You forgot the -H option to grep which comes in handy if you want to know WHICH file it found that string in. So:
find / -type f -exec grep -H somestring {} \;
The thing is that you know this for one of two reasons... either A) someone clued you in to doing this, or B) You read the manpage for find and found out that it has an -exec option which causes it to not print out the filenames but instead run whatever command you specify replacing {} with the filename and ending with a ; (which you then need to read the shell manpage to find out that you need to escape with a \ or do ';'). Then you remembered that grep is used to search text within a file given the text to search and the name of the file. However you try it without any other options and realize that you can't see which filename it is showing you the content of, so you man grep for info and find out that you can give it the -H option). Or maybe instead you don't know of this so you write a small script which runs grep and if it succeeds then prints the filename after it.
Option B is the true *nix users way of doing things and is what makes *nix so powerful. Option A is something you can tell a normal user to do if they need to search for text in all files in a directory (in this case/). Of course if the person is truly clued-in they will discover option B for other things and eventually realize the true power of the *nix CLI. At that point they will become a true believer and realize that all other systems are shit compared with this sort of power.
Do you really think that so many people would advocate Linux if they didn't think it was a better system? Conversely look at the people who advocate Windows. They too think that Windows is a better system and therefore advocate Windows. Although from my point of view I see their logic as being mostly flawed. I.e. arguments like "But Windows is so much easier!" don't agree with me. Windows is generally easier to learn because of it's GUI interface to everything. Linux with a GUI like that could easily match Windows ease of use and have more power for those who want to take advantage of it. The only drawback is that yes, it is unfamiliar territory. OTOH if you spend about 5 minutes and let a person know that it's still a GUI (you could even use a term such as windows-like) they may clue-in to the fact that if the play around with it for a second it won't be that hard to figure out how to do basic tasks. And since they already have experience with a GUI from Windows they will realize it's really not all that different.
The shitty thing is that you will still have those occasional few who have bought-in to the UNIX is old and outdated argument and don't realize that just about everything good about Windows comes from UNIX with the exception of the GUI which Billy-boy stole from the Mac and did a piss-poor job of stealing it. I.e. if UNIX is so bad why is the Windows TCP/IP stack based on the BSD TCP/IP code. Why does Win2K use kerberos for its domain logon/resource access functionality? Why does MS make a huge press release when they finally get around to implementing something that UNIX had 20 years ago and call it a brand new technology (e.g. symlinks anyone?)? Or here's another one: the ability to mount a filesystem in a directory of another filesystem instead of on it's own drive letter? It's all in the image. Some people really and truly believe all the marketing that MS spews and don't bother to learn anything else. Some people don't know the specifics of working with UNIX, but do know that MS is full of shit. And some people know the specifics of UNIX, know that MS is full of shit, and wonder why everyone else can't figure it out.:-)
Well, I like the premise of the article. I, as a POSIX person, fully understand this. In fact, I understand it so well that all I can say is "well DUH!" However, I think the author makes a very glaring assumption in the very beginning that a UNIX-like system is more stable than a Windows system. People who work with UNIX know this. People who work with Windows are of the mentality that it is stable enough. And I finally have caught the premise of the article that stable-enough really isn't. That the real thing is that the stability problems/do/ cost money. That the fact that your users don't trust their computers to work properly costs money.
What UNIX people don't realize is that most people don't understand this. And unfortunately this article sounded too much like pro-UNIX anti-Windows for any Windows person to read it. They would have stopped in the first few paragraphs saying how dare he say that Windows isn't stable. Hell, all computers make mistakes.
See, right now I am taking a full MCSE program from a local place where I live. Most of the people in my class think like this. The whole, well, computers have problems mentality. These people know nothing of UNIX. They know nothing of administration period.. with the exception of me and one other guy who also gets it w/ respect to how to be a good admin in general (he is/was a netware admin and his company is paying for him to do this). The only thing that all of these people know is how to get around a Windows 9x/Me system. Some of them were Gateway employees who did tech support or built systems.
Also, a lot of these people repeatedly utter the phrase, "Is this going to be on the test". In other words they see this as drop $10,000+ and get a full MCSE training and cert and go out and make lots of dough. Hell, it's advertised that way. What they don't realize is that you cannot be an effective administrator unless you know how the system works. And whats even more heart-sickening is that the MOC (MS Official Curriculum) actually does go into pretty decent detail about the how's and why's of Windows but these people don't give a fuck.
In general I feel that the MOC is a very good guide on administering Win2K systems. It's very practical and also explains the basic concepts. But again, these people only care about passing the exams. In fact, the instructors usual response is that "No, but you'll need to know it for work". Occaisonally it is "No, don't worry about it too much, just keep it in your head" because honestly, some of the stuff is a little too "Gee Wiz, look at what we've done, see how good our software is that we have this and this and this and this feature".
Anyway, the bottom line is that if everyone actually followed what the MS texts say that Win2K would not be horrible. And I think that the author of this article should note that. I understand that he does in a round-about way say that he is assuming a professional IS team for both Windows and UNIX, and that even with a good IS team for Windows you still have the basic problems that the software sucks in the first place.
Anyway, at some point soon here I will have all this Win2K knowledge and can actually give a straight honest-to-god answer that YES, I/HAVE/ done Windows.
One thing I have noted is that Win2K's strength lies in the fact that MS has basically designed your network for you. I know this sounds silly, but MS really has done a decent job of this. This is why Win2K is supposedly easier to administrate. You don't have to put much thought into design, you already have default groups like backup operators and so on. You can easily delegate authority to other members of the company. If you know what you are doing and understand the security implications it is easy to setup the system.. hell, the point that you by default allow nothing to happen and you grant people permissions to view/modify things is even stressed. This is good network philosophy. I actually applaud MS for this setup.
However, at the same time, their software sucks. I could just as easily setup default groups and so on for kerberos with Red Hat and sell that as a product. It'd be a lot more stable too.
Already we are seeing peoples computers running Win2K Advanced Server have stupid things happen. We have already had to reinstall one dudes machine because it somehow got fucked up. No one knows, no one cares, in this case it was easier just to reinstall from the ghost image because there was nothing important on the machine.
Now, at the same time, it seems that most of the instructors there do have a clue. They may not know UNIX for lack of time, they may only have experience with it in some college courses. But at least they understand why people use it, and would learn it if they had the time. The bottom line is that they spend so much time fucking around with the Windows boxes that they don't have time to learn Linux:-(.
Anyway, that is just my 2 cents on this whole thing. BTW, there are also some people there who have bought into the whole UNIX is outdated and sucks and Windows is new and cool and is better. Only a few have deluded themselves that badly though. That is of the people working there. In the class no one really has experience with anything so they don't really care.
Problem is that apparently SPA in Outlook is an MS specific thing. Well, what do you want them to do. The only way for outlook to support not sending the login in cleartext is to use SPA.
What about RFC 1734 [faqs.org] and 2095 [faqs.org]?
Sorry, I should have been more clear, what I meant is that Outlook (Express) already had support for MS SPA and not the other standards, and therefore MSN's only choice was to implement SPA or upgrade everyone to a different client. They did exactly this.
Basically all that needs to be done is for other mail clients to support MS SPA
This is where the problem is. There are existing protocols to deal with secure POP/IMAP authenticaion, but MS goes ahead and writes their own, and then people say "why doesn't everyone just support the Microsoft format" -- it is this line of reasoning that has led to nearly every proprietary closed format/protocol.
If MS doesn't want to pass passwords in cleartext, they should be using one of the existing and open methods of encryption, not forcing other people to use their software.
You are so right, they should have chosen an open standard for Outlook in the first place. Although I am not sure how many were widely accepted when OE 4 (first version named OE and not MSIMN) was released. Fast forward to modern times and MS still has no incentive to implement another SPA protocol.
And really, how many mail clients do you know of that support secure password authentication? I know that Netscape and Mozilla do not. Fetchmail does support quite a lot of authentication protocols. Interestingly, it does suport NTLM authentication for use with Exchange. Maybe this is a similar method. Knowing Microsoft they wouldn't reinvent anything but just use some crappy system they already have (i.e. NTLM).
HOWEVER: MS could VERY SIMPLY allow SSL connections which most mail clients do support and which would encrypt everything including the mail from the server to you. And furthermore OE does support SSL connections without MS SPA. However, this would be a lot of load on the mail servers doing all of that encryption.
So I suppose note well anyone who wants to give MSN a valid option, that is it. However, believe me I am well aware of the fact that MSN does not give a flying fsck about their customers, so YMMV.
I am hoping this is a joke and not for real. If this is supposed to be for real, then these people at the EFF need to be beaten severely with a clue-stick.
There is absolutely no guarantee that mail goes through. E-mail is not government regulated, period. I agree with them that there should not be stupid laws that try to limit SPAM and yet do something else. Currently, the best solution is something like MAPS.
What is so wrong about MAPS? The concept is sound. If you are connecting to my mail server (i.e. you are essentially treading on my property) then I have every legal right to tell you what you can and cannot do. I can also choose to not let you use my property. This is what MAPS does, it says, "Sorry, you can't use my server." With MAPS when you have fixed your server, you will be allowed to use mine again.
The argument that a system administrator cannot change her mail server is totally bogus. That is like MS saying they couldn't possibly remove IE from Windows. If the sysadmins like to be used as open relays, then I can let them know that they won't be allowed to send to my server. When they fix their server and notify MAPS then they will again be allowed to use my server. The same goes true for servers in the RBL. If your server sends a lot of SPAM and I don't want to incur costs to process it then I will simply deny everything from you until you fix your problem.
The statement that users get no notification of this is also FALSE. Assuming you can receive bounce messages (i.e. you don't have a problem with your mail configuration) then you will receive notice that the mail was rejected. Unless of course the mail server silently accepts the mail then tosses it, but that is not the point of MAPS.
The Internet is not government controlled, it's not entirely corporation controlled. It is a community network.
I don't understand how denying me the right to use MAPS is a good thing? I am supposed to accept mail from anyone regardless? Am I also supposed to let every jackass who walks onto my property and starts fishing in the bay do so? Can't I tell them that I own this property and you are not allowed to use it, so go somewhere else? Isn't this a fundamental right as well?
I especially love the other comments that we should replace SMTP with Jabber or ICQ because it is impossible to make SMTP SPAM-free. That's true if you still believe that every SMTP machine is supposed to relay everywhere. The reality of it is that an SMTP machine should 1) Accept mail destined locally. 2) Accept mail destined for a machine that it is a backup for, and 3) accept mail from a local user destined for anywhere. What MAPS is trying to do is get everyone to abide by these very simple rules. Those who run open relays get put in the MAPS RSS. Those who are connecting through dial-up connections get put in the MAPS DUL. Those who are themselves sending the SPAM or are allowing their customers to send SPAM are put into the MAPS RBL. What is so hard to grasp about this.
Oh, wait, didn't peacefire get blocked because their hosting company was allowing spammers to use machines in the same netblock. Tough shit. Go get a different hosting company, you aren't required to use them. Why not find one a bit less friendly to spammers. This is called free-market economy. Your e-mail getting blocked because your outgoing server allows relaying or is allowing spammers to use it regularly. Go find a different outgoing server. Anyone can sign up for a free hotmail account if it is just absolutely necessary to send a message that very second. Big deal, so you have been incovenienced. No one has taken away your rights though.. You have NO right to send e-mail to everyone you want to, that is the bottom line. You have a priviledge to do so. PERIOD.
I recently got laid off at a leading teleservices corporation that did technical support for MSN. This is because they completely dropped the contract with MSN (for what reasons I have only heard speculation and will not repeat here). I can assure you though that it was not because our standards were not good. Although it sounds like I am tooting my own horn we had probably the best call-center for all MSN service judging by the number of people calling back with ticket numbers started by people in other centers. I also judge this by the way the people wrote up their tickets without specifying what in the hell they did forcing me to go back through all the troubleshootings steps. (end rant)
First of all, MSN has two types of mail. They have the "legacy" POP3 system and the new web-based e-mail. You can find this information at MSN Support Services.
Web-based mail is kind of like what it sounds. It uses the same mechanism (XML over HTTP) that Outlook Express >=5 uses to access hotmail. However the server for @msn.com accounts is different from the server for @hotmail.com accounts. If you have an @msn.com web-based account you can go to http://supportsevices.msn.com/us/oeconfig/ to configure OE and then go to tools accounts and read the server name out of there. Note, this is also true for free @msn.com accounts. Note that only Outlook Express 5 or greater or Outlook XP can use this mail. Obviously MSN Explorer and the hotmail.com website itself are compatible with this.
Anyway, it seems the real issue is that these people would like to use their new MSN POP3 accounts with e.g. fetchmail. To correctly configure Outlook Express for MSN POP3 e-mail you must use the outgoing (POP3) server of pop3.email.msn.com (go figure) or the incoming (SMTP) server of smtp.email.msn.com. Furthermore you must select the "Log on using Secure Password Authentication" option as well as select the option under Outgoing Mail server that "My server requires authentication". You then must press the settings button and be sure it is using the same settings as the outgoing mail server. That is it logs on using SPA with the same U/P as the POP3 server.
Because of this MSN states that you MUST use Outlook to get your MSN POP3 mail. This is not entirely correct. What you must have is a client that supports SPA. Why is MSN doing this? MSN's reason: to reduce SPAM. However they tell customers simply this because most of their customers are rather computer illeterate (especially the former AOL lusers). The real reason is that since they contract out Dial-up Points of Presences (Pops, not to be confused with POP3 e-mail) that either A) they must use the POP3 before SMTP hack, or B) You must login to the SMTP server to send mail. If they didn't do this then any jackass dialing into one of those POPs even with another ISP would be able to send tons of SPAM through MSN servers. There have been plenty of/. articles about this before and anyone familiar with how contracted out POPs interfere with the ability to allow SMTP access to only your subscribers should know what I am talking about.
Now, MSN/could/ have simply kept the plaintext login POP3 and only required you to use a plaintext login for SMTP. However they decided that not only should they require a login for SMTP but at the same time they should require secure password authentication for both POP3 and SMTP. In other words, if they were going to have to have people change their Outlook mail settings they might as well knock out the ability of people to sniff the packets and retrive their users passwords while they are at it.
Problem is that apparently SPA in Outlook is an MS specific thing. Well, what do you want them to do. The only way for outlook to support not sending the login in cleartext is to use SPA. So therefore they enabled SPA on their mailservers and disabled clear-text logins. Of course theoretically they could include some other more open method of secure password authentication for use with other clients, or they could open up the MS SPA protocol. Or they could just say the hell with it because they only officially support MSN using MS software on Windows OSes (which actually does NOT include WinCE, you must contact your OEM for WinCE support with MSN).
Basically all that needs to be done is for other mail clients to support MS SPA. How to do this I am not really sure as I have not put much thought into it as I don't use MSN myself except for free accounts. All the free accounts use hotmail based e-mail.
There is of course another option. You could always "upgrade" your account to web-based from POP3 and then either go to the hotmail website to get your e-mail or use Outlook Express >=5 or Outlook XP to get your email in a real mail client (if you can call Outlook a real mail client, but hey, at least's it better than www.hotmail.com). There does exist a script (PERL I think) for retrieving mail from hotmail but I have looked at that code and it is really really crappy (apologies to the guy who wrote it, but I am sure he also knows that it is nothing more than a quick hack). Theoretically there is no reason that Evolution should not support the MS HTTPmail protocol. Turn on HTTP logging in the Advanced tab of OE properties and then open up the log in notepad. You will notice that the schema is relatively easy to figure out even though to the best of my knowledge it is not published anywhere. Evolution already uses XML extensively and has all of the framework necessary for parsing XML. I assume it also has the framework necessary for accessing an HTTP server in general. Therefore it should be rather trivial to write an MS HTTPmail backend for Evolution. In fact, I am surprised that no one has done so (I guess none of the developers use hotmail). I have toyed with the idea of doing one myself but 1) I use balsa, and 2) I have not done any programming with XML. However now that evolution is fairly stable I may go ahead and write this. Hell, I don't have anything pressing to do until Monday except clean the garage so we'll see. There's never a bad time to learn more programming techniques, and XML is one of the most popular things today so not only would I personally benefit from learning XML but also benefit with being able to access hotmail from evolution. And note well... if I do write this I do intend to support the advertisement properties as best as possible (i.e. opening up a small frame at the bottom and displaying a webpage in it). I know it seems stupid, but hey, they deserve to get paid even if they are MSN. And if anyone really wants to they can just change the code later to take out the ads.
Anyway, I hope this clears up a lot of the confusion people are having with this. I see at this point over 600 comments have been posted, a few reasonable, most along the lines of fsck Microsoft. People, I hate MS as much as everyone here. They are theives and crooks and must be beaten. However, as the cliche says: You catch more flies with honey than with vinegar. The only way MS will be beaten is when people stop bitching about them and just go do better than them. Every time I bitch about MS to my mother she reminds me: Then go write something better. While everyone has argued this point to death the bottom line is that in some respects MS software is "better" than open-source/free software. Even if only in the marketing sense of better.
On my server I have a shell script run by the webserver which echos the REMOTE_ADDR variable along with the current time into/tmp/cripfifo (Code Red IP FIFO). From there I have a script that I run as another user that does and endless loop of cat/tmp/cripfifo | while read i which then loops and for each line it inputs (which is actually probably only 1 since the fifo is closed each time the webserver script writes into it). With each line of input it first checks to see if there is a file named $REMOTE_ADDR* in the errorips directory. If not it runs another script in the background which uses wget redirecting output into errorips/${REMOTE_ADDR}_TIME. If it succeeds it moves that file into the processed/ directory which means that if the machine attacks again it will repeat this processes each time it attacks.
Yeah, I know there are some race conditions if the same IP gets processed before the other one has had a chance to write the logfile. It's not a big deal though, and the chances of that happening are slim anyway unless someone is just bombarding you with attacks in which case it still doesn't matter.
See http://kernelrpm.sourceforge.net/codered/
for the scripts. Note, it may be a better idea to use the Location directive for/default.ida and point it to the script rather than making the whole damn directory ExecCGI. Although I don't have a problem doing that since only I have access to it anyway.
So, just to be clear this isn't "another gay-ass Ask Slashdot that could've been answered in 5 seconds if the poster had figured out how to use Google".
Obviously you think this or you wouldn't be asking the question. However, this isn't far from the truth:-(. First of all, the program you probably want to use is mgetty which is included with Red Hat. In 7.1 there are four mgetty packages on the second disk. You definitely want mgetty installed, and probably mgetty-voice if you intend to use this as an answering machine type system. Now, I have browsed through the documentation directories for mgetty and I do agree it is pretty lame. In addition trying to search for mgetty is not very helpful. The first result returned from google is http://www.leo.org/~doering/mgetty/ which seems to be about the best reference you will find (so google seems to be doing its job). However the distinctive ring is buried in the caller-ID section and I am not exactly sure how to go about getting this to work. My advice is to read the docs thoroughly and also look for comments in the config files. There is also a mention in an mgetty FAQ on faqs.org of a mailing list for mgetty. You will probably get a better response on a mailing list for the software than on a general forum such as slahsdot.
A few things you should know is how to put your modem into distinctive ring mode. You will probably need to change an S register so that the modem will report RING 1, RING 2, RING 3, etc. instead of just plain RING. Use minicom to test this out. You may be able to do an AT$ or ATS$ to get lists of commands (works on USR modems, and I think also some Winmodems have something similar, maybe just the USR/3com ones.
I hope that helps you out in your quest. If nothing else start viewing the mgetty source and if necessary patch it to better support distinctive ring. Or maybe it has good support but is just poorly documented, so document it so the next guy can figure it out.
From the looks of your post you are just another troll. You say you find Linux to be lacking and that in all the key areas of business it is clearly inferior to Windows 2000. However, you don't actually go on to explain why Windows 2000 is so superior to Linux in key areas of business. The most you do is say that Linux is obscure, hard to use, and inconsistant. Did you read that off of a Microsoft website or out of a ZD net magazine?
For the purposes of the rest of this comment I am going to assume that you do in fact have reasons you have just neglected to mention them.
You mention that you make your living providing solutions based on Microsoft technology-- good for you. I am sure you have quite a bit of knowledge relating to Microsoft products. However, I am also fairly certain that you have not even begun to learn UNIX in the same way that you have learned Microsoft technology. If you actually took some time to evaluate Linux and brought in a real Linux guru, you may find that for many applications Linux is clearly a better choice. Likewise for many applications Windows is a better choice.
You mention that if businesses were not making money from their MS systems they wouldn't stick with them. That is entirely true. If the system is in place and meeting the needs of the business it would be absolutely idiotic to switch to Linux. However, if the system is not meeting all the needs of the business then it is prudent to evaluate all options before proceeding with an information system upgrade. An entry level systems analysis and design class will tell you that. Using another technology may be a viable option, and is at least worthy of being included as one of the proposed solutions.
You also mention that if Linux offered any competitive advantage that businesses would install redhat straight away. This is also not correct for the same reasons as above. Haven't you ever heard the cliche "if it ain't broke, don't fix it?" This is especially true for information systems. If it would cost more money to migrate to Linux than would be saved by migrating to Linux then it is not worth it. That is basic business practice. You can also find that out in any entry level systems analysis/design class or in a more general form in any entry level business course.
You say Linux is obscure, hard to use, and inconsistant. Have you taken a look at Windows lately? Or even MacOS for that matter. Every system is somewhat inconsistant, hard to use, and obscure. Those are pretty vague adjectives, try to come up with some specifics.
In the past Microsoft has pushed out several competitors. You mention OS/2 and Amiga. For one thing, Amiga is a different hardware platform entirely and was never really considered to be a serious business machine. Most businesses used IBM computers and turned to IBM compatible PCs when they became available, so I can't see how Microsoft kicked Amiga's ass. Microsoft did kick OS/2's ass pretty hard. However, OS/2 had some drawbacks as well. Version 1.x was IBMs baby although MS did much of the coding and design. IBMs original purpose with OS/2 was to provide a better operating environment to entice consumers to switch to IBM computers running OS/2. Version 1.x had such shitty hardware support for anything not IBM that almost no one could use it. During that time Windows was not even anywhere near usable. It wasn't until MS decided to split from the project (and go work on Windows 3.0 which could now run in protected mode) that IBM decided to compete on the software level.
At that point Microsoft launched a massive marketing campaign to put Windows on every desktop and have every application written for Windows. They also had Windows preinstalled on every new PC and penalized manufacturers who installed other operating systems. Of course no one used OS/2. Why spend another $100+ when you already have Windows on the machine? It's a wonder OS/2 lasted as long as it did as a desktop OS. However, you may note that OS/2 is still used for many embedded devices and that a new version has just been released renamed eComStation.
In your final paragraph you mention that geeks who want to succeed in the information economy should study some business courses. That's extremely good advice. You also say that their time would be better spent getting an MCSE instead of rebuilding the kernel every 2 minutes. That's true too. If you want a professional quality Linux certification, they are available.
First of all, IE is not my main browser. For this comment I am using Mozilla. However it just so happened that I was testing out IE on a newly compiled CVS build so I figured I'd write the comment.
I would have to say that I am not very thrilled with VMware from my personal experiences. IE running under Wine runs faster than Mozilla. IE running under VMware runs slow as shit. IE under Win4Lin is a bit slower too (but not very noticably so). Using Wine as a Win4Lin/VMware type thing is not such a bad idea. Sure it is not quite as compatible at the system level, but at the application level most things are working. Like I said, one of the major drawbacks to Wine is the USER/GDI code. With working core libraries in place WINE would provide a better solution than Win4Lin or VMware. Especially if a company (CodeWeavers would be the first Wine related company to come to mind) did some more work and packaged Wine as a Windows applications runner. However, I don't think you'll see them doing that, it doesn't seem to fit in to their business model from what I gather. However, Jeremy White and friends did do a very good job of packaging the Codeweavers releases of Wine.
About the license thing. Most (possibly all) of the DLLs I used native versions of actually came with IE because it updates the system. So theoretically if you can get the installer to work then you shouldn't need Windows at all.
Yes, that's right. With the latest CVS updates the painting is better too! It now doesn't white out the toolbars when you load a new page. The toolbars still start out white sometimes, but nothing a quick rollup and rolldown can't fix (which used to not be true either).
My IE 5.5 is installed on a copy of Win95 OSR2 with updates (including Y2K) under Win4Lin. Unfortunately, since Win4Lin doesn't work with kernel 2.4 yet I am unable to use it, but it makes a nice installed Windows tree with native DLLs for Wine to use. Currently I am running with native commctrl,shell, and ole.
There are still some issues with the user code in Wine. It's old. Recently Alexandre Juliard has submitted some changes to better seperate the Wine graphics driver out of the user code. I believe Alexandre is looking to do some major improvement to user. Well, at least I hope so, the current user code is very old and crufty and is the cause of lots of painting issues. Even solitaire doesn't exactly paint right at the moment (move a window around over top of it and you'll see what I mean).
Wine has come a long way and if you haven't had the opportunity to use it lately, I suggest that you try it again. IE seems to be working just fine for posting this.
So how exactly are you supposed to install this other RPM on a machine? Does it contain the source code in pristine form with patches and then must be compiled? It would seem that that would be the only way to satisfy the conditions of the license. If that is true then what the hell do you do when your server doesn't have a compiler (which it shouldn't).
I suppose now you will pull the argument that qmail is uber secure and thus you don't need the rest of the system locked down. That is total BS.
Red Hat makes a binary distribution using open source code and their own patches to better integrate components. The qmail license almost directly contradicts with the purpose of a binary linux distribution. It also contradicts the GPL.
It is entirely true that Red Hat is not allowed to include a modified binary in their distribution. Sure they could include some dumb hack to get around the license, but why should they.
You fucking people need to get it through to your small ass frontal lobes that businesses need to make money. Red Hat provides for their customers a prepackaged linux distro and automated package updates. Red Hat is not obligated to include every package under the sun. They include what they see as useful to their customers. If you want qmail on Red Hat then fucking grab the source and install it. Damn!
Let me get this straight, you admit that you can't distribute modified binaries and must distribute patches instead of modified source. Before this you say that these programs are free in the GNU sense of the word. I am sure Stallman would love to rip you a new asshole for that comment. I cannot believe that your shit post got modded up to 5.
And about this whole RedHat shipping Mozilla instead of Netscape: years ago when they included the Netscape binary only package it was said that part of the motivation was that netscape released most of the code. The original intention was always to switch to Mozilla from day 1 of including Netscape. This article is really old news, although I am sure most people have forgotten that by now.
Have fun trolling. I certainly hope your message comes up for meta-mod.
Since when did the term hacker become a catch all for criminals who use a computer? I am absolutely appalled by this. While it is true that words should and will naturally evolve in a language, this is really not the case here.
It seems to me that the author or editor decided to use the word hacker in the title for some sensationalism. This story does not even describe what an average citizen would call a hacker. Most people think of a hacker as someone who infiltrates a computer system. Most people also think that hackers steal information and damage property (e.g. delete or destroy data).
In all reality, a hacker is someone who is extremely skilled with computers. A hacker also has knowledge about how to break into a system. For instance, any system administrator worth a damn is a hacker. That is to say that they are skilled with computers and know how to break into the system. This puts them in a position to keep computer criminals (note the word choice, criminals not hackers) from breaking into the system.
In this article, the computer criminals are just that, criminals. They are simply modern-day scam artists. But hey, "Watch a computer scam-artist work the system" probably doesn't have that same ring to it.
Originally I was going to send an e-mail to Bob Sullivan (the author) but if you do a search for the word hacker on the page you will not find it in his article, only in the headline and in links to other MSNBC headlines. Apparently this is just some editor with his head way up his ass.
No sophomore CS major is going to spend time researching the tax code of Indiana.
Well, if he's an Indiana resident, he most likely will. Likewise so would residents of other states. You could theoretically make a sourceforge project out of this if you could somehow garner enough interest from many people.
Best thing to do would probably have a core team work out the basics of a tax program and then have other teams make modules for the different states. Probably in some scripting language (python would be a good choice given good GTK bindings and support as a scripting language in GNOME).
Trying to compare RPM to apt-get is illogical. Debian's apt-get takes a package name and goes through your list of servers and looks for the package. The format of these packages is.deb. Apt will automatically get the.deb you requested plus any.debs it depends on.
Most other distros have standardized on the RPM package format-- possibly because a lot of them were originally based on Red Hat at some point in their life. RPM is a package manager. It serves the same purpose as the debian package manager (which is not apt). Unfortunately, there has not been an equivilant for Apt on RPM-based distros.
However, recently Connectiva or someone modified Apt to work with RPMs as well as DEBs, and if you read some of the other comments here supposedly it works pretty well. Also, Ximian (formerly Helix-Code) has come up with the new Red Carpet Updater. Red Carpet also serves the same purpose as apt. Unfortunately, Red Carpet is completely GUI from what I have seen and there is no command line equivilant. This kind of bothers me as it would be nice to use parts of it in scripts. On the other hand, it is free software so anyone could make some good command-line tools to go with it.
I have been using Red Carpet 0.9 since its release and I really like it. A big warning though: It does not like unsatisfied dependencies at all. In fact, I had NVIDIA_GLX (which depends on NVIDIA_kernel) installed but had no NVIDIA_kernel package because I rolled my own kernel and named it and the NVIDIA driver differently than usual. Well, because of this unsatisfied dependency (that is, NVIDIA_GLX depends on NVIDIA_kernel, but there was no packaged named NVIDIA_kernel on the system) it wanted to removed NVIDIA_GLX. And for some stupid reason it decided that it also wanted to removed all the XFree86 packages and other stuff. I of course declined and then eventually traced it back to having no NVIDIA_kernel package.
Basically I would say that Red Carpet is a bit to stringent about dependencies. On the other hand, this is designed to ensure the system stays stable. The bad thing is that I could just see someone following its advice (which I chose not to) and having it remove XFree86 and a bunch of stuff and then being unable to run Red Carpet to get it all back because it requires X. Whoops, must have forgetten a dependency there!
I suppose that is more than you wanted to know, but if you have an RPM based distro you should really check out Ximian's Red Carpet. And Petrely needs to get his damn head on straight and realize that RPM is comparable to deb and apt-get is something different. At this point with Ximian Red Carpet, there is basically now apt-get for everyone so this whole article is a moot point.
And oh, as another poster said, I rarely hear of any RPM users screaming that they want to switch to Debian. Personally I like RPM and from what I have seen of deb I see no reason to switch. The only good reason to switch to debian seems to be apt, but now that one company has made apt work with RPM and another has written a new program that works from the start with both deb and rpm I really don't see any reason to switch.
The main thing about @Home is that not all @Home is created equally. The cable company has a lot of freedom to do different things. In Indiana (where I live, with AT&T@Home) part of your monthly fee pays for the modem and you can get a $5 discount or something by using your own modem. In Virginia (where my parents have now moved, with Cox@Home) you must pay for the modem. AT&T charges $5.95 (IIRC) for additional IPs with the max being 6 total or something. Cox charges $6.95 and does not allow more than 4 total computers. In addition they charged a $10 "setup fee" for the IP. So I basically paid them $10 so I can continue paying them $6.95 every month. What the "setup fee" is for I don't know. I know it is an automated process to add the hostname and IP to the DNS server since it works immediately. However, DNS lookups did not immediately return the address and a reverse lookup on the IP returned someone elses hostname (presumably someone who had cancled service). It does now however work correctly (it's been a couple days). Its possible that they actually have someone update the DNS server manually which would explain the delay, or maybe they just don't restart it for a while.
Currently AT&T (and Cox from what I am seeing on my dads computer) use static addresses but also support DHCP. Most installers in my area of Indiana always type in static because the DHCP server tended to go down constantly. Ask a low-level tech from @Home about static and they fucking freak out because "we will be switching to dynamic addressing in the near future". Says the same thing when you order additional IPs that you need to set your hostname and use DHCP and that its bad to configure them statically. In all reality there are so many customers with static addresses that it would be an absolute support nightmare to switch. In addition, I have heard that AT&T could give a shit about static but that it is @Home that wants to go dynamic. That is way through the grapevine but I believe it because @Home's DHCP servers go down so often in my area in particular that its cheaper for AT&T if everything is setup static because it is less support calls.
Should the configuration be wiped it will be necessary to type in the correct hostname and/or type in the IP number/netmask/gateway/dns servers anyway so there is no advantage to going with dynamic addressing UNLESS they would just randomly assign IPs to any computer that asked for one via DHCP. In that case the support costs would go way down because basically as long as the DHCP server works and the computer simply has a DHCP client it will get an address and work. The disadvantage is that your IP would then be truly dynamic and changeable which I don't like. I prefer to pay my $5 or whatever and have a real static IP. OTOH, many DSL providers do a hybrid of this system. That is, they will give out random IPs to any computer that asks for one but you may also purchase static IPs for $5 a pop. In my opinion, that is what I wish AT&T would do, it seems like the most fair system.
As for the other DHCP server (the one that gives the modem its address) I seriously doubt that it works any differently and most likely you couldn't just plug a new modem in and expect it to work.
Your modem has a MAC address (two actually). The MAC address identifies the modem for use with DHCP. The DHCP server gives the modem a 10.x.x.x address and it is online (there is a whole bunch of other stuff involved, but essentially that is how it works). The modem must be provisioned to some account to work. There may be a way to provision a modem to a non-existant account but it would have to be a very inside job at @Home (in Colorado most likely).
The modem essentially functions as a bridge between the ethernet interface and the RF interface. Although they are two different protocols on the physical level, they both speak TCP/IP so I believe it would be most correct to call it a bridge. The MAC address of the ethernet interface on the modem (which is like never used) is always one more than the MAC address of the RF interface. You would be most concerned about the MAC address of the ethernet interface if you were screwing with your ARP cache I suppose, otherwise you really don't need to care about it.
Now, assuming your modem is connected as described above, then you essentially have an ethernet network like any other. Your computer has a hostname usually beginning with the letter C (hence called the Charlie number by some techs). There is then a 5-7 digit (usually 6) number identifying the subscriber. There is then a dash and a letter identifying the computer. So something like c780677-a (which is actually my hostname). My second IP address that I ordered is c780677-b. Your DHCP client contacts the DHCP server, gives it the hostname, and gets an IP address (standard DHCP procedure). They do not use your ethernet cards MAC address like most other DHCP/BOOTP setups because you may want to change network cards.
So, assuming what you said was true, then the only thing you would be able to do is connect multiple computers to one cable modem without paying for additional IP addresses. Actually, that would be a pretty sweet deal. Grab another box, set it's hostname to the same as yours but with a -B and see what happens. If you get the address without paying for it, more power to ya.
On a RedHat 7.0 style system (also 6.2 IIRC), set the DHCP_HOSTNAME variable in the/etc/sysconfig/network-scripts/ifcfg-INTERFACE file to the desired hostname. In Windows you set the computers name to the desired hostname, also you may want to enable DNS and set the correct hostname and domain as well. For example, hostname "c780677-a", domain "[cityname]1.[2letterstateabbrev].home.com". Also set the workgroup to "@HOME".
I think that is probably everything you ever wanted to know about cable modems.
It is my understanding that Christmas is supposed to be about celebrating the birth of Jesus Christ. So what exactly is the issue of using the F-word in a Christmas card to some personal friends. If it was part of an inside joke then apparently these are at least fairly close friends or were at one time.
I don't know about you, but I generally say the F-word all of the time around my friends but try to avoid its use when around other people. Think about the irony in that.
Err... shouldn't that be "This is a situation with which I will not put up." You can improve it further by switching to the active voice and saying, "I will not put up with this situation."
The original poster was right, it should be "Chip news on which to crunch." In spite of that, most people would agree with timothy and say "Chip news to crunch on." But just because most midwestern americans (myself included) use poor grammar like that, doesn't mean it's right.
However, one has to ask, does anybody really care?
I learned BASIC at age 5 on a PCjr. Read through that spiral bound book with the beige-ish cover entitled something like "Hands-on BASIC for the IBM PCjr".
I don't believe that learning BASIC screwed me over later on. Yes, BASIC is a somewhat screwed up language, especially the flavor that the IBM PCjr had without any concept of blocks of code or looping constructs. One of the main problems I had with BASIC is that as I became a better programmer, BASIC was not allowing me to do what I wanted to in an easy way. However, it does have the advantage that it somewhat resembles the english language, which is great for a newbie but gets in the way of an advanced programmer.
Modern versions of BASIC without line-numbers and the need to use GOTO just to do a loop or a block of code for the branch of an if statement may actually be a decent language to start with. I have an issue with arrays being indexed at 1 in BASIC which tends to lead the beginning programmer into a trap of off-by-one errors when he/she gets to more advanced languages. However you can for example specifically use "0 TO 9" to get an array of size 10 indexed at 0, but that also leads to a trap later since in most other languages you would specify "10" and not "0 to 9".
One thing about the PCjr's BASIC book is that it had some nice graphics/sound examples which made it a more interesting book because it fooled the reader into thinking that he/she was actually getting the computer to do something rather than going over programming theory.
I believe for a young kid, BASIC may be the way to go. However, for older students with some reasonable math experience and a decent attention span I see no reason not to start out with C++. I don't mean delving into objects and all that, but just starting with the basics. C++ has several classes to do common things and it's not necessary to know as a beginning programmer how it is done, just what it does. C++ is also a great language for doing data structures and implementing them as re-usable classes. However C++ does not quite have the instant gratification that BASIC does. Doing cool stuff like graphics/sound is usually not very easy in C++.
Another thing is that you probably shouldn't feed programming to anyone. Programming is something you really have to want to learn or you will not be any good at it.
I would suggest a very easy QBASIC book and the standard copy of QBASIC that I believe you can still find on a win9x CD. I would suggest doing the first few chapters with your son/daughter and then see if he/she takes off and starts learning on his/her own. I remember my dad trying to remember DOS commands when we first got the computer. It was actually funny as hell watching him type things and get "Bad command or filename" because he didn't know which commands to type (at least for a 5-year-old). I believe I was better than my dad at getting around DOS and programming in BASIC in a matter of weeks. What I didn't know until much later was that in his day he was actually a damn good programmer. Ironically, my mom is now better at using a PC than my dad is.
Did this several times for some friends in my dorm. Works wonders. Got a dudes report back (minus MS word formatting). That particular disk was even worse becaue the directories and part of the FAT tables were in bad areas, but I could still at least do some recovery.
Charge them for it too (well, not if you are an instructor, that's kinda sketchy). But seriously, after I let the dude know that he should keep backups and use better floppies to begin with (he had an old floppy that his mom got from work) he never came back with the same problem, and I think some other people on the floor found out about it and started backing up better.
The only solution to this problem is to properly educate your users of the merits of backing up data. It does not make any difference what type of technology you use for media, all of them will fail eventually. Removable media has the additional problems that it can be stepped on, cracked, crushed, spilled on, etc.
I would suggest that you let instructors know that they should remind their students to keep at least one backup copy of anything even remotely important. One of the simplest solutions has already been mentioned: keep data on two floppies. Another solution is to keep one copy on a floppy and another copy on your personal network drive. Keeping an additional copy on the hard-drive at home doesn't hurt either.
This solution is extremely easy to implement. My Systems Analysis and Design professor did exactly this when we began working on our 10 page+ reports. She said something like "Every year someone manages to loose their data and have to retype it all in to make the final project report. Be sure to keep multiple copies of everything." That one simply statement will fix most of your problems.
If you want to get a little more high-tech, there is a very good mod_dav for Apache that works with Microsofts Web Folders among other DAV clients. If you are unfamiliar with Web Folders, it's a lot like a network drive except it is an HTTP extension and unfortunately cannot be mapped to a drive letter. Also, I think the application needs to have some support for it to work like a real network drive. However, its absolutely great since you can download the content from ANY internet connected PC, and upload/download from internet connected PCs with MS web folders, this includes the users home PC. To the user it works basically like any other folder on the hard-drive. This is especially good if you are not a technically oriented campus.
In short, I suggest that you simply educate your users about floppy failures, explain how to use the network drive, and possibly set-up a DAV server for accessing content from anywhere on the net.
I beliebe the secure keyboard option is intended to prevent other X clients from receiving keyboard messages destined for the terminal window.
For instance, if I wrote an X client to capture all keyboard events and record them to a log file and then ran the client on your X server I could log all of your keystrokes and get your passwords. This is why it's important not to allow remote clients or even clients running on the same machine but as different users to connect to your X server.
For stupid users who might not have their X set up right (or for stupid distros that have incorrect settings) the secure keyboard option allows you to lock the keyboard into that terminal window. Try it, you'll find that no matter which window is focused the keyboard events go directly to the terminal window. With that said, I do not consider the option to be an alternative to a properly configured machine. For your own safety, keep your machine up-to-date with the latest bug fixes for your distribution.
Yeah, RPM definitely has some problems with different versions of packages. I could find no other easy way to upgrade to RPM 4 other than simply installing pinstripe. For upgrades from 6.0->6.1 or 6.1->6.2 I've usually been quite successfull with an rpm -Fvh [list most of the packages on the disk] and then going through and fixing the dependency problems. My main problem was that I couldn't find a way to upgrade the database from 3.x to 4.x format without running the installer
contrib.redhat.com pretty much blows goats because there are a lot of clueless people building packages that I definitely don't trust. Downloading the SRPM is about the best thing you can do, and it is definitely better to go that route than to build straight from.tar.gz, at least you can package it up then so all your dependency info works
Yes, packagers do get too zealous in their dependencies. My suggestion is to leave that alone and let RPM find the shared libraries the package depends on. If you have to list dependencies, do so by listing dependencies to the files you need (but not in absolute paths) and not the packages themselves. That way people who have installed things from.tar.gz don't get completely screwed.
Actually, the recent versions of glibc are able to be upgraded and still run programs linked against older versions because the symbols are versioned. With 7.0 there is a compat-glibc-6.2 package, but it is only for building packages destined for 6.2 on a 7.0 system, not running 6.2 packages on a 7.0 system. Running them is no problem even with the update to glibc.
I still think a tool that could "package" the make install phase of a program would be valuable, especially for people who like to configure and build everything themselves, or for people who are recompiling/installing the package constantly but still want it to be tracked by RPM and don't want to have to rebuild the whole damn thing each time.
Slashdot Mirror
You forgot the -H option to grep which comes in handy if you want to know WHICH file it found that string in. So:
find / -type f -exec grep -H somestring {} \;
The thing is that you know this for one of two reasons... either A) someone clued you in to doing this, or B) You read the manpage for find and found out that it has an -exec option which causes it to not print out the filenames but instead run whatever command you specify replacing {} with the filename and ending with a ; (which you then need to read the shell manpage to find out that you need to escape with a \ or do ';'). Then you remembered that grep is used to search text within a file given the text to search and the name of the file. However you try it without any other options and realize that you can't see which filename it is showing you the content of, so you man grep for info and find out that you can give it the -H option). Or maybe instead you don't know of this so you write a small script which runs grep and if it succeeds then prints the filename after it.
Option B is the true *nix users way of doing things and is what makes *nix so powerful. Option A is something you can tell a normal user to do if they need to search for text in all files in a directory (in this case /). Of course if the person is truly clued-in they will discover option B for other things and eventually realize the true power of the *nix CLI. At that point they will become a true believer and realize that all other systems are shit compared with this sort of power.
Do you really think that so many people would advocate Linux if they didn't think it was a better system? Conversely look at the people who advocate Windows. They too think that Windows is a better system and therefore advocate Windows. Although from my point of view I see their logic as being mostly flawed. I.e. arguments like "But Windows is so much easier!" don't agree with me. Windows is generally easier to learn because of it's GUI interface to everything. Linux with a GUI like that could easily match Windows ease of use and have more power for those who want to take advantage of it. The only drawback is that yes, it is unfamiliar territory. OTOH if you spend about 5 minutes and let a person know that it's still a GUI (you could even use a term such as windows-like) they may clue-in to the fact that if the play around with it for a second it won't be that hard to figure out how to do basic tasks. And since they already have experience with a GUI from Windows they will realize it's really not all that different.
The shitty thing is that you will still have those occasional few who have bought-in to the UNIX is old and outdated argument and don't realize that just about everything good about Windows comes from UNIX with the exception of the GUI which Billy-boy stole from the Mac and did a piss-poor job of stealing it. I.e. if UNIX is so bad why is the Windows TCP/IP stack based on the BSD TCP/IP code. Why does Win2K use kerberos for its domain logon/resource access functionality? Why does MS make a huge press release when they finally get around to implementing something that UNIX had 20 years ago and call it a brand new technology (e.g. symlinks anyone?)? Or here's another one: the ability to mount a filesystem in a directory of another filesystem instead of on it's own drive letter? It's all in the image. Some people really and truly believe all the marketing that MS spews and don't bother to learn anything else. Some people don't know the specifics of working with UNIX, but do know that MS is full of shit. And some people know the specifics of UNIX, know that MS is full of shit, and wonder why everyone else can't figure it out. :-)
Well, I like the premise of the article. I, as a POSIX person, fully understand this. In fact, I understand it so well that all I can say is "well DUH!" However, I think the author makes a very glaring assumption in the very beginning that a UNIX-like system is more stable than a Windows system. People who work with UNIX know this. People who work with Windows are of the mentality that it is stable enough. And I finally have caught the premise of the article that stable-enough really isn't. That the real thing is that the stability problems /do/ cost money. That the fact that your users don't trust their computers to work properly costs money.
What UNIX people don't realize is that most people don't understand this. And unfortunately this article sounded too much like pro-UNIX anti-Windows for any Windows person to read it. They would have stopped in the first few paragraphs saying how dare he say that Windows isn't stable. Hell, all computers make mistakes.
See, right now I am taking a full MCSE program from a local place where I live. Most of the people in my class think like this. The whole, well, computers have problems mentality. These people know nothing of UNIX. They know nothing of administration period.. with the exception of me and one other guy who also gets it w/ respect to how to be a good admin in general (he is/was a netware admin and his company is paying for him to do this). The only thing that all of these people know is how to get around a Windows 9x/Me system. Some of them were Gateway employees who did tech support or built systems.
Also, a lot of these people repeatedly utter the phrase, "Is this going to be on the test". In other words they see this as drop $10,000+ and get a full MCSE training and cert and go out and make lots of dough. Hell, it's advertised that way. What they don't realize is that you cannot be an effective administrator unless you know how the system works. And whats even more heart-sickening is that the MOC (MS Official Curriculum) actually does go into pretty decent detail about the how's and why's of Windows but these people don't give a fuck.
In general I feel that the MOC is a very good guide on administering Win2K systems. It's very practical and also explains the basic concepts. But again, these people only care about passing the exams. In fact, the instructors usual response is that "No, but you'll need to know it for work". Occaisonally it is "No, don't worry about it too much, just keep it in your head" because honestly, some of the stuff is a little too "Gee Wiz, look at what we've done, see how good our software is that we have this and this and this and this feature".
Anyway, the bottom line is that if everyone actually followed what the MS texts say that Win2K would not be horrible. And I think that the author of this article should note that. I understand that he does in a round-about way say that he is assuming a professional IS team for both Windows and UNIX, and that even with a good IS team for Windows you still have the basic problems that the software sucks in the first place.
Anyway, at some point soon here I will have all this Win2K knowledge and can actually give a straight honest-to-god answer that YES, I /HAVE/ done Windows.
One thing I have noted is that Win2K's strength lies in the fact that MS has basically designed your network for you. I know this sounds silly, but MS really has done a decent job of this. This is why Win2K is supposedly easier to administrate. You don't have to put much thought into design, you already have default groups like backup operators and so on. You can easily delegate authority to other members of the company. If you know what you are doing and understand the security implications it is easy to setup the system.. hell, the point that you by default allow nothing to happen and you grant people permissions to view/modify things is even stressed. This is good network philosophy. I actually applaud MS for this setup.
However, at the same time, their software sucks. I could just as easily setup default groups and so on for kerberos with Red Hat and sell that as a product. It'd be a lot more stable too.
Already we are seeing peoples computers running Win2K Advanced Server have stupid things happen. We have already had to reinstall one dudes machine because it somehow got fucked up. No one knows, no one cares, in this case it was easier just to reinstall from the ghost image because there was nothing important on the machine.
Now, at the same time, it seems that most of the instructors there do have a clue. They may not know UNIX for lack of time, they may only have experience with it in some college courses. But at least they understand why people use it, and would learn it if they had the time. The bottom line is that they spend so much time fucking around with the Windows boxes that they don't have time to learn Linux :-(.
Anyway, that is just my 2 cents on this whole thing. BTW, there are also some people there who have bought into the whole UNIX is outdated and sucks and Windows is new and cool and is better. Only a few have deluded themselves that badly though. That is of the people working there. In the class no one really has experience with anything so they don't really care.
Problem is that apparently SPA in Outlook is an MS specific thing. Well, what do you want them to do. The only way for outlook to support not sending the login in cleartext is to use SPA.
What about RFC 1734 [faqs.org] and 2095 [faqs.org]?
Sorry, I should have been more clear, what I meant is that Outlook (Express) already had support for MS SPA and not the other standards, and therefore MSN's only choice was to implement SPA or upgrade everyone to a different client. They did exactly this.
Basically all that needs to be done is for other mail clients to support MS SPA
This is where the problem is. There are existing protocols to deal with secure POP/IMAP authenticaion, but MS goes ahead and writes their own, and then people say "why doesn't everyone just support the Microsoft format" -- it is this line of reasoning that has led to nearly every proprietary closed format/protocol.
If MS doesn't want to pass passwords in cleartext, they should be using one of the existing and open methods of encryption, not forcing other people to use their software.
You are so right, they should have chosen an open standard for Outlook in the first place. Although I am not sure how many were widely accepted when OE 4 (first version named OE and not MSIMN) was released. Fast forward to modern times and MS still has no incentive to implement another SPA protocol.
And really, how many mail clients do you know of that support secure password authentication? I know that Netscape and Mozilla do not. Fetchmail does support quite a lot of authentication protocols. Interestingly, it does suport NTLM authentication for use with Exchange. Maybe this is a similar method. Knowing Microsoft they wouldn't reinvent anything but just use some crappy system they already have (i.e. NTLM).
HOWEVER: MS could VERY SIMPLY allow SSL connections which most mail clients do support and which would encrypt everything including the mail from the server to you. And furthermore OE does support SSL connections without MS SPA. However, this would be a lot of load on the mail servers doing all of that encryption.
So I suppose note well anyone who wants to give MSN a valid option, that is it. However, believe me I am well aware of the fact that MSN does not give a flying fsck about their customers, so YMMV.
I am hoping this is a joke and not for real. If this is supposed to be for real, then these people at the EFF need to be beaten severely with a clue-stick.
There is absolutely no guarantee that mail goes through. E-mail is not government regulated, period. I agree with them that there should not be stupid laws that try to limit SPAM and yet do something else. Currently, the best solution is something like MAPS.
What is so wrong about MAPS? The concept is sound. If you are connecting to my mail server (i.e. you are essentially treading on my property) then I have every legal right to tell you what you can and cannot do. I can also choose to not let you use my property. This is what MAPS does, it says, "Sorry, you can't use my server." With MAPS when you have fixed your server, you will be allowed to use mine again.
The argument that a system administrator cannot change her mail server is totally bogus. That is like MS saying they couldn't possibly remove IE from Windows. If the sysadmins like to be used as open relays, then I can let them know that they won't be allowed to send to my server. When they fix their server and notify MAPS then they will again be allowed to use my server. The same goes true for servers in the RBL. If your server sends a lot of SPAM and I don't want to incur costs to process it then I will simply deny everything from you until you fix your problem.
The statement that users get no notification of this is also FALSE. Assuming you can receive bounce messages (i.e. you don't have a problem with your mail configuration) then you will receive notice that the mail was rejected. Unless of course the mail server silently accepts the mail then tosses it, but that is not the point of MAPS.
The Internet is not government controlled, it's not entirely corporation controlled. It is a community network.
I don't understand how denying me the right to use MAPS is a good thing? I am supposed to accept mail from anyone regardless? Am I also supposed to let every jackass who walks onto my property and starts fishing in the bay do so? Can't I tell them that I own this property and you are not allowed to use it, so go somewhere else? Isn't this a fundamental right as well?
I especially love the other comments that we should replace SMTP with Jabber or ICQ because it is impossible to make SMTP SPAM-free. That's true if you still believe that every SMTP machine is supposed to relay everywhere. The reality of it is that an SMTP machine should 1) Accept mail destined locally. 2) Accept mail destined for a machine that it is a backup for, and 3) accept mail from a local user destined for anywhere. What MAPS is trying to do is get everyone to abide by these very simple rules. Those who run open relays get put in the MAPS RSS. Those who are connecting through dial-up connections get put in the MAPS DUL. Those who are themselves sending the SPAM or are allowing their customers to send SPAM are put into the MAPS RBL. What is so hard to grasp about this.
Oh, wait, didn't peacefire get blocked because their hosting company was allowing spammers to use machines in the same netblock. Tough shit. Go get a different hosting company, you aren't required to use them. Why not find one a bit less friendly to spammers. This is called free-market economy. Your e-mail getting blocked because your outgoing server allows relaying or is allowing spammers to use it regularly. Go find a different outgoing server. Anyone can sign up for a free hotmail account if it is just absolutely necessary to send a message that very second. Big deal, so you have been incovenienced. No one has taken away your rights though.. You have NO right to send e-mail to everyone you want to, that is the bottom line. You have a priviledge to do so. PERIOD.
I recently got laid off at a leading teleservices corporation that did technical support for MSN. This is because they completely dropped the contract with MSN (for what reasons I have only heard speculation and will not repeat here). I can assure you though that it was not because our standards were not good. Although it sounds like I am tooting my own horn we had probably the best call-center for all MSN service judging by the number of people calling back with ticket numbers started by people in other centers. I also judge this by the way the people wrote up their tickets without specifying what in the hell they did forcing me to go back through all the troubleshootings steps. (end rant)
First of all, MSN has two types of mail. They have the "legacy" POP3 system and the new web-based e-mail. You can find this information at MSN Support Services.
Web-based mail is kind of like what it sounds. It uses the same mechanism (XML over HTTP) that Outlook Express >=5 uses to access hotmail. However the server for @msn.com accounts is different from the server for @hotmail.com accounts. If you have an @msn.com web-based account you can go to http://supportsevices.msn.com/us/oeconfig/ to configure OE and then go to tools accounts and read the server name out of there. Note, this is also true for free @msn.com accounts. Note that only Outlook Express 5 or greater or Outlook XP can use this mail. Obviously MSN Explorer and the hotmail.com website itself are compatible with this.
Anyway, it seems the real issue is that these people would like to use their new MSN POP3 accounts with e.g. fetchmail. To correctly configure Outlook Express for MSN POP3 e-mail you must use the outgoing (POP3) server of pop3.email.msn.com (go figure) or the incoming (SMTP) server of smtp.email.msn.com. Furthermore you must select the "Log on using Secure Password Authentication" option as well as select the option under Outgoing Mail server that "My server requires authentication". You then must press the settings button and be sure it is using the same settings as the outgoing mail server. That is it logs on using SPA with the same U/P as the POP3 server.
Because of this MSN states that you MUST use Outlook to get your MSN POP3 mail. This is not entirely correct. What you must have is a client that supports SPA. Why is MSN doing this? MSN's reason: to reduce SPAM. However they tell customers simply this because most of their customers are rather computer illeterate (especially the former AOL lusers). The real reason is that since they contract out Dial-up Points of Presences (Pops, not to be confused with POP3 e-mail) that either A) they must use the POP3 before SMTP hack, or B) You must login to the SMTP server to send mail. If they didn't do this then any jackass dialing into one of those POPs even with another ISP would be able to send tons of SPAM through MSN servers. There have been plenty of /. articles about this before and anyone familiar with how contracted out POPs interfere with the ability to allow SMTP access to only your subscribers should know what I am talking about.
Now, MSN /could/ have simply kept the plaintext login POP3 and only required you to use a plaintext login for SMTP. However they decided that not only should they require a login for SMTP but at the same time they should require secure password authentication for both POP3 and SMTP. In other words, if they were going to have to have people change their Outlook mail settings they might as well knock out the ability of people to sniff the packets and retrive their users passwords while they are at it.
Problem is that apparently SPA in Outlook is an MS specific thing. Well, what do you want them to do. The only way for outlook to support not sending the login in cleartext is to use SPA. So therefore they enabled SPA on their mailservers and disabled clear-text logins. Of course theoretically they could include some other more open method of secure password authentication for use with other clients, or they could open up the MS SPA protocol. Or they could just say the hell with it because they only officially support MSN using MS software on Windows OSes (which actually does NOT include WinCE, you must contact your OEM for WinCE support with MSN).
Basically all that needs to be done is for other mail clients to support MS SPA. How to do this I am not really sure as I have not put much thought into it as I don't use MSN myself except for free accounts. All the free accounts use hotmail based e-mail.
There is of course another option. You could always "upgrade" your account to web-based from POP3 and then either go to the hotmail website to get your e-mail or use Outlook Express >=5 or Outlook XP to get your email in a real mail client (if you can call Outlook a real mail client, but hey, at least's it better than www.hotmail.com). There does exist a script (PERL I think) for retrieving mail from hotmail but I have looked at that code and it is really really crappy (apologies to the guy who wrote it, but I am sure he also knows that it is nothing more than a quick hack). Theoretically there is no reason that Evolution should not support the MS HTTPmail protocol. Turn on HTTP logging in the Advanced tab of OE properties and then open up the log in notepad. You will notice that the schema is relatively easy to figure out even though to the best of my knowledge it is not published anywhere. Evolution already uses XML extensively and has all of the framework necessary for parsing XML. I assume it also has the framework necessary for accessing an HTTP server in general. Therefore it should be rather trivial to write an MS HTTPmail backend for Evolution. In fact, I am surprised that no one has done so (I guess none of the developers use hotmail). I have toyed with the idea of doing one myself but 1) I use balsa, and 2) I have not done any programming with XML. However now that evolution is fairly stable I may go ahead and write this. Hell, I don't have anything pressing to do until Monday except clean the garage so we'll see. There's never a bad time to learn more programming techniques, and XML is one of the most popular things today so not only would I personally benefit from learning XML but also benefit with being able to access hotmail from evolution. And note well... if I do write this I do intend to support the advertisement properties as best as possible (i.e. opening up a small frame at the bottom and displaying a webpage in it). I know it seems stupid, but hey, they deserve to get paid even if they are MSN. And if anyone really wants to they can just change the code later to take out the ads.
Anyway, I hope this clears up a lot of the confusion people are having with this. I see at this point over 600 comments have been posted, a few reasonable, most along the lines of fsck Microsoft. People, I hate MS as much as everyone here. They are theives and crooks and must be beaten. However, as the cliche says: You catch more flies with honey than with vinegar. The only way MS will be beaten is when people stop bitching about them and just go do better than them. Every time I bitch about MS to my mother she reminds me: Then go write something better. While everyone has argued this point to death the bottom line is that in some respects MS software is "better" than open-source/free software. Even if only in the marketing sense of better.
-Dave
If anyone is interested in buying this and has a spare half a mill laying around in one of his/her checking accounts, go for it!
On my server I have a shell script run by the webserver which echos the REMOTE_ADDR variable along with the current time into /tmp/cripfifo (Code Red IP FIFO). From there I have a script that I run as another user that does and endless loop of cat /tmp/cripfifo | while read i which then loops and for each line it inputs (which is actually probably only 1 since the fifo is closed each time the webserver script writes into it). With each line of input it first checks to see if there is a file named $REMOTE_ADDR* in the errorips directory. If not it runs another script in the background which uses wget redirecting output into errorips/${REMOTE_ADDR}_TIME. If it succeeds it moves that file into the processed/ directory which means that if the machine attacks again it will repeat this processes each time it attacks.
Yeah, I know there are some race conditions if the same IP gets processed before the other one has had a chance to write the logfile. It's not a big deal though, and the chances of that happening are slim anyway unless someone is just bombarding you with attacks in which case it still doesn't matter.
See http://kernelrpm.sourceforge.net/codered/ for the scripts. Note, it may be a better idea to use the Location directive for /default.ida and point it to the script rather than making the whole damn directory ExecCGI. Although I don't have a problem doing that since only I have access to it anyway.
Obviously you think this or you wouldn't be asking the question. However, this isn't far from the truth :-(. First of all, the program you probably want to use is mgetty which is included with Red Hat. In 7.1 there are four mgetty packages on the second disk. You definitely want mgetty installed, and probably mgetty-voice if you intend to use this as an answering machine type system. Now, I have browsed through the documentation directories for mgetty and I do agree it is pretty lame. In addition trying to search for mgetty is not very helpful. The first result returned from google is http://www.leo.org/~doering/mgetty/ which seems to be about the best reference you will find (so google seems to be doing its job). However the distinctive ring is buried in the caller-ID section and I am not exactly sure how to go about getting this to work. My advice is to read the docs thoroughly and also look for comments in the config files. There is also a mention in an mgetty FAQ on faqs.org of a mailing list for mgetty. You will probably get a better response on a mailing list for the software than on a general forum such as slahsdot.
A few things you should know is how to put your modem into distinctive ring mode. You will probably need to change an S register so that the modem will report RING 1, RING 2, RING 3, etc. instead of just plain RING. Use minicom to test this out. You may be able to do an AT$ or ATS$ to get lists of commands (works on USR modems, and I think also some Winmodems have something similar, maybe just the USR/3com ones.
I hope that helps you out in your quest. If nothing else start viewing the mgetty source and if necessary patch it to better support distinctive ring. Or maybe it has good support but is just poorly documented, so document it so the next guy can figure it out.
From the looks of your post you are just another troll. You say you find Linux to be lacking and that in all the key areas of business it is clearly inferior to Windows 2000. However, you don't actually go on to explain why Windows 2000 is so superior to Linux in key areas of business. The most you do is say that Linux is obscure, hard to use, and inconsistant. Did you read that off of a Microsoft website or out of a ZD net magazine?
For the purposes of the rest of this comment I am going to assume that you do in fact have reasons you have just neglected to mention them.
You mention that you make your living providing solutions based on Microsoft technology-- good for you. I am sure you have quite a bit of knowledge relating to Microsoft products. However, I am also fairly certain that you have not even begun to learn UNIX in the same way that you have learned Microsoft technology. If you actually took some time to evaluate Linux and brought in a real Linux guru, you may find that for many applications Linux is clearly a better choice. Likewise for many applications Windows is a better choice.
You mention that if businesses were not making money from their MS systems they wouldn't stick with them. That is entirely true. If the system is in place and meeting the needs of the business it would be absolutely idiotic to switch to Linux. However, if the system is not meeting all the needs of the business then it is prudent to evaluate all options before proceeding with an information system upgrade. An entry level systems analysis and design class will tell you that. Using another technology may be a viable option, and is at least worthy of being included as one of the proposed solutions.
You also mention that if Linux offered any competitive advantage that businesses would install redhat straight away. This is also not correct for the same reasons as above. Haven't you ever heard the cliche "if it ain't broke, don't fix it?" This is especially true for information systems. If it would cost more money to migrate to Linux than would be saved by migrating to Linux then it is not worth it. That is basic business practice. You can also find that out in any entry level systems analysis/design class or in a more general form in any entry level business course.
You say Linux is obscure, hard to use, and inconsistant. Have you taken a look at Windows lately? Or even MacOS for that matter. Every system is somewhat inconsistant, hard to use, and obscure. Those are pretty vague adjectives, try to come up with some specifics.
In the past Microsoft has pushed out several competitors. You mention OS/2 and Amiga. For one thing, Amiga is a different hardware platform entirely and was never really considered to be a serious business machine. Most businesses used IBM computers and turned to IBM compatible PCs when they became available, so I can't see how Microsoft kicked Amiga's ass. Microsoft did kick OS/2's ass pretty hard. However, OS/2 had some drawbacks as well. Version 1.x was IBMs baby although MS did much of the coding and design. IBMs original purpose with OS/2 was to provide a better operating environment to entice consumers to switch to IBM computers running OS/2. Version 1.x had such shitty hardware support for anything not IBM that almost no one could use it. During that time Windows was not even anywhere near usable. It wasn't until MS decided to split from the project (and go work on Windows 3.0 which could now run in protected mode) that IBM decided to compete on the software level.
At that point Microsoft launched a massive marketing campaign to put Windows on every desktop and have every application written for Windows. They also had Windows preinstalled on every new PC and penalized manufacturers who installed other operating systems. Of course no one used OS/2. Why spend another $100+ when you already have Windows on the machine? It's a wonder OS/2 lasted as long as it did as a desktop OS. However, you may note that OS/2 is still used for many embedded devices and that a new version has just been released renamed eComStation.
In your final paragraph you mention that geeks who want to succeed in the information economy should study some business courses. That's extremely good advice. You also say that their time would be better spent getting an MCSE instead of rebuilding the kernel every 2 minutes. That's true too. If you want a professional quality Linux certification, they are available.
First of all, IE is not my main browser. For this comment I am using Mozilla. However it just so happened that I was testing out IE on a newly compiled CVS build so I figured I'd write the comment.
I would have to say that I am not very thrilled with VMware from my personal experiences. IE running under Wine runs faster than Mozilla. IE running under VMware runs slow as shit. IE under Win4Lin is a bit slower too (but not very noticably so). Using Wine as a Win4Lin/VMware type thing is not such a bad idea. Sure it is not quite as compatible at the system level, but at the application level most things are working. Like I said, one of the major drawbacks to Wine is the USER/GDI code. With working core libraries in place WINE would provide a better solution than Win4Lin or VMware. Especially if a company (CodeWeavers would be the first Wine related company to come to mind) did some more work and packaged Wine as a Windows applications runner. However, I don't think you'll see them doing that, it doesn't seem to fit in to their business model from what I gather. However, Jeremy White and friends did do a very good job of packaging the Codeweavers releases of Wine.
About the license thing. Most (possibly all) of the DLLs I used native versions of actually came with IE because it updates the system. So theoretically if you can get the installer to work then you shouldn't need Windows at all.
Yes, that's right. With the latest CVS updates the painting is better too! It now doesn't white out the toolbars when you load a new page. The toolbars still start out white sometimes, but nothing a quick rollup and rolldown can't fix (which used to not be true either).
My IE 5.5 is installed on a copy of Win95 OSR2 with updates (including Y2K) under Win4Lin. Unfortunately, since Win4Lin doesn't work with kernel 2.4 yet I am unable to use it, but it makes a nice installed Windows tree with native DLLs for Wine to use. Currently I am running with native commctrl,shell, and ole.
There are still some issues with the user code in Wine. It's old. Recently Alexandre Juliard has submitted some changes to better seperate the Wine graphics driver out of the user code. I believe Alexandre is looking to do some major improvement to user. Well, at least I hope so, the current user code is very old and crufty and is the cause of lots of painting issues. Even solitaire doesn't exactly paint right at the moment (move a window around over top of it and you'll see what I mean).
Wine has come a long way and if you haven't had the opportunity to use it lately, I suggest that you try it again. IE seems to be working just fine for posting this.
So how exactly are you supposed to install this other RPM on a machine? Does it contain the source code in pristine form with patches and then must be compiled? It would seem that that would be the only way to satisfy the conditions of the license. If that is true then what the hell do you do when your server doesn't have a compiler (which it shouldn't).
I suppose now you will pull the argument that qmail is uber secure and thus you don't need the rest of the system locked down. That is total BS.
Red Hat makes a binary distribution using open source code and their own patches to better integrate components. The qmail license almost directly contradicts with the purpose of a binary linux distribution. It also contradicts the GPL.
It is entirely true that Red Hat is not allowed to include a modified binary in their distribution. Sure they could include some dumb hack to get around the license, but why should they.
You fucking people need to get it through to your small ass frontal lobes that businesses need to make money. Red Hat provides for their customers a prepackaged linux distro and automated package updates. Red Hat is not obligated to include every package under the sun. They include what they see as useful to their customers. If you want qmail on Red Hat then fucking grab the source and install it. Damn!
Let me get this straight, you admit that you can't distribute modified binaries and must distribute patches instead of modified source. Before this you say that these programs are free in the GNU sense of the word. I am sure Stallman would love to rip you a new asshole for that comment. I cannot believe that your shit post got modded up to 5.
And about this whole RedHat shipping Mozilla instead of Netscape: years ago when they included the Netscape binary only package it was said that part of the motivation was that netscape released most of the code. The original intention was always to switch to Mozilla from day 1 of including Netscape. This article is really old news, although I am sure most people have forgotten that by now.
Have fun trolling. I certainly hope your message comes up for meta-mod.
Since when did the term hacker become a catch all for criminals who use a computer? I am absolutely appalled by this. While it is true that words should and will naturally evolve in a language, this is really not the case here.
It seems to me that the author or editor decided to use the word hacker in the title for some sensationalism. This story does not even describe what an average citizen would call a hacker. Most people think of a hacker as someone who infiltrates a computer system. Most people also think that hackers steal information and damage property (e.g. delete or destroy data).
In all reality, a hacker is someone who is extremely skilled with computers. A hacker also has knowledge about how to break into a system. For instance, any system administrator worth a damn is a hacker. That is to say that they are skilled with computers and know how to break into the system. This puts them in a position to keep computer criminals (note the word choice, criminals not hackers) from breaking into the system.
In this article, the computer criminals are just that, criminals. They are simply modern-day scam artists. But hey, "Watch a computer scam-artist work the system" probably doesn't have that same ring to it.
Originally I was going to send an e-mail to Bob Sullivan (the author) but if you do a search for the word hacker on the page you will not find it in his article, only in the headline and in links to other MSNBC headlines. Apparently this is just some editor with his head way up his ass.
Well, if he's an Indiana resident, he most likely will. Likewise so would residents of other states. You could theoretically make a sourceforge project out of this if you could somehow garner enough interest from many people.
Best thing to do would probably have a core team work out the basics of a tax program and then have other teams make modules for the different states. Probably in some scripting language (python would be a good choice given good GTK bindings and support as a scripting language in GNOME).
Trying to compare RPM to apt-get is illogical. Debian's apt-get takes a package name and goes through your list of servers and looks for the package. The format of these packages is .deb. Apt will automatically get the .deb you requested plus any .debs it depends on.
Most other distros have standardized on the RPM package format-- possibly because a lot of them were originally based on Red Hat at some point in their life. RPM is a package manager. It serves the same purpose as the debian package manager (which is not apt). Unfortunately, there has not been an equivilant for Apt on RPM-based distros.
However, recently Connectiva or someone modified Apt to work with RPMs as well as DEBs, and if you read some of the other comments here supposedly it works pretty well. Also, Ximian (formerly Helix-Code) has come up with the new Red Carpet Updater. Red Carpet also serves the same purpose as apt. Unfortunately, Red Carpet is completely GUI from what I have seen and there is no command line equivilant. This kind of bothers me as it would be nice to use parts of it in scripts. On the other hand, it is free software so anyone could make some good command-line tools to go with it.
I have been using Red Carpet 0.9 since its release and I really like it. A big warning though: It does not like unsatisfied dependencies at all. In fact, I had NVIDIA_GLX (which depends on NVIDIA_kernel) installed but had no NVIDIA_kernel package because I rolled my own kernel and named it and the NVIDIA driver differently than usual. Well, because of this unsatisfied dependency (that is, NVIDIA_GLX depends on NVIDIA_kernel, but there was no packaged named NVIDIA_kernel on the system) it wanted to removed NVIDIA_GLX. And for some stupid reason it decided that it also wanted to removed all the XFree86 packages and other stuff. I of course declined and then eventually traced it back to having no NVIDIA_kernel package.
Basically I would say that Red Carpet is a bit to stringent about dependencies. On the other hand, this is designed to ensure the system stays stable. The bad thing is that I could just see someone following its advice (which I chose not to) and having it remove XFree86 and a bunch of stuff and then being unable to run Red Carpet to get it all back because it requires X. Whoops, must have forgetten a dependency there!
I suppose that is more than you wanted to know, but if you have an RPM based distro you should really check out Ximian's Red Carpet. And Petrely needs to get his damn head on straight and realize that RPM is comparable to deb and apt-get is something different. At this point with Ximian Red Carpet, there is basically now apt-get for everyone so this whole article is a moot point.
And oh, as another poster said, I rarely hear of any RPM users screaming that they want to switch to Debian. Personally I like RPM and from what I have seen of deb I see no reason to switch. The only good reason to switch to debian seems to be apt, but now that one company has made apt work with RPM and another has written a new program that works from the start with both deb and rpm I really don't see any reason to switch.
The main thing about @Home is that not all @Home is created equally. The cable company has a lot of freedom to do different things. In Indiana (where I live, with AT&T@Home) part of your monthly fee pays for the modem and you can get a $5 discount or something by using your own modem. In Virginia (where my parents have now moved, with Cox@Home) you must pay for the modem. AT&T charges $5.95 (IIRC) for additional IPs with the max being 6 total or something. Cox charges $6.95 and does not allow more than 4 total computers. In addition they charged a $10 "setup fee" for the IP. So I basically paid them $10 so I can continue paying them $6.95 every month. What the "setup fee" is for I don't know. I know it is an automated process to add the hostname and IP to the DNS server since it works immediately. However, DNS lookups did not immediately return the address and a reverse lookup on the IP returned someone elses hostname (presumably someone who had cancled service). It does now however work correctly (it's been a couple days). Its possible that they actually have someone update the DNS server manually which would explain the delay, or maybe they just don't restart it for a while.
Currently AT&T (and Cox from what I am seeing on my dads computer) use static addresses but also support DHCP. Most installers in my area of Indiana always type in static because the DHCP server tended to go down constantly. Ask a low-level tech from @Home about static and they fucking freak out because "we will be switching to dynamic addressing in the near future". Says the same thing when you order additional IPs that you need to set your hostname and use DHCP and that its bad to configure them statically. In all reality there are so many customers with static addresses that it would be an absolute support nightmare to switch. In addition, I have heard that AT&T could give a shit about static but that it is @Home that wants to go dynamic. That is way through the grapevine but I believe it because @Home's DHCP servers go down so often in my area in particular that its cheaper for AT&T if everything is setup static because it is less support calls.
Should the configuration be wiped it will be necessary to type in the correct hostname and/or type in the IP number/netmask/gateway/dns servers anyway so there is no advantage to going with dynamic addressing UNLESS they would just randomly assign IPs to any computer that asked for one via DHCP. In that case the support costs would go way down because basically as long as the DHCP server works and the computer simply has a DHCP client it will get an address and work. The disadvantage is that your IP would then be truly dynamic and changeable which I don't like. I prefer to pay my $5 or whatever and have a real static IP. OTOH, many DSL providers do a hybrid of this system. That is, they will give out random IPs to any computer that asks for one but you may also purchase static IPs for $5 a pop. In my opinion, that is what I wish AT&T would do, it seems like the most fair system.
As for the other DHCP server (the one that gives the modem its address) I seriously doubt that it works any differently and most likely you couldn't just plug a new modem in and expect it to work.
Okay, here is how it works:
Your modem has a MAC address (two actually). The MAC address identifies the modem for use with DHCP. The DHCP server gives the modem a 10.x.x.x address and it is online (there is a whole bunch of other stuff involved, but essentially that is how it works). The modem must be provisioned to some account to work. There may be a way to provision a modem to a non-existant account but it would have to be a very inside job at @Home (in Colorado most likely).
The modem essentially functions as a bridge between the ethernet interface and the RF interface. Although they are two different protocols on the physical level, they both speak TCP/IP so I believe it would be most correct to call it a bridge. The MAC address of the ethernet interface on the modem (which is like never used) is always one more than the MAC address of the RF interface. You would be most concerned about the MAC address of the ethernet interface if you were screwing with your ARP cache I suppose, otherwise you really don't need to care about it.
Now, assuming your modem is connected as described above, then you essentially have an ethernet network like any other. Your computer has a hostname usually beginning with the letter C (hence called the Charlie number by some techs). There is then a 5-7 digit (usually 6) number identifying the subscriber. There is then a dash and a letter identifying the computer. So something like c780677-a (which is actually my hostname). My second IP address that I ordered is c780677-b. Your DHCP client contacts the DHCP server, gives it the hostname, and gets an IP address (standard DHCP procedure). They do not use your ethernet cards MAC address like most other DHCP/BOOTP setups because you may want to change network cards.
So, assuming what you said was true, then the only thing you would be able to do is connect multiple computers to one cable modem without paying for additional IP addresses. Actually, that would be a pretty sweet deal. Grab another box, set it's hostname to the same as yours but with a -B and see what happens. If you get the address without paying for it, more power to ya.
On a RedHat 7.0 style system (also 6.2 IIRC), set the DHCP_HOSTNAME variable in the /etc/sysconfig/network-scripts/ifcfg-INTERFACE file to the desired hostname. In Windows you set the computers name to the desired hostname, also you may want to enable DNS and set the correct hostname and domain as well. For example, hostname "c780677-a", domain "[cityname]1.[2letterstateabbrev].home.com". Also set the workgroup to "@HOME".
I think that is probably everything you ever wanted to know about cable modems.
It is my understanding that Christmas is supposed to be about celebrating the birth of Jesus Christ. So what exactly is the issue of using the F-word in a Christmas card to some personal friends. If it was part of an inside joke then apparently these are at least fairly close friends or were at one time.
I don't know about you, but I generally say the F-word all of the time around my friends but try to avoid its use when around other people. Think about the irony in that.
Err... shouldn't that be "This is a situation with which I will not put up." You can improve it further by switching to the active voice and saying, "I will not put up with this situation."
The original poster was right, it should be "Chip news on which to crunch." In spite of that, most people would agree with timothy and say "Chip news to crunch on." But just because most midwestern americans (myself included) use poor grammar like that, doesn't mean it's right.
However, one has to ask, does anybody really care?
I learned BASIC at age 5 on a PCjr. Read through that spiral bound book with the beige-ish cover entitled something like "Hands-on BASIC for the IBM PCjr".
I don't believe that learning BASIC screwed me over later on. Yes, BASIC is a somewhat screwed up language, especially the flavor that the IBM PCjr had without any concept of blocks of code or looping constructs. One of the main problems I had with BASIC is that as I became a better programmer, BASIC was not allowing me to do what I wanted to in an easy way. However, it does have the advantage that it somewhat resembles the english language, which is great for a newbie but gets in the way of an advanced programmer.
Modern versions of BASIC without line-numbers and the need to use GOTO just to do a loop or a block of code for the branch of an if statement may actually be a decent language to start with. I have an issue with arrays being indexed at 1 in BASIC which tends to lead the beginning programmer into a trap of off-by-one errors when he/she gets to more advanced languages. However you can for example specifically use "0 TO 9" to get an array of size 10 indexed at 0, but that also leads to a trap later since in most other languages you would specify "10" and not "0 to 9".
One thing about the PCjr's BASIC book is that it had some nice graphics/sound examples which made it a more interesting book because it fooled the reader into thinking that he/she was actually getting the computer to do something rather than going over programming theory.
I believe for a young kid, BASIC may be the way to go. However, for older students with some reasonable math experience and a decent attention span I see no reason not to start out with C++. I don't mean delving into objects and all that, but just starting with the basics. C++ has several classes to do common things and it's not necessary to know as a beginning programmer how it is done, just what it does. C++ is also a great language for doing data structures and implementing them as re-usable classes. However C++ does not quite have the instant gratification that BASIC does. Doing cool stuff like graphics/sound is usually not very easy in C++.
Another thing is that you probably shouldn't feed programming to anyone. Programming is something you really have to want to learn or you will not be any good at it.
I would suggest a very easy QBASIC book and the standard copy of QBASIC that I believe you can still find on a win9x CD. I would suggest doing the first few chapters with your son/daughter and then see if he/she takes off and starts learning on his/her own. I remember my dad trying to remember DOS commands when we first got the computer. It was actually funny as hell watching him type things and get "Bad command or filename" because he didn't know which commands to type (at least for a 5-year-old). I believe I was better than my dad at getting around DOS and programming in BASIC in a matter of weeks. What I didn't know until much later was that in his day he was actually a damn good programmer. Ironically, my mom is now better at using a PC than my dad is.
-Dave
Did this several times for some friends in my dorm. Works wonders. Got a dudes report back (minus MS word formatting). That particular disk was even worse becaue the directories and part of the FAT tables were in bad areas, but I could still at least do some recovery.
Charge them for it too (well, not if you are an instructor, that's kinda sketchy). But seriously, after I let the dude know that he should keep backups and use better floppies to begin with (he had an old floppy that his mom got from work) he never came back with the same problem, and I think some other people on the floor found out about it and started backing up better.
The only solution to this problem is to properly educate your users of the merits of backing up data. It does not make any difference what type of technology you use for media, all of them will fail eventually. Removable media has the additional problems that it can be stepped on, cracked, crushed, spilled on, etc.
I would suggest that you let instructors know that they should remind their students to keep at least one backup copy of anything even remotely important. One of the simplest solutions has already been mentioned: keep data on two floppies. Another solution is to keep one copy on a floppy and another copy on your personal network drive. Keeping an additional copy on the hard-drive at home doesn't hurt either.
This solution is extremely easy to implement. My Systems Analysis and Design professor did exactly this when we began working on our 10 page+ reports. She said something like "Every year someone manages to loose their data and have to retype it all in to make the final project report. Be sure to keep multiple copies of everything." That one simply statement will fix most of your problems.
If you want to get a little more high-tech, there is a very good mod_dav for Apache that works with Microsofts Web Folders among other DAV clients. If you are unfamiliar with Web Folders, it's a lot like a network drive except it is an HTTP extension and unfortunately cannot be mapped to a drive letter. Also, I think the application needs to have some support for it to work like a real network drive. However, its absolutely great since you can download the content from ANY internet connected PC, and upload/download from internet connected PCs with MS web folders, this includes the users home PC. To the user it works basically like any other folder on the hard-drive. This is especially good if you are not a technically oriented campus.
In short, I suggest that you simply educate your users about floppy failures, explain how to use the network drive, and possibly set-up a DAV server for accessing content from anywhere on the net.
I beliebe the secure keyboard option is intended to prevent other X clients from receiving keyboard messages destined for the terminal window.
For instance, if I wrote an X client to capture all keyboard events and record them to a log file and then ran the client on your X server I could log all of your keystrokes and get your passwords. This is why it's important not to allow remote clients or even clients running on the same machine but as different users to connect to your X server.
For stupid users who might not have their X set up right (or for stupid distros that have incorrect settings) the secure keyboard option allows you to lock the keyboard into that terminal window. Try it, you'll find that no matter which window is focused the keyboard events go directly to the terminal window. With that said, I do not consider the option to be an alternative to a properly configured machine. For your own safety, keep your machine up-to-date with the latest bug fixes for your distribution.
Yeah, RPM definitely has some problems with different versions of packages. I could find no other easy way to upgrade to RPM 4 other than simply installing pinstripe. For upgrades from 6.0->6.1 or 6.1->6.2 I've usually been quite successfull with an rpm -Fvh [list most of the packages on the disk] and then going through and fixing the dependency problems. My main problem was that I couldn't find a way to upgrade the database from 3.x to 4.x format without running the installer
contrib.redhat.com pretty much blows goats because there are a lot of clueless people building packages that I definitely don't trust. Downloading the SRPM is about the best thing you can do, and it is definitely better to go that route than to build straight from .tar.gz, at least you can package it up then so all your dependency info works
Yes, packagers do get too zealous in their dependencies. My suggestion is to leave that alone and let RPM find the shared libraries the package depends on. If you have to list dependencies, do so by listing dependencies to the files you need (but not in absolute paths) and not the packages themselves. That way people who have installed things from .tar.gz don't get completely screwed.
Actually, the recent versions of glibc are able to be upgraded and still run programs linked against older versions because the symbols are versioned. With 7.0 there is a compat-glibc-6.2 package, but it is only for building packages destined for 6.2 on a 7.0 system, not running 6.2 packages on a 7.0 system. Running them is no problem even with the update to glibc.
I still think a tool that could "package" the make install phase of a program would be valuable, especially for people who like to configure and build everything themselves, or for people who are recompiling/installing the package constantly but still want it to be tracked by RPM and don't want to have to rebuild the whole damn thing each time.
-Dave