I'll be frank up front too. I'm an Independant, with a capital (I). I believe in clean drinking water, not totally fucking the eco system, keeping the economy rolling, keeping my gun rights, moving away from fossil fuels, nuclear energy, electric cars, the space program, an adequate military budget, and all the other stuff from both sides of the fence that actually make sense.
Now that we've gotten that out of the way, explain to me how Fox News is actually sided in the political schema. Seriously now, I'm not being sarcastic. I'd really like to know why you believe that channel has a Republican agenda.
Let's look at their primetime lineup.... Brit Hume - The serious looking old guy. Covers news he reads off the cue-o-matic machine, doesn't add any of his own comments.
Shepard Smith - Appeals to the ladies, the guys think he's funny. Covers new (cue-o-matic! (tm)), adds his own comments, but they're always smart ass remarks trying to be funny.
Bill O'Reilly - The big O. This is the funny one. All the Dems swear he's pro Rep, and the Reps swear he's pro Dem. Independants (like me) like him because he walks both sides of the fence by making logical decisions, rather than simply believing hype. If someone proves him wrong, he admits his fault, and re-evaluates his position on a subject. I've even seen him change his opinion on the air (rarely happens as he is quite well informed, but it does occur from time to time). He's said plenty bad about Bush, and plenty good about him too. Same goes for Kerry. Good show, people should watch it, even if they don't like it. You'll change your opinion of the show after a month or two, I guarantee it.
Hannity & Colmes - Hannity is openly a Republican, Colmes is openly a Democrat. They share a show so the world can watch them argue over issues that they feel are important. Sounds like fairness and balance to me.
Greta Van Susteren - Supposedly one of the best trial reporters around. Why you'd have a reporter specifically for covering trials, I don't know. Do I care what the atmosphere of the Kobee Bryant courtroom was like today? Not really. In fact, I don't even care what the outcome is. Someone else's trial in some other state, does not directly mean anything to my life, so why should it be my business. She doesn't seem to go out of her way to get into political issues.
So let's see now. We have a bunch of news shows that cover (imagine this) the news. None of them really going out of their way to bring politics into the mix as far as I've ever seen. And two shows that actually specifically cover political issues. One is Bill O, and the other is H & C. The Democrats do seem to be saying that Fox News is pro Republican lately. Personally, I chalk that up to the upcoming election. I have noticed a liberal slant to CNN, and I'm sure they'd much rather people watch that instead.
Yea, I'd have to agree with you on that one. I've been typing for over 15 years now myself. I was the master at the 4 finger typing skill on my old 286 long before I took typing when I was in highschool. I, of course, played ball to the bullshit typing methods that they tried to teach me, but very quickly made up my own modified style that allows me to type all day every day with no problems at all.
Hands perpendicular to the keyboard... Elbows off the table... Don't rest your palms on the table...
Yea, and make sure to sign up for medical treatment in advance...
My hands are almost always at a 30-40 degree angle away from being perpendicular to the keyboard, and I lazily rest both my palms and arms, and elbows on the table (I have a pretty swift desk I built for myself that emulates having two desks make a corner. It helps a lot with the whole resting of elbows;) So the keys aren't in the "magic hand-key relative location (tm)". Bah. I still type over 50wpm and have well less than 3% errors. In my opinion, typing is a skill that you need to have, but everyone is comfortable in different positions. The most difficult part of typing is learning where the keys are, so simply teaching computer skills to people and peaking an interest in them will be more than sufficient motivation for them to learn the keys. Which of course will allow them to make up their own bullshit typing skill that works well for them (like apparently we both did), and everyone is happy.
Dust & dirt. I would imagine that at such low voltage levels, induced current would require a damn near perfect level of alignment between the chip and the "socket". This is admitted in the article. What they don't admit is that it's going to be nearly impossible to get the damn thing in the socket without letting dust or dirt inbetween the chip and the socket.
And a more interesting topic is their consistent mentioning of taking the cache of the chip. That's a nice dream and all, but where the hell are you going to put it then? Hardwired onto the motherboard? That's going to dramatically increase the cost of mobo's (so they are simply shifting who gets to eat the high sticker price on their products). And what if I buy a quad capable mobo, but only put 2 processors on it, I'm effectively wasting 2 sets of cache, rather than simply wasting 2 cpu sockets, and the sockets are a hell of a lot cheaper than the cache. I suppose you could fix this by going back to COAST (cache on a stick, yeah i know you remember that nasty stuff). But that brings in a whole new problem: These days, cache is only fast because it's so close to the cpu. If they move it off the die, it's just going to be put back on in 2 years because we can't access the cache fast enough ever since we moved it off the die.
I'm no super computer engineer, but these guys better have an entire family of rabbits they plan on pulling out of their asses or this fucker's gonna flop.
You seem to miss the fact that in my solution you're already logging traffic destined to port 25. When a user's traffic reached a threshold, it's flagged. There's a reason I didn't require the use of a proxy in my scenario. Mainly because a single proxy would die in a matter of seconds as the thousands of emails send by a large customer base slammed into it, and customers would start bitching because they can't send email. With a packet sniffer, you have unobtrusive access to the same information you're going to get from a proxy, without the detrimental results if packets are lost.
Whoah there buddy. This is a good idea, but you have to check to make sure you aren't about to use a hammer to fix this problem. I can think of one cable modem provider (*cough* Charter) that uses this approach, but does it quite poorly. They firewall their customers on port 25 both inbound and outbound. The only acceptable traffic on that port is from the outside world to their mailserver, and from their internal network to their mailserver. This causes major problems when their customers are trying to use another provider's mail servers to send mail.
A more elegant solution requires a few steps. Firewall port 25 inbound to their customers (so they can't run their own mailservers). 99% of end users do NOT know whey they are doing with MTA's, and you will end up with lots of open relays on your network. Next you need to somehow log your user's smtp traffic. The best way is to put a packet sniffer on the network, and then have it tally up traffic and total sent emails by source IP address. This can be used to find out who is sending the most individual emails. Now you set up the packet sniffer to perform more detailed sniffing on that specific user's outbound email. Particularly, find out how many mails they send that are only 10% different, and how many times they send email to the same set of 5,000 addresses in a month. If you determine it's spam, tell them to stop. If they don't, cut their service.
This really isn't rocket science. They just need to be very careful how they resolve this problem. People don't like suffering because a few bad eggs amongh them were causing problems.
You seem to fail to see that a lot of our cool gadgets and high technology items have their roots in military research. Radar guns for instance, velcro, etc.
Plus military research has a habit of taking previous ideas and prototypes, and turning them into reliable machinery, like large aircraft, submarines (think exploratory subs), and even cars and trucks.
If you want someone to blame for the loss of scientific dominance, then blame ourselves. We allow government institutions like the patent office to continue stiffling innovation with meaningless patents (like software patents). We allow our representatives to draft and pass crazy laws like the DMCA that prevent reverse engineering so that our aspiring engineers cannot learn from the works of those that preceeded them.
Don't blame the budget, don't blame the government, blame us. We are the ones that allowed it to happen.
Re:Rugrats..
on
Robosaurus
·
· Score: 2, Informative
What information are you basing this on? If MS bought a license, and they are less clear than Sun, what makes you think that buying a license has anything at all to do with being clear? From the sounds of it, for all you know, HP could have been the one that shelled out for the license, and Sun is still in talks.
"4 big vendors" is not very specific. Try not to spread unfounded information. It doesn't help the situation at all. I honestly don't see why Sun would care anyways. They are primarily a hardware and systems support vendor. Their OS is just there to give their platform something to run. Solaris for x86 only exists to provide their customers an easy upgrade path to SPARC-based systems.
At any rate, I'd really like to see where you get your info, or if you are just speculating.
I must admit, I've never been very good at setting up printers in Unix. If I don't have access to the Redhat printconf gui utility, I'm pretty lost. This is bad considering that I'm a unix admin at heart. I guess I've never really had the need to configure many printers on Unix boxes, and when I do, it is always conveniently enough a RedHat box.
I might just have to pick this book up. Anyone have any other suggestions on how to demystify printing in Unix? I understand how to use the lpr command, and how to kill jobs with lprm and list them with lpstat, but I'm pretty much a noob at configuring printers. A complete guide on how it all works would be nice. I'm pretty sketchy on the whole "filters" idea, and wouldn't know where to start to set up CUPS or LPD if all I had was a command line available.
I usually use rsync for one way backups, and unison where I need 2 way synchronization. Rsync is nice because you can update lots of files very quickly, as it only moves binary diff's between files. Also, if it is a costly network link, you have the option to specify max transfer rates, so you don't kill your pipe when it runs from your cron job. Unison is nice because it is pretty smart about determining which files should be moved, and can correctly handle new and deleted files on either end of the link. Plus it supports doing all of it's comm via ssh, so it's secure.
The downside to both of these being that neither of them are instantaneous. However, I've had much success running both of these as often as every 5 minutes. Just make sure that you launch them from a script that is smart enough to check for already running instances before it starts trying to move data.
Technically, FPS is likely to be more limited by your GPU than the processor. In these benchmarks, the CPU is going to spend more time compressing textures, shifting data around, and trying to push data down the AGP bus than anything else. I'll be more impressed when they start using software rendering when they do these benchmarks.
Realistically, almost all of these game based benchmarks are crap. If you want real statistics, have it render CGI movie frames. THEN see how it compares to other CPUs. Until you do that, you aren't really looking at anything other than the CPU's ability to feed the graphics card.
Apparently your magic-8-ball has never had the pleasure to use it personally.
My magic-8-ball has been formerly introduced to the product. When I asked it's opinion, I recieved "Outlook not good".
Try exposing your 8-ball, and rechecking it's results. If the message does not change, you may have a defective 8-ball, and should call the manufacturer for a replacement.
actually, the parent you replied to is correct. Neither scp or rcp are "defying" what he is saying. They don't copy files with the execute bit turned on. Rather, they copy the file, and its permissions. Same goes for cp and mv. They all try to maintain permissions as best as possible. That isn't a default of the open() posix function, it is an option that the application in question (scp, rcp, cp, mv, etc) all choose to use.
So like he said, unless the email client sets the execute bit for you (which would be really dumb), attachments won't be executable until you make them so manually.
Please share your method of remotely administering Windows boxes. I'd really like to know how to do this, especially with Win2k. Seriously! This is only a mild poke at your comment! If there is some magic way to do what you are saying with Win2k that I don't know of, I'd REALLY like to know, because it would save me hundreds of hours a year trucking my ass from PC to PC.
I'm not familiar with many (or any) of Larry's speaches, but realistically I can't think of anything better than replacing PC with dummy/smart terminals. Let's face it, network administration was MUCH easier when you could simply replace the junk terminal with a new one, rather than having to reinstall an OS every 2 weeks. (Which is basically what it amounts to when you have about 100 PCs to support. Some PC, somewhere, goes junk, and they are too expensive to simply replace with new ones when they die.) Not to mention the single point of administration. I much prefer SUN's methodology. The network is the computer. You have a terminal/workstation, and lots of services all split up all over. But the key is, your workstation doesn't have to really provide much, or any of it. All it needs to do is give you a method of input and output. Mouse/keyboard, and screen. Beyond that, you have workgroup printers on the network, fileservers in the server room, and the admin can install any new software (accessible to all machines that need it) from one location and in one fell swoop.
Much much better than trucking my ass to each PC so that I can install the new app when there are 100 PCs involved. Even better when the number of machines in question reaches the thousands.
Anyways, the software industry isn't dead. There will always be a need for new software. Business models are different between companies. That's how they compete. They excel in different areas, and to do that they need different software. Software that more closely meets their needs. Saying software is dead is akin to Steve Balmer saying that opensource and free software don't innovate, but Microsoft innovates all the time. MS buys what they think is cool, and reshrink wrap it with a new label. Free software is honestly the ONLY place where innovation occurs. Someone has an idea, and they run with it. The idea may not be polished, the software may not even be implemented that well, but it isn't the software that is being questioned, it is the innovation. The software becomes polished when some company buys/steals the idea, or when a new company is formed specifically to flesh the idea out.
If you people actually believe that all the software ever necessary is already written, then please do get out of this industry. Go write a book or something, please. Let the rest of us innovate in peace.:)
(BTW: I'm not trying to attack the parent thread, I'm just spewing ideas.:)
One quick question: How does giving the first 100 or so free prevent you from paying for email worms that continuously spam everyone in your address book? I have about 300 contacts in my address book at work. But even if I only had 5, by your idea, after 20 cycles of spams (which I would likely never see happening) I'd still start to pay for it.
This won't work for two reasons. Open mail relays and forged message headers.
If you can't track the source, you can't bill them. So then who do you bill? The company with the open relay? Some would say that's a good way to promote good system administration, but remember that the bill imposed could easily put a company out of business and into bankruptcy. Sounds a little strong to me.
I still feel that we are better off not having a mandatory tax. Instead, set up third party message verification systems. Emailers can, for a fee, have their message ran through an intense one way hashing/encryption system to create a special "Registed Email" message header, which is then sent along with the original message to the intended recipient(s). Using this system is entirely optional, but read on for the benefits of using it at least once per recipient.
Upon reciept, the recieving email client will see the special header, check it's validity with the issuer, and place it in the user's inbox. If the message does not have the 'registed email' header, then the sender's name is checked against a list of known users. If the user is known (from having been manually entered or already recieved 1 registed email in the past, and not in the blocked senders list, the mail will again go into the users's inbox. All other mail is automatically placed in a folder of the user's choice. If that means the trash, fine.
There you go. Don't need to even care about open mail relays, because if you've never heard of them before, and they don't send registered email, you'll never see their penis enlarging message. I've thrown this idea out before, but I thought I'd see if I could get more feedback on it.
Suppose now that the clients update the same record on the two servers in an incompatible way... you could imagine what will happen when the servers become visible to each other again...
This is why you have transaction logs that are timestamped. When the sytems resync, they merge their transaction logs, rollback to the last synced state, and then re-execute every transaction until they are current. The end result is that the newer row updates will overwrite the older row updates. This may or may not be the desited effect, so you must allow this operation to be turned on or off. What most systems do is forcefully synchronize all write operations amongst master servers. By this I mean that if a master server becomes unreachable, the update query is blocked until either it becomes available again or times out with an error.
Most problems with replication, though, have more to do with poor software design than the database. In order to make the multiple-masters system work correctly in an uncertain environment (pretty much every case), you should really have two tables. One local, and one global. The local table is non-replicated and queues operations to be posted to the global (replicated) table, and a background process runs on the local server to try to push them to the global table as connectivity permits. This allows you to have pending operations, and also allows you to also have a conflict table to store conflicting updates, which can then be later resolved by flesh-n-blood v1.0. Also, SQL queries that insert into one table and then update or insert another using the last insert id from the previous are often written incorrectly. You should really never call the db's function for last insert as part of a separate select, and then use the resulting hard number in the next table's insert/update. Instead, you should incorporate the function call into the 2nd table's query, so that if the insert occurs on an unsynched system, the result of replaying the transaction log will still result in an accurate id number (thus preserving the row relationships). The id may have been changed, but at least it will still have the relation intact. Obviously, all of this needs to occur within a single transaction to make it work. Coincidentally, because of this, the id column makes for a poor unique row id if it is an integer or some magically created string by the client software. You're much better off using GUIDs (or UIDs in linux), as they are based off of more unique algorithm (usually, high res date/time and NIC MAC & some random number).
On a related side note, 'master server' is kind of a bad label to use, since true RDBMS's have better granulatiry than that. Usually you can set master/slave properties at the table level, and still most even let you set master/slave at the row level (as defined by a where clause), as well as published data at the column level (as defined by a non-joining select statement).
In all honesty, I do agree with your argument about replication not being easy, but I take the stance that the truly hard (and critical) part is writing software that correctly handles being in a replicating environment. It's easy to write software that uses a database backend. but it can be a bitch to make it run correctly in all cases in a replicating environment. There is only so much that DB replicating can do for you, the rest is up to the software to keep the data sane during disconnected states.
Isn't clustering supposed to be a function of the database system, not the software you use to access it?
I mean, this is neat and all, but I really don't want to have to use this interface just so that I can cluster my database. You're much better off placing clustering functions within the database itself. Then you can access the data by any method (ODBC, native libraries, hell even with the provided command line interface).
Take a look at how MS SQL Server performs clustering sometime. Everything (and I mean EVERYTHING) is performed via triggers and tsql. All the clustering setup does is set up a bunch of known working trigger scripts to propagate the data. You can even edit them to your liking afterwards if you wish. Now I'm not saying that MS's solution for clustering is the cat's ass. Personally, I think it is kind of hackish, but then again I believe that clustering should be something you simply turn on, and shouldn't be able to fuss with. Realistically, I can't think of any good reason to change the cookie cutter tsql scripts that perform the clustering, so I only see the ability to modify them as a potential way to fsck it up (that being an obviously bad thing).
Clustering really isn't that hard to implement. I'm pretty surprised that MySQL and Postgres don't have better support for it. Especially Postgres, since transaction support is really the one big key that makes clustering possible. Maybe no one has really had an itch to make it heppen yet. Hopefully it will happen soon, since I'd love clustering to be another argument for why OSS databases can play with the big kids just as easily.
You must also consider the implications that using C# has on your scripting language choices. I can't think of many big games that don't have a scripting language of some sort. Usually a company will write or license an engine, and then extend it into the game they are trying to create. This involves scripting on some level. A lot of games currently use Lua, I've seen some that use Python, and some that use their own custom C-like language. The down side with C# is that I'm not aware of any C# bindings for the popular scripting languages. Obviously a custom script engine could be made to have C# bindings (or could even be written in C#), but most games don't use custom script languages, and will instead use an existing language simply because it lessens their development and bug finding time.
Correction: American corporations have too much power.
The problem with the US is not that they have too much power, it is that the government is corrupted by private interests. Like you stated, Rumsfeld used to head a pharmaceutical company, and probably still recieves kickbacks. Thus his views are bias. If you remove the corporate interests from the equation, the US really aren't that bad. But yes, in its current form, the US is a pain in the ass.
And surprise surprise, I'm an American. And yes, I actually believe that our government is overrun by politicians that are too easily swayed by pocket stuffers, and are more interested in getting re-elected than doing the right thing. Unfortunately, short of a complete revolt, there isn't much that we the people can really do about it. It is almost impossible to elect anyone who isn't already corrupt or will become correupted, and those few elected officials that are not corrupt are powerless to change the current system, as they are severely outnumbered by their "peers".
Personally, if shit gets too wierd over here, I'll just move. I like some of the laws and freedoms I have over here and I'm used to the language, but realistically I'm still relatively young, and could adapt to a new country with a fair amount of success. But, I hope it doesn't come to that.
I can name at least 20 non-mainstream bands that I listen to. Some have recording contracts, some just sell CDs off their website. So how do you determine which of these bands gets any of the compulsory license fees? All of them, since they all sell music for profit? Only the ones with record deals? None of them since they are not mainstream? What is the criteria for getting paid? It seems to me that compulsory licensing would never work, since you really can't even decide on who to pay.
actually, i have a reason for you. Prism2 support. I have a DWL-650 (2nd gen), based on the prism2. I had used wlan-ng in redhat 7.2(or was it 3?), but gave up on wlan-ng as it was pretty lame on the configuration side, and too much of a bitch to implement. (Kernel recompiles necessary, ripping out all the original pcmcia support and replacing it, etc) However, RedHat 9 supports my dwl-650 right out of the box... er... bitorrent acquired iso burned to cheap cdr media...;)
So there you go. Good reason right there. I used to have to run XP to use my wireless card, now I don't. Yay for me.
Oh, and gnome 2.2 is actually cleaner than 2.0. Expecially the fact that meta themes are now officially implemented, and the new menu system isn't as freaky as that funk ass "Extras" submenu.
As usual, your mileage may vary, but all in all I'm quite happy with RH9. If I wasn't dirt ass broke, I'd probably go out and buy it just to have real media.
I don't think that there are too many free software projects out there. Have you ever gone to download.com and just looked around at all of the thousands of Windows applications? Many many of them perform the same function. Some better than others, some are innovative, some are not. That doesn't seem to stop people from downloading and using them.
Quantity isn't the problem. The problem is quality. Well, percieved quality anyways. Unix has a different paradigm when it comes to software installation. That's a fact. There is no 'Program Files' folder that everything is installed to. Of course, there's always/opt!:) Not to mention the lack of automated installers for most projects. The installers exist, people just don't use them. I'll admit, some of them are a little lacking (a scriptable installer ala InstallShield could be helpful), but I have plenty of Linux games that use those simple installers, and they work great, despite kernel upgrades and distribution changes.
IMHO, people just rely on./configure too much these days. I've always felt that build scripts like./configure were useful for the developers and hackers, but the general public really shouldn't have to have gcc installed just to install new software.
It also doesn't help that we still don't have a hard definition of what a useable base installation of Linux entails. Yes, we have LSB, but it really doesn't seem to cover enough ground. They waste a lot of time documenting exported functions, when really they should simply state library versions, and maybe even keep a copy of the appropriate source (even a precompiled copy?) available on their site so there is no question about what version they are referring to.
In addition to fixing the LSB, distros really should start obeying it. It certainly would make things easier for us end users. Is RedHat 9 even LSB compatible at all? I never see anything on their website about it, but I've continually heard from various sources that "the next version is LSB compliant".
I'm sure my remarks have pissed more than a few people off, who will undoubtedly attack my credibility. So for the record, Linux could stay non standardized for all eternity, and I'd still have no problem using it myself. I'm only putting these arguments forth since I feel that they are the real reason that free software isn't as mainstream as we would all like.
Rebuttles and counter arguments are, of course, always welcome.:)
Slashdot Mirror
I'll be frank up front too. I'm an Independant, with a capital (I). I believe in clean drinking water, not totally fucking the eco system, keeping the economy rolling, keeping my gun rights, moving away from fossil fuels, nuclear energy, electric cars, the space program, an adequate military budget, and all the other stuff from both sides of the fence that actually make sense.
Now that we've gotten that out of the way, explain to me how Fox News is actually sided in the political schema. Seriously now, I'm not being sarcastic. I'd really like to know why you believe that channel has a Republican agenda.
Let's look at their primetime lineup....
Brit Hume - The serious looking old guy. Covers news he reads off the cue-o-matic machine, doesn't add any of his own comments.
Shepard Smith - Appeals to the ladies, the guys think he's funny. Covers new (cue-o-matic! (tm)), adds his own comments, but they're always smart ass remarks trying to be funny.
Bill O'Reilly - The big O. This is the funny one. All the Dems swear he's pro Rep, and the Reps swear he's pro Dem. Independants (like me) like him because he walks both sides of the fence by making logical decisions, rather than simply believing hype. If someone proves him wrong, he admits his fault, and re-evaluates his position on a subject. I've even seen him change his opinion on the air (rarely happens as he is quite well informed, but it does occur from time to time). He's said plenty bad about Bush, and plenty good about him too. Same goes for Kerry. Good show, people should watch it, even if they don't like it. You'll change your opinion of the show after a month or two, I guarantee it.
Hannity & Colmes - Hannity is openly a Republican, Colmes is openly a Democrat. They share a show so the world can watch them argue over issues that they feel are important. Sounds like fairness and balance to me.
Greta Van Susteren - Supposedly one of the best trial reporters around. Why you'd have a reporter specifically for covering trials, I don't know. Do I care what the atmosphere of the Kobee Bryant courtroom was like today? Not really. In fact, I don't even care what the outcome is. Someone else's trial in some other state, does not directly mean anything to my life, so why should it be my business. She doesn't seem to go out of her way to get into political issues.
So let's see now. We have a bunch of news shows that cover (imagine this) the news. None of them really going out of their way to bring politics into the mix as far as I've ever seen. And two shows that actually specifically cover political issues. One is Bill O, and the other is H & C. The Democrats do seem to be saying that Fox News is pro Republican lately. Personally, I chalk that up to the upcoming election. I have noticed a liberal slant to CNN, and I'm sure they'd much rather people watch that instead.
Yea, I'd have to agree with you on that one. I've been typing for over 15 years now myself. I was the master at the 4 finger typing skill on my old 286 long before I took typing when I was in highschool. I, of course, played ball to the bullshit typing methods that they tried to teach me, but very quickly made up my own modified style that allows me to type all day every day with no problems at all.
;) So the keys aren't in the "magic hand-key relative location (tm)". Bah. I still type over 50wpm and have well less than 3% errors. In my opinion, typing is a skill that you need to have, but everyone is comfortable in different positions. The most difficult part of typing is learning where the keys are, so simply teaching computer skills to people and peaking an interest in them will be more than sufficient motivation for them to learn the keys. Which of course will allow them to make up their own bullshit typing skill that works well for them (like apparently we both did), and everyone is happy.
Hands perpendicular to the keyboard...
Elbows off the table...
Don't rest your palms on the table...
Yea, and make sure to sign up for medical treatment in advance...
My hands are almost always at a 30-40 degree angle away from being perpendicular to the keyboard, and I lazily rest both my palms and arms, and elbows on the table (I have a pretty swift desk I built for myself that emulates having two desks make a corner. It helps a lot with the whole resting of elbows
Dust & dirt. I would imagine that at such low voltage levels, induced current would require a damn near perfect level of alignment between the chip and the "socket". This is admitted in the article. What they don't admit is that it's going to be nearly impossible to get the damn thing in the socket without letting dust or dirt inbetween the chip and the socket.
And a more interesting topic is their consistent mentioning of taking the cache of the chip. That's a nice dream and all, but where the hell are you going to put it then? Hardwired onto the motherboard? That's going to dramatically increase the cost of mobo's (so they are simply shifting who gets to eat the high sticker price on their products). And what if I buy a quad capable mobo, but only put 2 processors on it, I'm effectively wasting 2 sets of cache, rather than simply wasting 2 cpu sockets, and the sockets are a hell of a lot cheaper than the cache. I suppose you could fix this by going back to COAST (cache on a stick, yeah i know you remember that nasty stuff). But that brings in a whole new problem: These days, cache is only fast because it's so close to the cpu. If they move it off the die, it's just going to be put back on in 2 years because we can't access the cache fast enough ever since we moved it off the die.
I'm no super computer engineer, but these guys better have an entire family of rabbits they plan on pulling out of their asses or this fucker's gonna flop.
You seem to miss the fact that in my solution you're already logging traffic destined to port 25. When a user's traffic reached a threshold, it's flagged. There's a reason I didn't require the use of a proxy in my scenario. Mainly because a single proxy would die in a matter of seconds as the thousands of emails send by a large customer base slammed into it, and customers would start bitching because they can't send email. With a packet sniffer, you have unobtrusive access to the same information you're going to get from a proxy, without the detrimental results if packets are lost.
Whoah there buddy. This is a good idea, but you have to check to make sure you aren't about to use a hammer to fix this problem. I can think of one cable modem provider (*cough* Charter) that uses this approach, but does it quite poorly. They firewall their customers on port 25 both inbound and outbound. The only acceptable traffic on that port is from the outside world to their mailserver, and from their internal network to their mailserver. This causes major problems when their customers are trying to use another provider's mail servers to send mail.
A more elegant solution requires a few steps. Firewall port 25 inbound to their customers (so they can't run their own mailservers). 99% of end users do NOT know whey they are doing with MTA's, and you will end up with lots of open relays on your network. Next you need to somehow log your user's smtp traffic. The best way is to put a packet sniffer on the network, and then have it tally up traffic and total sent emails by source IP address. This can be used to find out who is sending the most individual emails. Now you set up the packet sniffer to perform more detailed sniffing on that specific user's outbound email. Particularly, find out how many mails they send that are only 10% different, and how many times they send email to the same set of 5,000 addresses in a month. If you determine it's spam, tell them to stop. If they don't, cut their service.
This really isn't rocket science. They just need to be very careful how they resolve this problem. People don't like suffering because a few bad eggs amongh them were causing problems.
You seem to fail to see that a lot of our cool gadgets and high technology items have their roots in military research. Radar guns for instance, velcro, etc.
Plus military research has a habit of taking previous ideas and prototypes, and turning them into reliable machinery, like large aircraft, submarines (think exploratory subs), and even cars and trucks.
If you want someone to blame for the loss of scientific dominance, then blame ourselves. We allow government institutions like the patent office to continue stiffling innovation with meaningless patents (like software patents). We allow our representatives to draft and pass crazy laws like the DMCA that prevent reverse engineering so that our aspiring engineers cannot learn from the works of those that preceeded them.
Don't blame the budget, don't blame the government, blame us. We are the ones that allowed it to happen.
For those who don't know, the "burninating" reference is from Strongbad. Very funny shit. I highly recommend it.
http://homestarrunner.com/sbemail.html
What information are you basing this on? If MS bought a license, and they are less clear than Sun, what makes you think that buying a license has anything at all to do with being clear? From the sounds of it, for all you know, HP could have been the one that shelled out for the license, and Sun is still in talks.
"4 big vendors" is not very specific. Try not to spread unfounded information. It doesn't help the situation at all. I honestly don't see why Sun would care anyways. They are primarily a hardware and systems support vendor. Their OS is just there to give their platform something to run. Solaris for x86 only exists to provide their customers an easy upgrade path to SPARC-based systems.
At any rate, I'd really like to see where you get your info, or if you are just speculating.
I must admit, I've never been very good at setting up printers in Unix. If I don't have access to the Redhat printconf gui utility, I'm pretty lost. This is bad considering that I'm a unix admin at heart. I guess I've never really had the need to configure many printers on Unix boxes, and when I do, it is always conveniently enough a RedHat box.
I might just have to pick this book up. Anyone have any other suggestions on how to demystify printing in Unix? I understand how to use the lpr command, and how to kill jobs with lprm and list them with lpstat, but I'm pretty much a noob at configuring printers. A complete guide on how it all works would be nice. I'm pretty sketchy on the whole "filters" idea, and wouldn't know where to start to set up CUPS or LPD if all I had was a command line available.
I usually use rsync for one way backups, and unison where I need 2 way synchronization.
Rsync is nice because you can update lots of files very quickly, as it only moves binary diff's between files. Also, if it is a costly network link, you have the option to specify max transfer rates, so you don't kill your pipe when it runs from your cron job.
Unison is nice because it is pretty smart about determining which files should be moved, and can correctly handle new and deleted files on either end of the link. Plus it supports doing all of it's comm via ssh, so it's secure.
rsync
unison
The downside to both of these being that neither of them are instantaneous. However, I've had much success running both of these as often as every 5 minutes. Just make sure that you launch them from a script that is smart enough to check for already running instances before it starts trying to move data.
Technically, FPS is likely to be more limited by your GPU than the processor. In these benchmarks, the CPU is going to spend more time compressing textures, shifting data around, and trying to push data down the AGP bus than anything else. I'll be more impressed when they start using software rendering when they do these benchmarks.
Realistically, almost all of these game based benchmarks are crap. If you want real statistics, have it render CGI movie frames. THEN see how it compares to other CPUs. Until you do that, you aren't really looking at anything other than the CPU's ability to feed the graphics card.
Apparently your magic-8-ball has never had the pleasure to use it personally.
My magic-8-ball has been formerly introduced to the product. When I asked it's opinion, I recieved "Outlook not good".
Try exposing your 8-ball, and rechecking it's results. If the message does not change, you may have a defective 8-ball, and should call the manufacturer for a replacement.
actually, the parent you replied to is correct. Neither scp or rcp are "defying" what he is saying. They don't copy files with the execute bit turned on. Rather, they copy the file, and its permissions. Same goes for cp and mv. They all try to maintain permissions as best as possible. That isn't a default of the open() posix function, it is an option that the application in question (scp, rcp, cp, mv, etc) all choose to use.
So like he said, unless the email client sets the execute bit for you (which would be really dumb), attachments won't be executable until you make them so manually.
anyhow, cheers!
Please share your method of remotely administering Windows boxes. I'd really like to know how to do this, especially with Win2k. Seriously! This is only a mild poke at your comment! If there is some magic way to do what you are saying with Win2k that I don't know of, I'd REALLY like to know, because it would save me hundreds of hours a year trucking my ass from PC to PC.
I'm not familiar with many (or any) of Larry's speaches, but realistically I can't think of anything better than replacing PC with dummy/smart terminals. Let's face it, network administration was MUCH easier when you could simply replace the junk terminal with a new one, rather than having to reinstall an OS every 2 weeks. (Which is basically what it amounts to when you have about 100 PCs to support. Some PC, somewhere, goes junk, and they are too expensive to simply replace with new ones when they die.) Not to mention the single point of administration. I much prefer SUN's methodology. The network is the computer. You have a terminal/workstation, and lots of services all split up all over. But the key is, your workstation doesn't have to really provide much, or any of it. All it needs to do is give you a method of input and output. Mouse/keyboard, and screen. Beyond that, you have workgroup printers on the network, fileservers in the server room, and the admin can install any new software (accessible to all machines that need it) from one location and in one fell swoop.
:)
:)
Much much better than trucking my ass to each PC so that I can install the new app when there are 100 PCs involved. Even better when the number of machines in question reaches the thousands.
Anyways, the software industry isn't dead. There will always be a need for new software. Business models are different between companies. That's how they compete. They excel in different areas, and to do that they need different software. Software that more closely meets their needs. Saying software is dead is akin to Steve Balmer saying that opensource and free software don't innovate, but Microsoft innovates all the time. MS buys what they think is cool, and reshrink wrap it with a new label. Free software is honestly the ONLY place where innovation occurs. Someone has an idea, and they run with it. The idea may not be polished, the software may not even be implemented that well, but it isn't the software that is being questioned, it is the innovation. The software becomes polished when some company buys/steals the idea, or when a new company is formed specifically to flesh the idea out.
If you people actually believe that all the software ever necessary is already written, then please do get out of this industry. Go write a book or something, please. Let the rest of us innovate in peace.
(BTW: I'm not trying to attack the parent thread, I'm just spewing ideas.
One quick question: How does giving the first 100 or so free prevent you from paying for email worms that continuously spam everyone in your address book? I have about 300 contacts in my address book at work. But even if I only had 5, by your idea, after 20 cycles of spams (which I would likely never see happening) I'd still start to pay for it.
This won't work for two reasons.
Open mail relays and forged message headers.
If you can't track the source, you can't bill them. So then who do you bill? The company with the open relay? Some would say that's a good way to promote good system administration, but remember that the bill imposed could easily put a company out of business and into bankruptcy. Sounds a little strong to me.
I still feel that we are better off not having a mandatory tax. Instead, set up third party message verification systems. Emailers can, for a fee, have their message ran through an intense one way hashing/encryption system to create a special "Registed Email" message header, which is then sent along with the original message to the intended recipient(s). Using this system is entirely optional, but read on for the benefits of using it at least once per recipient.
Upon reciept, the recieving email client will see the special header, check it's validity with the issuer, and place it in the user's inbox. If the message does not have the 'registed email' header, then the sender's name is checked against a list of known users. If the user is known (from having been manually entered or already recieved 1 registed email in the past, and not in the blocked senders list, the mail will again go into the users's inbox. All other mail is automatically placed in a folder of the user's choice. If that means the trash, fine.
There you go. Don't need to even care about open mail relays, because if you've never heard of them before, and they don't send registered email, you'll never see their penis enlarging message. I've thrown this idea out before, but I thought I'd see if I could get more feedback on it.
Suppose now that the clients update the same record on the two servers in an incompatible way... you could imagine what will happen when the servers become visible to each other again...
This is why you have transaction logs that are timestamped. When the sytems resync, they merge their transaction logs, rollback to the last synced state, and then re-execute every transaction until they are current. The end result is that the newer row updates will overwrite the older row updates. This may or may not be the desited effect, so you must allow this operation to be turned on or off. What most systems do is forcefully synchronize all write operations amongst master servers. By this I mean that if a master server becomes unreachable, the update query is blocked until either it becomes available again or times out with an error.
Most problems with replication, though, have more to do with poor software design than the database. In order to make the multiple-masters system work correctly in an uncertain environment (pretty much every case), you should really have two tables. One local, and one global. The local table is non-replicated and queues operations to be posted to the global (replicated) table, and a background process runs on the local server to try to push them to the global table as connectivity permits. This allows you to have pending operations, and also allows you to also have a conflict table to store conflicting updates, which can then be later resolved by flesh-n-blood v1.0. Also, SQL queries that insert into one table and then update or insert another using the last insert id from the previous are often written incorrectly. You should really never call the db's function for last insert as part of a separate select, and then use the resulting hard number in the next table's insert/update. Instead, you should incorporate the function call into the 2nd table's query, so that if the insert occurs on an unsynched system, the result of replaying the transaction log will still result in an accurate id number (thus preserving the row relationships). The id may have been changed, but at least it will still have the relation intact. Obviously, all of this needs to occur within a single transaction to make it work. Coincidentally, because of this, the id column makes for a poor unique row id if it is an integer or some magically created string by the client software. You're much better off using GUIDs (or UIDs in linux), as they are based off of more unique algorithm (usually, high res date/time and NIC MAC & some random number).
On a related side note, 'master server' is kind of a bad label to use, since true RDBMS's have better granulatiry than that. Usually you can set master/slave properties at the table level, and still most even let you set master/slave at the row level (as defined by a where clause), as well as published data at the column level (as defined by a non-joining select statement).
In all honesty, I do agree with your argument about replication not being easy, but I take the stance that the truly hard (and critical) part is writing software that correctly handles being in a replicating environment. It's easy to write software that uses a database backend. but it can be a bitch to make it run correctly in all cases in a replicating environment. There is only so much that DB replicating can do for you, the rest is up to the software to keep the data sane during disconnected states.
Isn't clustering supposed to be a function of the database system, not the software you use to access it?
I mean, this is neat and all, but I really don't want to have to use this interface just so that I can cluster my database. You're much better off placing clustering functions within the database itself. Then you can access the data by any method (ODBC, native libraries, hell even with the provided command line interface).
Take a look at how MS SQL Server performs clustering sometime. Everything (and I mean EVERYTHING) is performed via triggers and tsql. All the clustering setup does is set up a bunch of known working trigger scripts to propagate the data. You can even edit them to your liking afterwards if you wish. Now I'm not saying that MS's solution for clustering is the cat's ass. Personally, I think it is kind of hackish, but then again I believe that clustering should be something you simply turn on, and shouldn't be able to fuss with. Realistically, I can't think of any good reason to change the cookie cutter tsql scripts that perform the clustering, so I only see the ability to modify them as a potential way to fsck it up (that being an obviously bad thing).
Clustering really isn't that hard to implement. I'm pretty surprised that MySQL and Postgres don't have better support for it. Especially Postgres, since transaction support is really the one big key that makes clustering possible. Maybe no one has really had an itch to make it heppen yet. Hopefully it will happen soon, since I'd love clustering to be another argument for why OSS databases can play with the big kids just as easily.
You must also consider the implications that using C# has on your scripting language choices. I can't think of many big games that don't have a scripting language of some sort. Usually a company will write or license an engine, and then extend it into the game they are trying to create. This involves scripting on some level. A lot of games currently use Lua, I've seen some that use Python, and some that use their own custom C-like language. The down side with C# is that I'm not aware of any C# bindings for the popular scripting languages. Obviously a custom script engine could be made to have C# bindings (or could even be written in C#), but most games don't use custom script languages, and will instead use an existing language simply because it lessens their development and bug finding time.
Correction:
American corporations have too much power.
The problem with the US is not that they have too much power, it is that the government is corrupted by private interests. Like you stated, Rumsfeld used to head a pharmaceutical company, and probably still recieves kickbacks. Thus his views are bias. If you remove the corporate interests from the equation, the US really aren't that bad. But yes, in its current form, the US is a pain in the ass.
And surprise surprise, I'm an American. And yes, I actually believe that our government is overrun by politicians that are too easily swayed by pocket stuffers, and are more interested in getting re-elected than doing the right thing. Unfortunately, short of a complete revolt, there isn't much that we the people can really do about it. It is almost impossible to elect anyone who isn't already corrupt or will become correupted, and those few elected officials that are not corrupt are powerless to change the current system, as they are severely outnumbered by their "peers".
Personally, if shit gets too wierd over here, I'll just move. I like some of the laws and freedoms I have over here and I'm used to the language, but realistically I'm still relatively young, and could adapt to a new country with a fair amount of success. But, I hope it doesn't come to that.
I can name at least 20 non-mainstream bands that I listen to. Some have recording contracts, some just sell CDs off their website. So how do you determine which of these bands gets any of the compulsory license fees? All of them, since they all sell music for profit? Only the ones with record deals? None of them since they are not mainstream? What is the criteria for getting paid? It seems to me that compulsory licensing would never work, since you really can't even decide on who to pay.
actually, i have a reason for you. ;)
Prism2 support.
I have a DWL-650 (2nd gen), based on the prism2. I had used wlan-ng in redhat 7.2(or was it 3?), but gave up on wlan-ng as it was pretty lame on the configuration side, and too much of a bitch to implement. (Kernel recompiles necessary, ripping out all the original pcmcia support and replacing it, etc) However, RedHat 9 supports my dwl-650 right out of the box... er... bitorrent acquired iso burned to cheap cdr media...
So there you go. Good reason right there. I used to have to run XP to use my wireless card, now I don't. Yay for me.
Oh, and gnome 2.2 is actually cleaner than 2.0. Expecially the fact that meta themes are now officially implemented, and the new menu system isn't as freaky as that funk ass "Extras" submenu.
As usual, your mileage may vary, but all in all I'm quite happy with RH9. If I wasn't dirt ass broke, I'd probably go out and buy it just to have real media.
I don't think that there are too many free software projects out there. Have you ever gone to download.com and just looked around at all of the thousands of Windows applications? Many many of them perform the same function. Some better than others, some are innovative, some are not. That doesn't seem to stop people from downloading and using them.
/opt! :) Not to mention the lack of automated installers for most projects. The installers exist, people just don't use them. I'll admit, some of them are a little lacking (a scriptable installer ala InstallShield could be helpful), but I have plenty of Linux games that use those simple installers, and they work great, despite kernel upgrades and distribution changes.
./configure too much these days. I've always felt that build scripts like ./configure were useful for the developers and hackers, but the general public really shouldn't have to have gcc installed just to install new software.
:)
Quantity isn't the problem. The problem is quality. Well, percieved quality anyways. Unix has a different paradigm when it comes to software installation. That's a fact. There is no 'Program Files' folder that everything is installed to. Of course, there's always
IMHO, people just rely on
It also doesn't help that we still don't have a hard definition of what a useable base installation of Linux entails. Yes, we have LSB, but it really doesn't seem to cover enough ground. They waste a lot of time documenting exported functions, when really they should simply state library versions, and maybe even keep a copy of the appropriate source (even a precompiled copy?) available on their site so there is no question about what version they are referring to.
In addition to fixing the LSB, distros really should start obeying it. It certainly would make things easier for us end users. Is RedHat 9 even LSB compatible at all? I never see anything on their website about it, but I've continually heard from various sources that "the next version is LSB compliant".
I'm sure my remarks have pissed more than a few people off, who will undoubtedly attack my credibility. So for the record, Linux could stay non standardized for all eternity, and I'd still have no problem using it myself. I'm only putting these arguments forth since I feel that they are the real reason that free software isn't as mainstream as we would all like.
Rebuttles and counter arguments are, of course, always welcome.
Cheers!
Probably because Freshmeat.net already beat them to the punch.
;)