have a sniffer onboard, and you want to see what the heck it's doing, a hub is a handy tool to have on your shelf. I use them quite often to intercept traffic while debugging software and hardware at work.
They also allow you to run a trace on a separate machine, so as not to interfere with the unit-under-test.
Yes, you could use a monitor-port, but that assumes that your employer will spring for fancy managed switches AND will allow you log into them and mess around. Mine doesn't, but they're more than happy to buy me a $40 hub and a couple patch cables
I'm not a net. admin either, I'm a software engineer.
When you're writing network software, or software that uses the network, you often run into wierd and hard-to-debug problems. The task of finding the cause of these bugs if often expedited by looking at the packets on the wire.
For example, you think you're sending a particular pattern of bits (1's and 0's) -- that's what you think you coded in your program. But for some reason, the other end doesn't understand your packets. You could put a bunch of debugging statements in your program, recompile, and hope you can see the problem, or, you can simply sniff the packets and see what's really going out on the wire.
As another person mentioned, sniffing is also useful for reverse-engineering closed-source software that uses the network. That's how those guys implimented clients for AIM -- they just figured out what messages to send back and forth. (Sadly, AOHell decided to change the protocol every 2 minutes so the open-source clients don't work very well.)
As far as security, sniffing or analyzing traffic is one of best tools available to see what's passing through your network. It's analagous to the security cameras in the local stop-and-rob (gas station) or in a casino -- they let the security guys watch what's going on, review it after the fact, and find/identify the bad guys.
You know, Sun really pissed me off in this particular area. Snoop is basically tcpdump with all the options and syntax changed. The output is a little prettier, but it's essentially the same. Switching between the two used to give me no end of headaches. Why reinvent the darned wheel?
:) )
(Note: I'm not griping at you, and I love Solaris... Sun just does some things that really annoy me
I guess I'm oldschool, but I still use tcpdump for most day-to-day things. It's handy, it's fast, and it runs on just about every OS (including Windows (google for windump)). The output is ugly, but once you get used to it, you hardly notice.
:) ), but couldn't/didn't want to install all kinds of GUI tools, etc. This is where tcpdump really shines. You can capture to a binary file and read the file with tcpdump, ethereal, Etherpeek, and many other packages. As long as you can get the file off the machine, you can analyze the data.
:)
When I really need to analyze a stream or set of streams, or I'm going to be staring at packets for more than about 10 minutes, I switch to ethereal. Again, it's free, runs on most OS's (including Windows, again), and the GUI is a little clunky, but quite usable. As several people have mentioned, the capture filter syntax is identical to tcpdump. The display filter syntax is different and I find is a little tricky to get right, so I try to prefilter (or filter with tcpdump beforehand) as much as possible.
One handy feature is the ability to analyze certain types of streams, such as a TCP session (filter out the whole session and see all the data in one window) and SIP (analyze jitter, loss, extract audio session, etc.). It's also open-source, so if it doesn't understand some kind of traffic, you can write your own extension. I haven't had to do this yet, but I know people who have, and it seems easy enough for a compitent programmer.
My employer has a site license for WildPackets Etherpeek (it comes in several versions... I think we have one of the higher-end ones). Frankly, it's prettier than ethereal, but, at least for the debugging I do, provides very little extra functionality. The capture filters are embedded in a GUI which I find makes it hard to see how they're configured.
Etherpeek is pretty and may be easier for novices to use. But I wouldn't waste the money unless it has some quirky feature you just can't live without.
Something to keep in mind: often, the place where you capture packets is not where you'd like to analyze them. For example, I've had situations where I needed to sniff traffic on a remote server -- I had ssh access to the server (and root, of course
There are also handy tools for managing and analyzing tcpdump files, such as tcpslice, which breaks up large dumps by time, date, etc.; there is a tool that "anonomizes" (sp?) packets so that you can analyze streams without violating anyone's privacy (this is largely for academic use, but if, for example, you wanted to do some kind of traffic analysis on your uplink, you could do so without ruffling as many feathers).
Finally, note that tcpdump will sniff on pretty much any interface that supports libpcap. Tools like Etherpeek only talk to certain (ethernet) adapters, for example. Caveat emptor.
Bottom line: pick the right tool for the job
Your point is well taken and apparently, I didn't make mine clearly enough.
:)
:)
What I was trying to do is to explore the possible sources of and solution to the gentleman's problem. You are correct that his cat might just like eating cables, just as my dog liked eating USB headsets. So, you try correcting the animal or using sour apple or hotsauce, as you suggest. (FYI: my dog likes hot sauce, so that wouldn't work for him!), in addition to providing him with acceptible toys.
However, if these techniques don't work, then it could be a deeper issue.
What I should have said was that, chances are, it's NOT a deeper issue and that the other techniques should work.
One thing to remember -- don't blame the pet, blame the owner. Nothing against you, myself, or the poster; instead, when your pet acts badly, the question you should ask is, "what did I do wrong" or "what can I do to teach him that this is bad." Pet's don't intrinsically know right and wrong, or what is appropriate or inappropriate behavior. You have to teach them (just like children). It's not the dog's fault that he ate your $120 headset -- he didn't know that it wasn't a toy.
One final thought, which applies more to dogs than to cats (who are more agile and better climbers) -- consider "puppy-proofing" your house (or "cat-proofing", if you will). This is to say that you should basically crawl around on the floor and look for things that your pet would have access to which you don't want him getting at. We're talking wires, open trash cans, food products, etc. If you've ever raised a kid, you know the term "baby-proofing" or "toddler-proofing" -- this is the same idea. Remove attractive nusiances and you will be less stressed and your pet will be safer.
[Sorry -- formatting got fouled up on first post. Here's the full post.]
:)
:)
I had the same problem with my golden retriever when he was young (8mos). Among other things, he chewed through:
o The power adapter for my laptop
o An AC computer power cable (luckily, it wasn't live at the time, though that might have broken him of the habit)
o Several sets of $100 USB headsets
When animals chew on inappropriate things, it can mean one or more of several things. [Caveat: my experience is largely with dogs, though I expect some of the insights will apply to felines too.]
o They are teething or have a dental problem. This is normal when they are young (and new teeth are coming in).
o They are bored or feel abandoned -- inappropriate chewing, marking, etc. can be a compulsive disorder for dogs, kind of like ADHD. Sometimes it is an experssion of anxiety (particularly separation anxiety) or lack of attention. There are resources available for training this out of dogs (and cats, too, I assume)
o In the case of both cats and dogs, it can mean they are pissed off at you, either for leaving them alone, not playing with them, or whatever.
o Finally, it may indicate that they are hungry (is your cat *eating* the wire or just chewing it to bits?). If they are eating what they are chewing up, you'll find... evidence... routed though their South end in a couple days...
Deterrance and correction. The best course of action is to catch them in the act and correct them, usually with a loud "DON'T!" or something similar. In the case of dogs in particular, this is a good opportunity to establish dominance -- e.g., roll the dog on his back, pin him down (primarily by the neck or mouth), and hold him there until he stops struggling. He'll understand that you're the Alpha and that he's done something unappropriate.
[Note: this is a fight that you absolutely have to win -- if he escapes, catch him; if he tries to get away, restrain him. If you don't win the engagement, the dog will conclude that the's the Alpha and behavior problems will persist. Also, I'm not advocating *hurting* or being violent with your pet. The technique I described mimics, at least for dogs, actual dominant dog behavior, but does not inflict injury or physical pain. It sends the right message with a minimum of violence and a maximum of effectiveness and is a well-accepted training technique.]
One thing that is not effective is correcting the animal after the fact -- dogs and cats have short memories. If he chewed up your cables and you yell at him even 5 minutes later, he won't understand -- most animals don't have the sophisticated reasoning system to connect the correction with the past event. You have to catch him in the act and issue the correction post haste.
One way to ensure that you can catch them in the act is to set up a "trap" -- leave a cable on the floor and keep an eye on the animal. If, or when, he goes after it, nail him then and there. If you do this a couple times, then hopefully he'll get the idea that cables are not food/playthings
Another technique is to use a product such as "Sour Apple", which you coat on surfaces (such as cables) that the animal tends to eat. The product is non-toxic, but tastes quite nasty to them (very bitter/sour) and will deter them. Your local pet shop should have a selection of such products. I never had to use them with my dog, but I have friends who have used it with great success.
The behavior problem may also be due to the fact that the animal does not have any or enough appropriate toys to chew on. Go to your pet shop and buy a selection of animal- and breed- (read: size) appropriate toys. If the problem is chewing, then get chew-toys. If the problem is scratching, get scratching posts or similar products.
Introduce the toys to your pet and praise or reward him when he chews on them and plays with them. This will build a positive association between playing with the appropriate toys. To