For some reason, I'm reminded of a thing I heard about that happened (I believe) sometime in the 1970's or so. The scientists who were working on recombinant DNA became worried about what they were doing. They entered into a worldwide moratorium on many types of research, for fear of what the technology would lead to. That research stopped for many years, until the researchers discovered two important things. First of all, the DNA work they had stopped doing was still going on, in nature, as it had been for millions of years. Second, what they had stopped doing was effectively the same as the selective breeding that has been going on since humans discovered agriculture (albeit faster and more precise).
That's what we have here. A faster and more precise version of something we're already doing. As Jon pointed out, people already spend quite a bit on the seeds of a person with an attractive phenotype (and therefore probably attractive genotype). So you want your children to have blond hair and blue eyes. OK. Why? Because you think that is attractive. Look over at your SO, the other half of the gene pool from which your child will spring. Blond. Blue-eyed. "Attractive". Coincidence? I think not. People have been weeding out undesirable genes for as long as the species has existed. This is merely a scalpel to replace the broadsword.
The other thing that interests me is this: There is a belief that when a species becomes intelligent, begins to use tools etc., evolution stops. That species is no longer shaped by the environment, but instead shapes its environment around it. If you think about it, natural selection is kind of passe these days. Natural selection is only concerned with birth and death. We have for some time had the ability to choose whether or not to have children. And, even as much death as there is in the industrialized world, only a small percentage of people die befor they reach the end of their childbearing years. What does this mean? It means that the death rate has little effect on evolution. And many of the strongest and smartest among us choose to have few children, or none at all. There is an evolutionary preference towards people who are too stupid to understand birth control. (also towards those who are smart enough to use it, but choose not to. That's a different thing, and I assume them to be a reasonably random distribution among the population, therefore an insignificant pressure). But this leads to an interesting thought. It's nature's own bootstrap code. First, the environment evolves us into intelligence. Intelligence halts evolution. But then that same intelligence gives us the tools to shape our own evolution.
More accurately, what is currently called a switch is a multiport bridge. Well, actually an Ethernet switch is a multiport bridge. An ATM switch or Frame Relay switch is more like a real switch.
But, to be more informative, and to relate to the original question, routers will not be replaced by switches (read: bridges). Bridges theoretically have to know the address of every device addressable on their network (read: broadcast domain) Routers deal with layer 3 protocols which have (or at least should have) hierachical addressing. They do not have to keep a forwarding table for every device, only one for every subnetwork, network, or supernet. The current internet has on order of 80K routes. That stretches the capabilities of some routers. I know of few devices that could keep a forwarding table of a billion addresses (one quarter of the IPv4 address space.... which is running out.) and do anything usefull with it. What will happen (and is happening) is that routing functions will begin to be integrated with switching functions.
The way Cisco's "Layer 3 Switches" currently do things is this: The switch is connected to a Route Processor, usually a 7000 series router or an RSP blade. If a packet comes in that is destined for the MAC of the router, it's layer 3 destination is examined and placed into a "flow." When that packet comes back out of the router, that "flow" is completed, and further packets from the same port with the same destination get sent out the port that the original packet went out of, rewritten slightly to reflect what would have happened if it had gone through the router. This increases speed in two ways. First, the packet didn't have to put up with the latency in the router. Two, it only had to traverse the backplane once. There are tradeoffs, of course, but it is a nifty technology. Note, however, that you still need a router in this scenario. And when you don't need a router, it's because your switch is acting like a really stupid (but really fast) router.
Electricity is essentially the movement of electrons, not protons. Protons do not (normally) move from atom to atom. Electrons certainly DO NOT travel faster than the speed of light. Electrons have finite mass; if they traveled at the speed of light, the Lorentz transformation equations assert that they would infinite mass. I assure you that the electrons in your computer do not have infinite mass. In fact, electrons usually travel far slower than that. The conduction speed of copper wire is in the range of a few thousand meters per second at best. Light (well, actually photons) travels faster than anything else. The speed of light is an absolute speed limit in our universe and is a fundamental physical constraint.
OK, I'm nitpicking and somewhat offtopic, but...
As someone else pointed out, the speed of propagation of an electrical signal has little to do with electrons moving. In normal applications, electrons move pretty slowly. The electromagnetic forces from them move much faster, IIRC they propagate at the same speed as light.
The Lorentz equations do not assert that a particle with finite rest mass would have infinite mass at the speed of light - as it approaches the speed of light, its mass approaches infinity, at the speed of light it is undefined (division by zero, and the universe dumps core), and above the speed of light it is a complex number. I believe the factor is 1/(sqrt(c^2-v^2), but I may well be wrong.
The speed of light is not an absolute speed limit. The speed of light in a vacuum, as far as we know, is. there are applications where electrons will travel through some medium at a speed faster than the speed of light in that same medium, although slower than the speed of light in a vacuum. There are media through which the speed of light is very slow. (I recall somewhere recently I saw an article about researchers who had found a medium through which light travelled at something like 38 m/s.) This is the source of what is known as Cherenkov radiation. (you can find a small blurb about this here).
Well, no one likes a nitpicker, but I hate to see someone so close to being right.
I agree that SMTP AUTH has a high coolness factor, but the one thing I can't seem to find anywhere on the sendmail site is a list of clients that support it. Netscape messenger? Pine? Outlook? Does anyone know?
It's so interesting to see other points of view. My dad started out teaching college psychology, working on his doctorate. He decided he was sick of it, and dropped out to work for Roto-Rooter, and was happy as a clam. Instead of going in each day to face students who didn't want to be there, and who certainly didn't want *him* to be there, he went in each day and worked with people who were *really glad* to see him. Then he started doing construction work, eventually became an inspector for the city, and liked it well enough, but didn't like having a boss. So he and my mom started up their own electrical contracting business, and he was happy as a clam again, because he was his own boss, and was pretty much doing something that he was good at and that he liked doing. Eventually, he and my mom split up, and she got the company (no hard feelings about it, it just had to go to someone) and he started working for another contractor - and loved it. He doesn't have to worry about how well the company is doing anymore, doesn't have to worry about much of anything, they just give him work, and he just does it.
Then there's me, I'm a network engineer, and this past year I paid more in taxes than my dad made, actually more than my parents made together in any year when I was growing up with them. But that's OK by them. My dad's doing what he likes to do, and making enough that he doesn't have to worry about money too much. I'm doing what I like to do, and will continue to do it no matter how much (or how little) they pay me, so long as I can pay my bills OK. I don't feel guilty for what I make, because I firmly believe that the worth of a thing is what the market will bear. Right now, that's a lot.
Strangely enough, the only thing I regret (job-wise) is that I'm not doing manual labor. I remember when I was in high school, I spent a summer laying concrete block. Hauling 12" block around, lugging 5 gallon buckets of mortar up 30' of scaffolding, taking a 15 lb. or so block and placing it *carefully* in place. I was in better shape than I've ever been in! And when I came home, I was tired. I could sleep. It used to seem wierd to me that people would pay thousands of dollars to go to a gym and, basically, do manual labor. But maybe it's not that wierd. But, I'll continue to do what I do, because I still have more fun exercising my mind than exercising my body. And who knows? Maybe if the tech market crashes hard, I'll go back to laying block. Endless possibilities....
Does it seem ironic to anyone else that the "crime" he is accused of basically boils down to figuring out the "passwords" (encryption keys) that DVD's use, and then when the police take him in for questioning, they not only seized his equipment, but they required him to give them his passwords?
Can they learn from it? The post office (and ups and fedex and etc...) do some nice things. Package delivery is one. Message delivery is another. (although I haven't sent a piece of actual mail in three years, but I have received several) What I would really like is for the post office to start to learn lessons from the net.
I have an internet address. it is 209.116.217.40. It is also phoenix.hppc.com. If I change ISP's, my ip will change, but phoenix will not. We all take this for granted, and most of us know how DNS works, but it's pretty nifty nonetheless.
I also have a postal address. If I move, it changes. I can do the whole change-of-address form, and tell people my new address, and so on, but in the end, chances are I'll lose some mail.
Wouldn't it be nice if I just had a logical address that mapped to a physical one? If I move, I just contact the authority for my address, and let them know that my physical address has changed. My logical address continues to work. And even better, if I want to give someone my address so that they can send me mail, I don't have to let them know where I live.
The USPS seems like a logical place to keep this database of logical to physical mappings. Does anyone else think this is a good idea?
I'm also an MCSE and Linux advocate. I'd have to say that you don't have many choices, and my first reaction is that if you don't like something about a package at first blush, chances are you'll find other things you hate about it deeper in. If you have problems like this, conversion to some other package might be your best option.
the next question is, do you have to upgrade? The version you have works, on the system you have. If when your vendor calls you and tells you that they've got their new version out and you should upgrade, just say "Sorry, but your upgrade won't run on our servers. As soon as it does, we'll look at it." Who knows, they may even pay attention to you. (but probably not. The company I work for is one of two dozen customers using the accounting package we have. Even though we represent about %5 of their market, they don't listen to us.)
But, in the end, I will have to say that adding an NT/MSSQL server shouldn't be too big of a problem. The biggest problem with most MS systems is admins and users mucking around with things they don't understand. If you just install NT, and install MSSQL, and pretty much leave the server alone after that, chances are it'll do OK. Microsoft makes the occasional great product, a few "fair to middlin'" ones, and many that really suck. NT Server and MSSQL seem to me to fall into that middle category. If you have reasonably competent people who know enough to run some other SQL server and are generally fairly careful, chances are you won't have any problems. The biggest headache I suspect you'll run into is figuring out how you want to do user authentication. Most accounting packages deal with this poorly. Very poorly. But now and then there's one that does, and if you want to be really slick you could set things up such that your users are logging into sql via integrated security (i.e. it authenticates against the NT OS password), in which case you may want to put your PDC on one of your linux boxen.
But that's just a suggestion. Personally I've always had a lot of fun with integrating Linux and NT networks. But then again, I'm kind of a sick and twisted person that way.:)
Finally, none of your choices is all that bad. If it were me, I'd ask the vendor to make changes, and if they ignore you, ask your bosses to try and find a better vendor, and if that doesn't work, oh well, there probably won't be too many problems (and you'll get something new to put on your resume)
I have a few friends who use SFNB, and they all seem quite happy with it. The only drawbacks are 1) the lack of brick and mortar branches (which really isn't a problem unless you want something like a safe deposit box.) and 2) the lack of SFNB ATM's. Most of the ATM's I've seen charge you some fee (usually $1.50) for their use if you're not a customer of the bank that put up the ATM. Of course, a lot of grocery stores etc. are taking atm transactions now, and will give you change on the transaction, with no transaction fee, so if you plan things reasonably well and go to the store often enough, it works out well.
Personally, I use First Union, and they've just recently gotten their heads out of their butts and become pretty decent in their online offerings. (It used to be that their online bill payment system used a netscape plugin that only worked on 95/98, not even NT, certainly not *nix).
The system they have now uses a pure web-based system (currently supplied by checkfree, but I've talked to them and they're building their own improved interface) for bill payment, bill presentment, etc. The interface they have is fairly nice. It's cool to be able to tell it "Ok, every month pull down my phone bill and if it's not over X dollars, pay it automagically." They have a nice system for statement download for checking, savings, and credit cards. They have a section for brokerage services, but I don't use it so I can't comment on it. Personally, I'm happy with them. If for no other reason than when I call their tech support, I usually get someone clueful.
They have demos set up of most of their functionality. Browse through it and see if it looks like what you want.
Well, it may not do much good, but by registering their domain, they do give up some privacy. whois lists their billing and administrative contact as admin@ETOYS.COM, and their technical contact (which would probably be the least helpful) as hostmaster@ETOYS.COM. One of these would probably be able to answer questions as to who to contact for "customer service", comments &cetera and their legal department.
Or, you could always send them a letter.
etoys (ETOYS3-DOM) 3100 Ocean Park Blvd., Suite 300 Santa Monica, CA 90405 US
They might even care. (but probably not).
Or start a rumor. That's the address that Santa lives at. spread it among the children.;)
Well, I'm sort of interested in this... From NetCraft, it seems that static and products.mars.ucla.edu are running apache on solaris (as the article says), however www.marspolarlander.com shows up as apache on BSD/OS. The article dosen't say what the NetCache servers are running, but it seems that it may be a BSD derivative.
Notice that the article wasn't a post from the webmaster, but from the NetApp company (which makes sense, it's good press for them and their product).
I also note that, contrary to the original post, they don't mention Linux anywhere.
I do question their choice of OS though. I love Solaris, but everyone I know who has run Solaris x86 has been disappointed by it. For sheer processing horsepower, and the ability to move a lot of data across a system bus, you just can't beat a SPARC. (although, there was mention of a linux port to the S/390, that might do it...)
But anyway, my overall impression of the setup is: Rox!
The software is still out there, and will probably run under dosemu just fine. Of course you won't get any added functionality beyond what the software originally had (well, things like hard links and nfs, but that's not really the application itself...)
Of course I've never tried it... and I don't even know how some things would work... I mean, WP made extensive use of the alt-F? key combos. How would you switch virtual terminals? I don't know, but it's a thought...
Well, the original problem seems to be that pricing updates come in mdb format. Although I can certainly agree with the original poster that I certainly wouldn't want to put an e-commerce server on NT, it seems to me that one NT workstation with a copy of MS Access shouldn't be too much of a problem. You create an mdb with linked tables, one linked to the file location where you will put your pricing updates, and one linked through the MySQL ODBC drivers to the MySQL (or PostgresSQL or whatever) databases. Then a little AutoExec macro, and you've got yourself an "update application".
The problem with this idea is that it doesn't actually enlarge the effective keyspace. It could be effective against someone trying to attack *your* entropy, but becomes less so when someone is trying to attack *everyone's* entropy. To put another way, you can do a dictonary attack against a password, and that may be effective. With a workload scheme, you feed your dictionary into the workload function *once* and create a new dictionary, which may be a bit larger than your original, but has the same odds of success.
The more people who use the same function, the more economical it is to do this.
Yes, trademark erosion is a bad thing for the companies whose trademarks get eroded. If your boss tells you go order a Xerox machine, are you going to make sure that it is from Xerox corporation? Or are you going to get a copy machine, perhaps the one with the best price or niftiest features, but maybe made by some other company? Same with Kleenex, Band-Aids, or, for that matter, Unix. People are buying products that they are calling by their brand name, but the company who makes that brand is getting nothing for it. That's bad for the company, and in many cases bad for the consumer. If on the other hand (let's see if I can think of a brand that hasn't eroded yet...) your boss tells you to get an APC Smart-UPS UPS, you will probably not just get a UPS, but one made by the company that produced the supposedly high quality products, and definitely high quality marketing that made the term "Smart-UPS" stick in his head.
The interesting thing is that most of the time, trademark dilution comes from a companies inability to get people to refer to their products as "{brand} of {item}". It would be much more straightforward and not brand dilution if you were to refer to "Band-Aid bandages" or "Xerox copiers", but when a brand goes from being an adjective to being a noun, the company is screwed.
IDG has some interesting issues here. Their product isn't getting "nouned" to death, but their trademark is certainly getting diluted. Unfortunately, there's not much a company can do about trademark dilution, because it mostly happens in people's minds. Eventually, they start getting overzealous, until there's nothing left of their trademark, and they eventually give up.
There is room for debate on this one. If someone other than IDG published a book called "Sendmail for dummies", it would clearly be a trademark violation. If someone created a website with that title (sendmailfordummies.com?) I would consider it trademark infringement, others would not. In this case, it's the subject of an e-mail, and I wouldn't consider that infringement. Others would. The point is, it's debatable, as we're proving by debating it right here and now.
I wonder if Hormel ever tried to enforce their trademark on Spam. That seems to be the most diluted trademark around these days....
(For those of you who don't know, at one point Spam was a meat food product, not junk e-mail.)
Re:One of the sequels was brilliant
on
Ender's Shadow
·
· Score: 1
At one point I had read all that I could find of Card, but it seems he's writing them faster than I can read them.:)
Anyway, if you read "Maps in a Mirror", it contains the "seeds" for most of his novels. (it also contains some extrordinarily good short fiction)
I definitely loved both Enders Game and Speaker for the Dead, but some of his lesser known novels are just as good: The Worthing Saga, Treason, Wyrms, and Hart's Hope, to name a few.
And I have to admit, he also did an incredible job on the novelization of a great movie: The Abyss. I thought the novel was as good as the movie, which is difficult to do when going across media. I have high hopes for an Ender's Game movie. I can't wait to see it.
Why do people on this site always have to relate something posted on Slashdot to Geeks? I don't love reading Enders Game because its a book on Geeks. I read it because its a damn fine read. Why do you people (aka Geeks) have to try to justify yourselves? I'm confused. That would be "F_ing Good Read", (working title), or "Damned Fine Novel" (release title). Get it straight. (Those who have "Maps in a Mirror: The Short Fiction of Orson Scott Card", and have thus read the short story "Damn Fine Novel", will understand... )
the problem with these is that they don't really understand ext2fs all that well, so it's doing, from what I understand, the same job as if you used dd if=/cdrom/image of=/dev/hda1.
This is fine so long as you've got a hard drive image that'll fit on a cd. But with a 650MB cdrom and a 18GB hard drive... I don't think it will work all that well. (note: I haven't tried this myself, it's just my understanding of how ghost works)
Basically, you know that whatever solution you use should probably use linux, as it should understand ext2fs, but once you get past there, there's a lot that you can do. dd an image from an nfs server, use kickstart, or for that matter, use tar. The biggest problem that I see is partitioning the drives in the first place. One of the nicest things about ghost is their gdisk program. (a completely command line scriptable replacement for fdisk. extremely customizable and very nice. If anyone knows of a similar tool under linux, let me know.)
make oldconfig does ask questions, however it should be pretty easy to write a small expect script to run make oldconfig for you, and answer yes to everything. I'm not suggesting that this will make a more stable kernel, but it will make one. Chances are you could even have the script call up the help on each option, and have that as output from the program, which would in turn get mailed back to you by cron. In the morning check and see if the answers made sense, if they did reboot, if they didn't recopy.config, make oldconfig again, and answer the questions the way you think they should be answered.
Now, as for the script itself... I have a few (lame) suggestions:
#!/bin/sh cd ~/
# replace this with the correct URL wget http://www.XX.kernel.org/pub/linux/kernel/v2.2/lin ux-2.2.12.tar.bz2
#I don't know about wget. It would be nice to be able to have it first look for #LATEST-IS-* and then get linux-(whatever).tar.bz2 #or even better, have your script add one to uname -r and try #to get that file. If it doesn't exist, exit.
# bzcat linux-2.2.12.tar.bz2 | tar xv # rm -f linux-2.2.12.tar.bz2 # cd ~/linux #I would suggest as an alternative keeping/usr/src/linux as a link to/usr/src/linux-{ver}
rm/usr/src/linux mkdir/usr/src/linux-`uname -r {+1}` #this doesn't work, I just forgot how to add. ln -s/usr/src/linux-`uname -r {+1}`/usr/src/linux rm oldlinux ln -s/usr/src/linux-`uname -r`/usr/src/oldlinux bzcat linux-{version}.tar.bz2 | tar x
# will this work? # not quite # cp/usr/src/linux/.config . cp/usr/src/oldlinux/.config/usr/src/linux /usr/local/bin/some-expect-script-that-calls-mak e-oldconfig
make dep make bzImage make modules make modules_install # and then cp/boot/vmlinux.new/boot/vmlinux.old cp/usr/src/linux/arch/i386/boot/compressed/vmlinux/boot/vmlinux.new lilo
Then you have a lilo.conf that looks like this:
# LILO configuration file # generated by 'liloconfig' # # Start LILO global section boot =/dev/sda append = "ether=0,0,eth1" #compact # faster, but won't work on all systems. delay = 50 vga = normal # force sane state # ramdisk = 0 # paranoia setting # Linux bootable partition config begins image =/vmlinuz.new root =/dev/sda1 label = lin read-only # Non-UMSDOS filesystems should be mounted read-only for checking # Linux bootable partition config ends # Linux bootable partition config begins image =/vmlinuz.old root =/dev/sda1 label = backup read-only # Non-UMSDOS filesystems should be mounted read-only for checking # Linux bootable partition config ends # Linux bootable partition config begins image =/vmlinuz.reallystable root =/dev/sda1 label = original read-only # Non-UMSDOS filesystems should be mounted read-only for checking # Linux bootable partition config ends # End LILO global section horizon:~#
have cron run this every night, and every once in a while you'll wake up and find e-mail telling you that you've got a new kernel. If everything looks OK, just reboot and it's active.
Personally, I haven't bought a stamp in approx. 3 years. I don't even remember how to send snail mail. Everyone I care about communicating with has an e-mail address. (and if I really wanted to get wierd, my pager can send "e-mail" to a phone number through a text-to-speech engine...) the thing is, I don't use encryption and digital sigs all that much, and it doesn't matter too much to me. Yes, I trust them much more than an ink signature. But that doesn't mean that I trust it all that much. Digital sigs I don't need at all, because I generally don't care who an e-mail came from. If it came from my dad, I can generally tell whether it's signed or not. The actual applications that I use these tools for are few and far between. I would prefer it if everyone used encryption, but things are getting better. There's a reasonable level of interworking between the mailers that I and my friends use (outlook, netscape, pgp-elm and eudora) I've had very few problems.
I guess my general answer is that it's happening already, and encryption is not an obstacle. Encryption is more secure than any snail mail you can send, and easier than your average certified mail.
We're running an NT SAN here, using the Vicom UltraLink 2000 SSA - SCSI routers. I know that they offically support Sun, HP, AIX, and several others, but the drives show up as SCSI ID/LUN. We've got the untralink's plugged into Adaptec cards. We haven't tried it, but there should be only a few problem with plugging this into a Linux box.
There are two problems that I see:
First, as the earlier poster mentioned, last time I checked linux assigns/dev/sda to the first scsi device it finds,/dev/sdb to the second, etc. I've never seen what linux does when, for instance, scsi id 5=/dev/sda and you add a device with id 3 and rescan the bus. Personally, I'd love to know what happens. It would be nice if there could be a switch from the way things are to a system more like (as I recall) Solaris uses which is not sequential, but definitive, i.e. incorporating into the device name the system bus/adapter/adapter bus/scsi ID/LUN. Of course that would be a pain for all of us that are used to the old way.
The second problem with the Vicom's is that you can only configure them from an NT box. In general this isn't too much of a problem, since you only set it up once, but if you really hate NT, it may be a problem.
They've got some interesting stuff. They are here.
Slashdot Mirror
For some reason, I'm reminded of a thing I heard about that happened (I believe) sometime in the 1970's or so. The scientists who were working on recombinant DNA became worried about what they were doing. They entered into a worldwide moratorium on many types of research, for fear of what the technology would lead to. That research stopped for many years, until the researchers discovered two important things. First of all, the DNA work they had stopped doing was still going on, in nature, as it had been for millions of years. Second, what they had stopped doing was effectively the same as the selective breeding that has been going on since humans discovered agriculture (albeit faster and more precise).
That's what we have here. A faster and more precise version of something we're already doing. As Jon pointed out, people already spend quite a bit on the seeds of a person with an attractive phenotype (and therefore probably attractive genotype). So you want your children to have blond hair and blue eyes. OK. Why? Because you think that is attractive. Look over at your SO, the other half of the gene pool from which your child will spring. Blond. Blue-eyed. "Attractive". Coincidence? I think not. People have been weeding out undesirable genes for as long as the species has existed. This is merely a scalpel to replace the broadsword.
The other thing that interests me is this: There is a belief that when a species becomes intelligent, begins to use tools etc., evolution stops. That species is no longer shaped by the environment, but instead shapes its environment around it. If you think about it, natural selection is kind of passe these days. Natural selection is only concerned with birth and death. We have for some time had the ability to choose whether or not to have children. And, even as much death as there is in the industrialized world, only a small percentage of people die befor they reach the end of their childbearing years. What does this mean? It means that the death rate has little effect on evolution. And many of the strongest and smartest among us choose to have few children, or none at all. There is an evolutionary preference towards people who are too stupid to understand birth control. (also towards those who are smart enough to use it, but choose not to. That's a different thing, and I assume them to be a reasonably random distribution among the population, therefore an insignificant pressure). But this leads to an interesting thought. It's nature's own bootstrap code. First, the environment evolves us into intelligence. Intelligence halts evolution. But then that same intelligence gives us the tools to shape our own evolution.
Fascinating!
More accurately, what is currently called a switch is a multiport bridge. Well, actually an Ethernet switch is a multiport bridge. An ATM switch or Frame Relay switch is more like a real switch.
But, to be more informative, and to relate to the original question, routers will not be replaced by switches (read: bridges). Bridges theoretically have to know the address of every device addressable on their network (read: broadcast domain) Routers deal with layer 3 protocols which have (or at least should have) hierachical addressing. They do not have to keep a forwarding table for every device, only one for every subnetwork, network, or supernet. The current internet has on order of 80K routes. That stretches the capabilities of some routers. I know of few devices that could keep a forwarding table of a billion addresses (one quarter of the IPv4 address space.... which is running out.) and do anything usefull with it. What will happen (and is happening) is that routing functions will begin to be integrated with switching functions.
The way Cisco's "Layer 3 Switches" currently do things is this: The switch is connected to a Route Processor, usually a 7000 series router or an RSP blade. If a packet comes in that is destined for the MAC of the router, it's layer 3 destination is examined and placed into a "flow." When that packet comes back out of the router, that "flow" is completed, and further packets from the same port with the same destination get sent out the port that the original packet went out of, rewritten slightly to reflect what would have happened if it had gone through the router. This increases speed in two ways. First, the packet didn't have to put up with the latency in the router. Two, it only had to traverse the backplane once. There are tradeoffs, of course, but it is a nifty technology. Note, however, that you still need a router in this scenario. And when you don't need a router, it's because your switch is acting like a really stupid (but really fast) router.
Actually, it's 1/sqrt(1-(v^2)/(c^2))
OK, I feel really stupid now... I didn't even check the obvious case. If the rest mass is not equal to the mass at v=0, something is definitely wrong.
Thanks for the correction.
Electricity is essentially the movement of electrons, not protons. Protons do not (normally) move from atom to atom. Electrons certainly DO NOT travel faster than the speed of light. Electrons have finite mass; if they traveled at the speed of light, the Lorentz transformation equations assert that they would infinite mass. I assure you that the electrons in your computer do not have infinite mass. In fact, electrons usually travel far slower than that. The conduction speed of copper wire is in the range of a few thousand meters per second at best. Light (well, actually photons) travels faster than anything else. The speed of light is an absolute speed limit in our universe and is a fundamental physical constraint.
OK, I'm nitpicking and somewhat offtopic, but...
As someone else pointed out, the speed of propagation of an electrical signal has little to do with electrons moving. In normal applications, electrons move pretty slowly. The electromagnetic forces from them move much faster, IIRC they propagate at the same speed as light.
The Lorentz equations do not assert that a particle with finite rest mass would have infinite mass at the speed of light - as it approaches the speed of light, its mass approaches infinity, at the speed of light it is undefined (division by zero, and the universe dumps core), and above the speed of light it is a complex number. I believe the factor is 1/(sqrt(c^2-v^2), but I may well be wrong.
The speed of light is not an absolute speed limit. The speed of light in a vacuum, as far as we know, is. there are applications where electrons will travel through some medium at a speed faster than the speed of light in that same medium, although slower than the speed of light in a vacuum. There are media through which the speed of light is very slow. (I recall somewhere recently I saw an article about researchers who had found a medium through which light travelled at something like 38 m/s.) This is the source of what is known as Cherenkov radiation. (you can find a small blurb about this here).
Well, no one likes a nitpicker, but I hate to see someone so close to being right.
I agree that SMTP AUTH has a high coolness factor, but the one thing I can't seem to find anywhere on the sendmail site is a list of clients that support it. Netscape messenger? Pine? Outlook? Does anyone know?
It's so interesting to see other points of view. My dad started out teaching college psychology, working on his doctorate. He decided he was sick of it, and dropped out to work for Roto-Rooter, and was happy as a clam. Instead of going in each day to face students who didn't want to be there, and who certainly didn't want *him* to be there, he went in each day and worked with people who were *really glad* to see him. Then he started doing construction work, eventually became an inspector for the city, and liked it well enough, but didn't like having a boss. So he and my mom started up their own electrical contracting business, and he was happy as a clam again, because he was his own boss, and was pretty much doing something that he was good at and that he liked doing. Eventually, he and my mom split up, and she got the company (no hard feelings about it, it just had to go to someone) and he started working for another contractor - and loved it. He doesn't have to worry about how well the company is doing anymore, doesn't have to worry about much of anything, they just give him work, and he just does it.
Then there's me, I'm a network engineer, and this past year I paid more in taxes than my dad made, actually more than my parents made together in any year when I was growing up with them. But that's OK by them. My dad's doing what he likes to do, and making enough that he doesn't have to worry about money too much. I'm doing what I like to do, and will continue to do it no matter how much (or how little) they pay me, so long as I can pay my bills OK. I don't feel guilty for what I make, because I firmly believe that the worth of a thing is what the market will bear. Right now, that's a lot.
Strangely enough, the only thing I regret (job-wise) is that I'm not doing manual labor. I remember when I was in high school, I spent a summer laying concrete block. Hauling 12" block around, lugging 5 gallon buckets of mortar up 30' of scaffolding, taking a 15 lb. or so block and placing it *carefully* in place. I was in better shape than I've ever been in! And when I came home, I was tired. I could sleep. It used to seem wierd to me that people would pay thousands of dollars to go to a gym and, basically, do manual labor. But maybe it's not that wierd. But, I'll continue to do what I do, because I still have more fun exercising my mind than exercising my body. And who knows? Maybe if the tech market crashes hard, I'll go back to laying block. Endless possibilities....
Does it seem ironic to anyone else that the "crime" he is accused of basically boils down to figuring out the "passwords" (encryption keys) that DVD's use, and then when the police take him in for questioning, they not only seized his equipment, but they required him to give them his passwords?
Evil Evil Evil.
Can they learn from it? The post office (and ups and fedex and etc...) do some nice things. Package delivery is one. Message delivery is another. (although I haven't sent a piece of actual mail in three years, but I have received several) What I would really like is for the post office to start to learn lessons from the net.
I have an internet address. it is 209.116.217.40. It is also phoenix.hppc.com. If I change ISP's, my ip will change, but phoenix will not. We all take this for granted, and most of us know how DNS works, but it's pretty nifty nonetheless.
I also have a postal address. If I move, it changes. I can do the whole change-of-address form, and tell people my new address, and so on, but in the end, chances are I'll lose some mail.
Wouldn't it be nice if I just had a logical address that mapped to a physical one? If I move, I just contact the authority for my address, and let them know that my physical address has changed. My logical address continues to work. And even better, if I want to give someone my address so that they can send me mail, I don't have to let them know where I live.
The USPS seems like a logical place to keep this database of logical to physical mappings. Does anyone else think this is a good idea?
I'm also an MCSE and Linux advocate. I'd have to say that you don't have many choices, and my first reaction is that if you don't like something about a package at first blush, chances are you'll find other things you hate about it deeper in. If you have problems like this, conversion to some other package might be your best option.
:)
the next question is, do you have to upgrade? The version you have works, on the system you have. If when your vendor calls you and tells you that they've got their new version out and you should upgrade, just say "Sorry, but your upgrade won't run on our servers. As soon as it does, we'll look at it." Who knows, they may even pay attention to you. (but probably not. The company I work for is one of two dozen customers using the accounting package we have. Even though we represent about %5 of their market, they don't listen to us.)
But, in the end, I will have to say that adding an NT/MSSQL server shouldn't be too big of a problem. The biggest problem with most MS systems is admins and users mucking around with things they don't understand. If you just install NT, and install MSSQL, and pretty much leave the server alone after that, chances are it'll do OK. Microsoft makes the occasional great product, a few "fair to middlin'" ones, and many that really suck. NT Server and MSSQL seem to me to fall into that middle category. If you have reasonably competent people who know enough to run some other SQL server and are generally fairly careful, chances are you won't have any problems. The biggest headache I suspect you'll run into is figuring out how you want to do user authentication. Most accounting packages deal with this poorly. Very poorly. But now and then there's one that does, and if you want to be really slick you could set things up such that your users are logging into sql via integrated security (i.e. it authenticates against the NT OS password), in which case you may want to put your PDC on one of your linux boxen.
But that's just a suggestion. Personally I've always had a lot of fun with integrating Linux and NT networks. But then again, I'm kind of a sick and twisted person that way.
Finally, none of your choices is all that bad. If it were me, I'd ask the vendor to make changes, and if they ignore you, ask your bosses to try and find a better vendor, and if that doesn't work, oh well, there probably won't be too many problems (and you'll get something new to put on your resume)
Good luck!
Personally, I use First Union, and they've just recently gotten their heads out of their butts and become pretty decent in their online offerings. (It used to be that their online bill payment system used a netscape plugin that only worked on 95/98, not even NT, certainly not *nix).
The system they have now uses a pure web-based system (currently supplied by checkfree, but I've talked to them and they're building their own improved interface) for bill payment, bill presentment, etc. The interface they have is fairly nice. It's cool to be able to tell it "Ok, every month pull down my phone bill and if it's not over X dollars, pay it automagically." They have a nice system for statement download for checking, savings, and credit cards. They have a section for brokerage services, but I don't use it so I can't comment on it. Personally, I'm happy with them. If for no other reason than when I call their tech support, I usually get someone clueful.
They have demos set up of most of their functionality. Browse through it and see if it looks like what you want.
Strange, I go exactly backwards of that...
;lkj;lkj;lkj;lkj;lkj
The other way seems somehow... wrong.
Well, it may not do much good, but by registering their domain, they do give up some privacy. whois lists their billing and administrative contact as admin@ETOYS.COM, and their technical contact (which would probably be the least helpful) as hostmaster@ETOYS.COM. One of these would probably be able to answer questions as to who to contact for "customer service", comments &cetera and their legal department.
;)
Or, you could always send them a letter.
etoys (ETOYS3-DOM)
3100 Ocean Park Blvd., Suite 300
Santa Monica, CA 90405
US
They might even care. (but probably not).
Or start a rumor. That's the address that Santa lives at. spread it among the children.
Notice that the article wasn't a post from the webmaster, but from the NetApp company (which makes sense, it's good press for them and their product).
I also note that, contrary to the original post, they don't mention Linux anywhere.
I do question their choice of OS though. I love Solaris, but everyone I know who has run Solaris x86 has been disappointed by it. For sheer processing horsepower, and the ability to move a lot of data across a system bus, you just can't beat a SPARC. (although, there was mention of a linux port to the S/390, that might do it...)
But anyway, my overall impression of the setup is: Rox!
Don't forget Analog Science Fiction / Science Fact. An excellent magazine.
The software is still out there, and will probably run under dosemu just fine. Of course you won't get any added functionality beyond what the software originally had (well, things like hard links and nfs, but that's not really the application itself...)
Of course I've never tried it... and I don't even know how some things would work... I mean, WP made extensive use of the alt-F? key combos. How would you switch virtual terminals? I don't know, but it's a thought...
Well, the original problem seems to be that pricing updates come in mdb format. Although I can certainly agree with the original poster that I certainly wouldn't want to put an e-commerce server on NT, it seems to me that one NT workstation with a copy of MS Access shouldn't be too much of a problem. You create an mdb with linked tables, one linked to the file location where you will put your pricing updates, and one linked through the MySQL ODBC drivers to the MySQL (or PostgresSQL or whatever) databases. Then a little AutoExec macro, and you've got yourself an "update application".
The problem with this idea is that it doesn't actually enlarge the effective keyspace. It could be effective against someone trying to attack *your* entropy, but becomes less so when someone is trying to attack *everyone's* entropy. To put another way, you can do a dictonary attack against a password, and that may be effective. With a workload scheme, you feed your dictionary into the workload function *once* and create a new dictionary, which may be a bit larger than your original, but has the same odds of success.
The more people who use the same function, the more economical it is to do this.
Yes, trademark erosion is a bad thing for the companies whose trademarks get eroded. If your boss tells you go order a Xerox machine, are you going to make sure that it is from Xerox corporation? Or are you going to get a copy machine, perhaps the one with the best price or niftiest features, but maybe made by some other company? Same with Kleenex, Band-Aids, or, for that matter, Unix. People are buying products that they are calling by their brand name, but the company who makes that brand is getting nothing for it. That's bad for the company, and in many cases bad for the consumer. If on the other hand (let's see if I can think of a brand that hasn't eroded yet...) your boss tells you to get an APC Smart-UPS UPS, you will probably not just get a UPS, but one made by the company that produced the supposedly high quality products, and definitely high quality marketing that made the term "Smart-UPS" stick in his head.
The interesting thing is that most of the time, trademark dilution comes from a companies inability to get people to refer to their products as "{brand} of {item}". It would be much more straightforward and not brand dilution if you were to refer to "Band-Aid bandages" or "Xerox copiers", but when a brand goes from being an adjective to being a noun, the company is screwed.
IDG has some interesting issues here. Their product isn't getting "nouned" to death, but their trademark is certainly getting diluted. Unfortunately, there's not much a company can do about trademark dilution, because it mostly happens in people's minds. Eventually, they start getting overzealous, until there's nothing left of their trademark, and they eventually give up.
There is room for debate on this one. If someone other than IDG published a book called "Sendmail for dummies", it would clearly be a trademark violation. If someone created a website with that title (sendmailfordummies.com?) I would consider it trademark infringement, others would not. In this case, it's the subject of an e-mail, and I wouldn't consider that infringement. Others would. The point is, it's debatable, as we're proving by debating it right here and now.
I wonder if Hormel ever tried to enforce their trademark on Spam. That seems to be the most diluted trademark around these days....
(For those of you who don't know, at one point Spam was a meat food product, not junk e-mail.)
At one point I had read all that I could find of Card, but it seems he's writing them faster than I can read them. :)
Anyway, if you read "Maps in a Mirror", it contains the "seeds" for most of his novels. (it also contains some extrordinarily good short fiction)
I definitely loved both Enders Game and Speaker for the Dead, but some of his lesser known novels are just as good: The Worthing Saga, Treason, Wyrms, and Hart's Hope, to name a few.
And I have to admit, he also did an incredible job on the novelization of a great movie: The Abyss. I thought the novel was as good as the movie, which is difficult to do when going across media. I have high hopes for an Ender's Game movie. I can't wait to see it.
Why do people on this site always have to relate something posted on Slashdot to Geeks? I don't love reading Enders Game because its a book on Geeks. I read it because its a damn fine read. Why do you people (aka Geeks) have to try to justify yourselves? I'm confused.
That would be "F_ing Good Read", (working title), or "Damned Fine Novel" (release title). Get it straight. (Those who have "Maps in a Mirror: The Short Fiction of Orson Scott Card", and have thus read the short story "Damn Fine Novel", will understand... )
the problem with these is that they don't really understand ext2fs all that well, so it's doing, from what I understand, the same job as if you used dd if=/cdrom/image of=/dev/hda1.
This is fine so long as you've got a hard drive image that'll fit on a cd. But with a 650MB cdrom and a 18GB hard drive... I don't think it will work all that well. (note: I haven't tried this myself, it's just my understanding of how ghost works)
Basically, you know that whatever solution you use should probably use linux, as it should understand ext2fs, but once you get past there, there's a lot that you can do. dd an image from an nfs server, use kickstart, or for that matter, use tar. The biggest problem that I see is partitioning the drives in the first place. One of the nicest things about ghost is their gdisk program. (a completely command line scriptable replacement for fdisk. extremely customizable and very nice. If anyone knows of a similar tool under linux, let me know.)
-joe
OK, here are some thoughts:
.config, make oldconfig again, and answer the questions the way you think they should be answered.
n ux-2.2.12.tar.bz2
/usr/src/linux as a link to /usr/src/linux-{ver}
/usr/src/linux /usr/src/linux-`uname -r {+1}` #this doesn't work, I just forgot how to add. /usr/src/linux-`uname -r {+1}` /usr/src/linux /usr/src/linux-`uname -r` /usr/src/oldlinux
/usr/src/linux/.config . /usr/src/oldlinux/.config /usr/src/linux k e-oldconfig
/boot/vmlinux.new /boot/vmlinux.old /usr/src/linux/arch/i386/boot/compressed/vmlinux /boot/vmlinux.new
/dev/sda /vmlinuz.new /dev/sda1 /vmlinuz.old /dev/sda1 /vmlinuz.reallystable /dev/sda1
make oldconfig does ask questions, however it should be pretty easy to write a small expect script to run make oldconfig for you, and answer yes to everything. I'm not suggesting that this will make a more stable kernel, but it will make one. Chances are you could even have the script call up the help on each option, and have that as output from the program, which would in turn get mailed back to you by cron. In the morning check and see if the answers made sense, if they did reboot, if they didn't recopy
Now, as for the script itself... I have a few (lame) suggestions:
#!/bin/sh
cd ~/
# replace this with the correct URL
wget http://www.XX.kernel.org/pub/linux/kernel/v2.2/li
#I don't know about wget. It would be nice to be able to have it first look for
#LATEST-IS-* and then get linux-(whatever).tar.bz2
#or even better, have your script add one to uname -r and try
#to get that file. If it doesn't exist, exit.
# bzcat linux-2.2.12.tar.bz2 | tar xv
# rm -f linux-2.2.12.tar.bz2
# cd ~/linux
#I would suggest as an alternative keeping
rm
mkdir
ln -s
rm oldlinux
ln -s
bzcat linux-{version}.tar.bz2 | tar x
# will this work?
# not quite
# cp
cp
/usr/local/bin/some-expect-script-that-calls-ma
make dep
make bzImage
make modules
make modules_install
# and then
cp
cp
lilo
Then you have a lilo.conf that looks like this:
# LILO configuration file
# generated by 'liloconfig'
#
# Start LILO global section
boot =
append = "ether=0,0,eth1"
#compact # faster, but won't work on all systems.
delay = 50
vga = normal # force sane state
# ramdisk = 0 # paranoia setting
# Linux bootable partition config begins
image =
root =
label = lin
read-only # Non-UMSDOS filesystems should be mounted read-only for checking
# Linux bootable partition config ends
# Linux bootable partition config begins
image =
root =
label = backup
read-only # Non-UMSDOS filesystems should be mounted read-only for checking
# Linux bootable partition config ends
# Linux bootable partition config begins
image =
root =
label = original
read-only # Non-UMSDOS filesystems should be mounted read-only for checking
# Linux bootable partition config ends
# End LILO global section
horizon:~#
have cron run this every night, and every once in a while you'll wake up and find e-mail telling you that you've got a new kernel. If everything looks OK, just reboot and it's active.
Personally, I haven't bought a stamp in approx. 3 years. I don't even remember how to send snail mail. Everyone I care about communicating with has an e-mail address. (and if I really wanted to get wierd, my pager can send "e-mail" to a phone number through a text-to-speech engine...) the thing is, I don't use encryption and digital sigs all that much, and it doesn't matter too much to me. Yes, I trust them much more than an ink signature. But that doesn't mean that I trust it all that much. Digital sigs I don't need at all, because I generally don't care who an e-mail came from. If it came from my dad, I can generally tell whether it's signed or not. The actual applications that I use these tools for are few and far between. I would prefer it if everyone used encryption, but things are getting better. There's a reasonable level of interworking between the mailers that I and my friends use (outlook, netscape, pgp-elm and eudora) I've had very few problems.
I guess my general answer is that it's happening already, and encryption is not an obstacle. Encryption is more secure than any snail mail you can send, and easier than your average certified mail.
There are two problems that I see:
First, as the earlier poster mentioned, last time I checked linux assigns /dev/sda to the first scsi device it finds, /dev/sdb to the second, etc. I've never seen what linux does when, for instance, scsi id 5=/dev/sda and you add a device with id 3 and rescan the bus. Personally, I'd love to know what happens. It would be nice if there could be a switch from the way things are to a system more like (as I recall) Solaris uses which is not sequential, but definitive, i.e. incorporating into the device name the system bus/adapter/adapter bus/scsi ID/LUN. Of course that would be a pain for all of us that are used to the old way.
The second problem with the Vicom's is that you can only configure them from an NT box. In general this isn't too much of a problem, since you only set it up once, but if you really hate NT, it may be a problem.
They've got some interesting stuff. They are here.