1) You're right that you can't delete tickets out of RT, but I don't understand why you'd want to. Just because an issue is dead, does't mean that all records of it should be purged from the database. "killed" is just another status, like open, stalled and resolved and just shows that this issue no longer requires action.
2) Your performance problems are totally whacked and don't correspond at all to my experience. I've seen RT 1.x systems with more than 500,000 tickets in them that don't have anything like the performance problems you're describing. Searches through the tickets only takes ~10s on moderate hardware on this system (with more than 500,000 tickets). Have you done something silly like not turned indexing on on the important tables?
Well I often stream from the Japan-A-Radio shoutcast stream. JAR has a store where you can purchase J-Pop and Anime music. Other than that my local used video game store carries a pretty fair selection of J-Pop and Anime at normal ($15) prices.
Say Hi to everyone in the Japan-A-Radio chatroom at #japan-a-radio on irc.d00mnet.com and most other IRC networks.
But unlike VM, the domains run in partitioned sets of hardware, so that (for instance) any CPU fault will only bring down a single instance, whereas with VM style logical partitioning you could lose a whole lot of domains with a single hardware fault.
I believe that you are mistaken. On an IBM system if a processor smokes all the tasks that wer e running on it will be automatically moved to a redundant spare and execution will resume from the last known good state. The Linux image would continute processing normally and would be unaware that anything at all had happened. I also believe that the machine would generate an exception alert for the local admins, as well as call IBM Global Services directly and report the problem so that a tech can already have been dispatched by the time the customer calls reporting a problem.
All the documentation that I have read indicates that IBM VM is significantly more advanced than Sun partitions. All the whining from Sun won't change that anytime soon, only several years of hard engineering will close the gap.
Linux on a mainframe can be great if you have a computing task that matches what mainframes are good at. It can also be good if you have an existing mainframe and wish to get more use out of it.
Context sensitive help should not merely repeat the error message, it should explain the issue
If I could add, have an error number along with the error message. For example instead of having a box pop up saying "You can't do this here", have it say "Error: ABC1234: You can't do this here". It is much easier to search by the error number in the online help and on Google then by the text of the message. This goes double when you app is localized in different languages.
While not exactally the information you were looking for, there is a perfectly good LaTeX macro set for making scripts. This many be a good base for creating an XML Scheme, knowing what tags you will need and such.
I would like to take exception to most of the points that you have made. They are not base problems with the technology, they are mainly problems endemic to DOS/Windows. Even running on the exact same hardware, most other OS's will not have these problems.
1. Upgrading one piece of software or one hardware component (e.g. video card) can easily turn into a cascade of upgrades and a week's worth of evenings. I've gotten afraid to upgrade; I don't want to mess with something that works.
The software component of your complaint is mainly a package management problem. Windows systems traditionally have terrible/non-existant software management tools and a very high level of interdependancy of software components that is usually completely non-obvious and opaque to both user and admin alike. To answer to the hardware part of your question this is generally just a Windows problem as well. For Mac's, hardware is easy (see the AC's post). On my desktop machine, all relevant drivers are included with the base OS and it has limited detection of new/changed hardware. I recently built a new Athalon MP system to replace my aging Pentium 233MMX system and merely moved the harddrive from one system to the other. It autodetected my new MB and NIC without any effort required on my part, changing my XF86Config-4 file consisted of adding a new Device section and changing one line. Less than 5 minutes effort all told. At no point did I have to scrounge on floppies/cdrom for drivers or have to update from the manufacturers website to get things working up to spec.
2. The rash of awful virii and worms that get released for whatever system provides the most opportunity (note: If Linux were on 95% of all desktops, there would be just as many Linux viruses; thinking otherwise is like thinking you have developed an unbreakable copy protection scheme). Keeping up with all the security patches and such has been a real headache. And unless I keep up with sites where these things are announced, I'd never know about them.
This has been hashed over again and again and I don't believe it to be true. Windows has a really bass-ackwards method for loading programs and in 9x versions absolutely no useful memory protection or system security. Last I checked NT based systems require many things to be world-writable, although SFP helps a little. Systems such as Linux have a completely different method for determining whether a program is loadable or not which significantly raises the bar for infection (ie. way higher than a Y/N popup from your web brower or email client) and can insure greater integrity of the system so that errors like rm -rf are more easily recoverable. With very little admin effort (which may be done by the vendor for you) you can make your system nearly invulnerable to everything short of a buffer overflow in your browser or MUA (and header and MIME parsing seems to be the only automatically exploitable vulnerability possible in an MUA). Anyway on a better designed system, without so many endemic design flaws, viruses have a much harder time infecting the system (propagating should be just as easy). Also there are even better security architectures that could be made default at any time (LSM w/Flask, SubDomain, Janus, POSIX, etc.) that could make the network a very hostile place for viruses. I predict that there will be no requirement for antivirus softare on Linux systems
3. There's still a general unreliability factor associated with PCs. Sometimes my PC doesn't boot completely, and I have to power down and try again. Ever run a game and hear the monitor click indicating a resolution change, and then nothing happens and even if you could kill the game you can't get your video card to reset without a reboot. This is a common occurrence in both Linux and Windows.
Ok, you've got me there. There is a lot of cheap, crappy hardware on the market with marginal reliability. It is generally difficult to tell the difference between unreliable hardware and unreliable OS software, but when one has a reliable OS, the unreliable hardware sticks out more. I do have to say, though, that simple operations like changing video resolution or from VGA text to graphics modes on my Linux machine is not an operation that causes system unreliability even though it is a more technologically intensive operation that what is possible under Windows (ie. virtual terminals with a mix of text and graphical heads).
4. 99% of the time there's a problem with a game or application, the response is "Do you have the latest video card drivers?" They seem to be released stealthily every few weeks. Who wants to deal with it? And whenever you upgrade there's a high probability of trouble with older software. See #1.
This is both a symptom of time-to-market pressure at the vidcard manufacturer and the poor GDI/display driver implementation in Windows. With XFree86 and its drastic seperation of video display and application, drivers are generally more stable, update less frequently and don't have application specific glitches. On my desktop at work I am still running the same copy of X that was started when I booted the machine last Oct, eg. 879 ? S 12033:14/etc/X11/X vt7 -auth/var/lib/gdm/:0.Xauth:0 . For games the only real requirement is an accelerated OpenGL and any card will work fine. Oh, and with X11, backwards compatability is pretty much a non-issue, in fact some of the standard X apps are pretty old and haven't needed any serious changes in years (5+).
Again, I think that your background (like mine 8^) has significantly colored your judgement, causing you to mistake endemic OS design and implementation issues with actual, hard, technology problems.
What was being compared was MkLinux (ie. Mach/Linux) and Darwin (ie. Mach/BSD). They are both designed by Apple and both use Mach. Mach has been known to be a slow and crappy microkernel implementation, but that doesn't seem to have impacted it's popularity very much.
On the other hand, BeOS uses its own microkernel, designed from the ground up for speed. This would be similar to other microkernel based OS's like NT and QNX (IIRC). At least for BeOS and QNX, nobody complains about kernel speed.
Re:Caldera's genius in buying DR Dos
on
Lineo Frees CP/M
·
· Score: 2
That's not entirely true. Calderadid purchase the assets of Digital Research from Novell, and begain a lawsuit against Microsoft, and settled for ~$250M. They must have needed the money badly because they settled, even though their case seemed very strong and an eventual win almost guaranteed. During this time Lineo (nee Caldera) did sell embedded solutions based on the DR-DOS code base. They also purchased the Arachne web browser for DOS, ported it to their Linux offering and sold it as DR-WebSpider. At the time they sold both DR-DOS and Linux based embedded packages, targeting the Kiosk market. They also made the source to DR-DOS (Caldera OpenDOS) available for the first release or two but closed it back up due to lack of interest, the difficulty of getting the build environment setup and business reasons.
DR-DOS lives on as the bootstrap for Novell Netware and I'm sure that there were a few other clients for embedded DOS (IIRC Kavouras used it, I can't remember others). DR-DOS, AFAIK, is still available for download and personal use, and Caldera has packaged it for use with DOSEmu. So while they did use DR-DOS for the lawsuit money (A perfectly valid and appropriate lawsuit if there ever was one) they also based the beginnings of their embedded offerings on it. Lineo is one of the better embedded companies right now, gunning for Wind River's marketshare, they are not going away.
Eh, you're probably right. For the first couple of years of my enslistment I was working out of an Army base in Germany and that sucked. This was circa 1997 and we were still using Unisys i386 machines (souped up!) for the management's desktops. Unisys made a great machine, built like a tank and very expandable but it was still a 386-25 (running Win95, eek!). My next assignment was to a real Air Force Base but the IT shop was as I described, understaffed and clueless, spending all their time trying to piss out 4 alarm fires. It was total chaos. We used some Navy products in our job and they were pretty spiffy, we definately preferred them to our AF equivilants. Maybe I just have a "pasture is greener" syndrome 8^)
I'm an OpenSRS reseller as well, we have about 300 domains hosted for our various clients. The few times that I've had to talk to their tech support they have been friendly, helpful and actually know their product. I think that OpenSRS is still a small shop, and their tech support people actually know the developers personally. I can't think of the last time that they have had an unscheduled service outage and their web CGI's seem to be well written and feature-complete.
I highly recommend finding an OpenSRS reseller and get rid of the useless dirtbags over at netwanksolutions.com
While I was in the Air Force and liked many things about the experience (I think that they are the least obnoxiously uptight of the services) what little of the Navy IT system impressed me. I know everyone is going to point out the NT-on-a-ship thing but that story is mostly bogus anyway.
Anyway, what little I had to deal with Navy-built systems I found several examples of them being better designed (from an IT perspective) than our own Air Force stuff. My experience with the AF IT system as an ADPE manager was pretty horrible. The people I met weren't very bright or well trained, constantly understaffed (they only had 12 people, including networking, server, desktop and manager people trying vainly to support over 1500 desktop systems all over base) and had very poor tools (who's bright idea was basing the entire IT infrastructure on MS Outlook and NT (even for secure messaging) anyway!?!!!). The worst part was that they were completely incapable of supporting desktop Windows users, they didn't even have a helpdesk, and everyone had to fend for themselves (including software and hardware purchasing.) Unless things have changed you would find the constant disaster very frustrating.
Oh, and I know everyone already told you this but Don't Trust Recruiters. When they aren't lying to you because they honestly don't know something they are lying to hide the ugly truth. If they tell you that they can't get your name into a slot for a certain job, implore you to select "Open/General" as a carreer and hope for the best don't sign! Get it in writing that your name is in a slot for the job you want before you sign a commitment. Don't cry too hard if the slot closes after you sign, shit happens, but at least they should try. Unfortunately the military isn't going to have you sitting on your butt waiting for annother slot to open up so you will probably be reclassed into something that they need at the time.
Well, since the subject was Netscape's failure to properly render CSS positioning, and since CSS wasn't in the HTML standards of "many years ago," you basically support my position. By the way, did you follow the link I gave [anybrowser.org]?
CSS1 has been a standard since
December
1996, CSS2 since May 1998. Don't give any crap about NS4 being older than the spec because it isn't really true. CSS didn't spring forth from the head of Berners-Lee, NS and MS both helped design the standard. At the time Netscape was pushing their proprietary HTML tags and CSS implementation and didn't bother to implement the standards. NS4 is intentionally broken.
I do blame Netscape for failing to implement the standards that they helped to create. It's very, very sad that browser implementations are 5 years behind their own standards, keeping web design in the dark ages. Blinking purple text on a blinking starfield background (with flames!!) here I come !!
Thanks, I'm using PINE 4.40 right now. I still don't find the filtering as useful as in Evolution, Mozilla, Netscape, Kmail, etc. Maybe I'm missing something but I don't see any way of specifying more than one match rule ( this _and_ that _but_not_ this other thing ) and/or more than one target rule ( autogenerate reply _and_ copy to folder foo ).
Luckilly I no longer require such precise filtering so PINE is working out fine.
I also was holding on to Netscape mail because I couldn't find a decent IMAP client for Linux. Several months ago I switched from NS Mail to PINE and have been pretty happy. PINE doesn't have the mail filtering support that NS does but it is by far the fastest and most efficient IMAP client I have ever used. Opening my Inbox folder with ~10,000 messages takes only a few seconds (server is a PII-300 IIRC) and the message finding/selecting/zooming feature is second to none. PINE seems to be highly optimized so that it issues the most efficient IMAP commands possible.
Even with current wiretapping laws capturing conversations of non-suspects is a problem. I seem to vaguely remember a scandal in the LAPD where they were illegally passing wiretap data that fell outside their warrent to Detectives who might find it interesting. This piece of legislation would make that problem worse.
Annother point is that to enact "roving" wiretaps you have to be able to rove. This implies that they will have automated and probably unmonitorred (in practice, maybe not according to the letter of the law) access to the phone system.
The legislation also seems to allow the ability to preemptively tap lines that a terrorist might use. Why don't you try proving to me that it would be impossible for a terrorist to use/borrow your phone. Remember that a terrorist is defined as anything the FBI wants it to be defined as. Basically this gives them the ability to tap any phone they want at any time with much less hassle.
Don't think about how this legislation might work in a perfect world, think about how it will fail in the real one.
Many deb packages, prolly including the ones on security.debian.org are GPG signed. You can get the Debian project's keyring, verify it and then know that the update you are getting is legit.
The only thing stopping it these days is Linux's smaller marketshare. (Worm propagation is one of those n squared problems).
That doesn't make the slightest bit of sense. Apache has 2 or 3 times the marketshare that IIS (including derivitives like PWS) has. An Apache exploit comperable to the IIS one would rip through the network like fire and could easily take down the majority of servers. One of the reasons that this hasn't happened is that Apache was coded far more carefully than IIS. Annother important reason is that Apache servers are not a monoculture. Apache runs on many different CPU arch's and many different OS's preventing the "One True Shellcode" from working (not that a worm coulnd't have a library of shellcode for many different platforms) Permissions on UNIX hosts tend to be slightly more sane out of the box as well, not great but better than the competition.
There have been several RedHat (not Linux or Unix in general) worms recently, but they just weren't that obnoxious (not that there weren't quite a few fire-and-forget RH6.x boxen around). RedHat isn't making the same mistakes again, RH7.x doens't turn on every installed service by default and optionally can setup firewalling rules that protect your machine from attack. Mandrake has a "make secure" button that does a pretty good job of locking a machine down and distro's like Debian try to err on the side of security whenever possible.
I'm rambling but Unix servser are generally more hostile to attacking worms than other environments. Any monkey can setup an IIS server but the results tend to be slipshod. Any monkey can setup an Apache server as well but the results tend to fair better when exposed to the open cesspool of the Internet.
Ok, I understand now. You are just not on the same plane of reality as the rest of the universe. I know it might hurt your head but you really should try thinking clearly and rationally some time. It is really much nicer when you don't believe that knowledge is out to "control" you.
The gist is this: If the road exists for the benefit of the big trucks, maybe we can ride our bikes on it. But we'll need to be wary of getting run off the road or turned into some form of roadkill. Look at the current Copyright fuss, and consider the incidental damage that may be done to peer2peer, simply because the RIAA thinks it's ALL about pirating music. Look at people accused of DMCA getting booted off of their ISPs, simply on suspicion.
Ok, that makes more sense. Right now we have huge problems with dynamic IP assignment, filtering of incoming traffic and absurd ToS agreements. This effectively, and maybe not intentionally, makes home computers "clients" and requires dedicated hosting to be a "server". This creates a situation where it is more difficult for a person to get web space that they fully control, if the data is stored on someone else's machine it is effectively under their control.
After reading the IPv6 article from the other day it seems that there is some hope. IPv6 seems to allow "roaming", much like cell phones you can transparently move your netblock between providers in around 60sec without losing any existing TCP connections. If your ISP doesn't work the way your want it is pretty easy to move to one who does. I just hope we can build it out the way we want to, one more step to Utopia 8^)
I use my IBM Model 8595 PS/2 for many things including Firewall, HTTP Proxy, NFS & Samba, DNS, VPN, DHCP, SMTP & IMAP, SSH, etc. It has ~27 GB of SCSI disks inside and outside of the frame. Not bad for a 486-50 with 64MB RAM.
It is easy to blame the victim--the People--for this failure to maintain vigilance over their freedoms. For my money, though, it is just as easy--and far more accurate--to blame the very foundation that allows us to have this conversation--the tech itself.
I'm sorry, could you say that again??!! You aren't going to blame the people but you are going to blame an inanimate object instead. How convenient. Like it has been said, the Internet will not be what you want it to be but what you make of it.
I'm not sure I know what you are talking about with your bicycles and 18-wheelers. The metaphor is not the argument, maybe you could explain using concrete terms.
While I wouldn't buy Trident for performance but they are traditionally very reliable. The people I know have built OEM whitebox computers in the past know which vendors are reliable and which aren't, they prefer Trident based boards because they are cheap and people don't have to return them. Other, whizzier, video chipsets tend to have more wierd issues with particular software or they overheat and die, but the Trident based boards didn't have these problems.
Hmm, I've been reading Slashdot for 3-4 years and I don't remember a time when it was anything other than CmdrTaco's personal plaything. You're right about one thing, stories don't often break here anymore, at least not before I am aware of them through other channels. That might just be different reading habits on my part, though.
I hear ya, brother. I had to let my PC Gamer subscription lapse 2-3 years ago because I just couldn't stand their crap anymore. In that timeframe they've had almost 100% turnover (everyone except for the RPG and Wargamer reviewers, last I checked.) They tried to keep an upbeat tone but it was obvious that it was forced. At one time their ads were handled seperately but as time went on it became obvious that they were editing the mag so that they could run adverts on the opposite page as the review.
I used to read boot as well but they started going south at around the same time. They changed their name to MaximumPC and were never quite the same again. As boot they would review nifty things like the BeBox, but as MaximumPC they would just pump out annother review for the latest whitebox Voodoo3. *gack*
Slashdot Mirror
I have two points with your post:
1) You're right that you can't delete tickets out of RT, but I don't understand why you'd want to. Just because an issue is dead, does't mean that all records of it should be purged from the database. "killed" is just another status, like open, stalled and resolved and just shows that this issue no longer requires action.
2) Your performance problems are totally whacked and don't correspond at all to my experience. I've seen RT 1.x systems with more than 500,000 tickets in them that don't have anything like the performance problems you're describing. Searches through the tickets only takes ~10s on moderate hardware on this system (with more than 500,000 tickets). Have you done something silly like not turned indexing on on the important tables?
Well I often stream from the Japan-A-Radio shoutcast stream. JAR has a store where you can purchase J-Pop and Anime music. Other than that my local used video game store carries a pretty fair selection of J-Pop and Anime at normal ($15) prices.
Say Hi to everyone in the Japan-A-Radio chatroom at #japan-a-radio on irc.d00mnet.com and most other IRC networks.
I believe that you are mistaken. On an IBM system if a processor smokes all the tasks that wer e running on it will be automatically moved to a redundant spare and execution will resume from the last known good state. The Linux image would continute processing normally and would be unaware that anything at all had happened. I also believe that the machine would generate an exception alert for the local admins, as well as call IBM Global Services directly and report the problem so that a tech can already have been dispatched by the time the customer calls reporting a problem.
All the documentation that I have read indicates that IBM VM is significantly more advanced than Sun partitions. All the whining from Sun won't change that anytime soon, only several years of hard engineering will close the gap.
Linux on a mainframe can be great if you have a computing task that matches what mainframes are good at. It can also be good if you have an existing mainframe and wish to get more use out of it.
If I could add, have an error number along with the error message. For example instead of having a box pop up saying "You can't do this here", have it say "Error: ABC1234: You can't do this here". It is much easier to search by the error number in the online help and on Google then by the text of the message. This goes double when you app is localized in different languages.
While not exactally the information you were looking for, there is a perfectly good LaTeX macro set for making scripts. This many be a good base for creating an XML Scheme, knowing what tags you will need and such.
I would like to take exception to most of the points that you have made. They are not base problems with the technology, they are mainly problems endemic to DOS/Windows. Even running on the exact same hardware, most other OS's will not have these problems.
The software component of your complaint is mainly a package management problem. Windows systems traditionally have terrible/non-existant software management tools and a very high level of interdependancy of software components that is usually completely non-obvious and opaque to both user and admin alike. To answer to the hardware part of your question this is generally just a Windows problem as well. For Mac's, hardware is easy (see the AC's post). On my desktop machine, all relevant drivers are included with the base OS and it has limited detection of new/changed hardware. I recently built a new Athalon MP system to replace my aging Pentium 233MMX system and merely moved the harddrive from one system to the other. It autodetected my new MB and NIC without any effort required on my part, changing my XF86Config-4 file consisted of adding a new Device section and changing one line. Less than 5 minutes effort all told. At no point did I have to scrounge on floppies/cdrom for drivers or have to update from the manufacturers website to get things working up to spec.
This has been hashed over again and again and I don't believe it to be true. Windows has a really bass-ackwards method for loading programs and in 9x versions absolutely no useful memory protection or system security. Last I checked NT based systems require many things to be world-writable, although SFP helps a little. Systems such as Linux have a completely different method for determining whether a program is loadable or not which significantly raises the bar for infection (ie. way higher than a Y/N popup from your web brower or email client) and can insure greater integrity of the system so that errors like rm -rf are more easily recoverable. With very little admin effort (which may be done by the vendor for you) you can make your system nearly invulnerable to everything short of a buffer overflow in your browser or MUA (and header and MIME parsing seems to be the only automatically exploitable vulnerability possible in an MUA). Anyway on a better designed system, without so many endemic design flaws, viruses have a much harder time infecting the system (propagating should be just as easy). Also there are even better security architectures that could be made default at any time (LSM w/Flask, SubDomain, Janus, POSIX, etc.) that could make the network a very hostile place for viruses. I predict that there will be no requirement for antivirus softare on Linux systems
Ok, you've got me there. There is a lot of cheap, crappy hardware on the market with marginal reliability. It is generally difficult to tell the difference between unreliable hardware and unreliable OS software, but when one has a reliable OS, the unreliable hardware sticks out more. I do have to say, though, that simple operations like changing video resolution or from VGA text to graphics modes on my Linux machine is not an operation that causes system unreliability even though it is a more technologically intensive operation that what is possible under Windows (ie. virtual terminals with a mix of text and graphical heads).
This is both a symptom of time-to-market pressure at the vidcard manufacturer and the poor GDI/display driver implementation in Windows. With XFree86 and its drastic seperation of video display and application, drivers are generally more stable, update less frequently and don't have application specific glitches. On my desktop at work I am still running the same copy of X that was started when I booted the machine last Oct, eg. /etc/X11/X vt7 -auth /var/lib/gdm/:0.Xauth :0
879 ? S 12033:14
. For games the only real requirement is an accelerated OpenGL and any card will work fine. Oh, and with X11, backwards compatability is pretty much a non-issue, in fact some of the standard X apps are pretty old and haven't needed any serious changes in years (5+).
Again, I think that your background (like mine 8^) has significantly colored your judgement, causing you to mistake endemic OS design and implementation issues with actual, hard, technology problems.
Have Fun!
What was being compared was MkLinux (ie. Mach/Linux) and Darwin (ie. Mach/BSD). They are both designed by Apple and both use Mach. Mach has been known to be a slow and crappy microkernel implementation, but that doesn't seem to have impacted it's popularity very much.
On the other hand, BeOS uses its own microkernel, designed from the ground up for speed. This would be similar to other microkernel based OS's like NT and QNX (IIRC). At least for BeOS and QNX, nobody complains about kernel speed.
That's not entirely true. Caldera did purchase the assets of Digital Research from Novell, and begain a lawsuit against Microsoft, and settled for ~$250M. They must have needed the money badly because they settled, even though their case seemed very strong and an eventual win almost guaranteed. During this time Lineo (nee Caldera) did sell embedded solutions based on the DR-DOS code base. They also purchased the Arachne web browser for DOS, ported it to their Linux offering and sold it as DR-WebSpider. At the time they sold both DR-DOS and Linux based embedded packages, targeting the Kiosk market. They also made the source to DR-DOS (Caldera OpenDOS) available for the first release or two but closed it back up due to lack of interest, the difficulty of getting the build environment setup and business reasons.
DR-DOS lives on as the bootstrap for Novell Netware and I'm sure that there were a few other clients for embedded DOS (IIRC Kavouras used it, I can't remember others). DR-DOS, AFAIK, is still available for download and personal use, and Caldera has packaged it for use with DOSEmu. So while they did use DR-DOS for the lawsuit money (A perfectly valid and appropriate lawsuit if there ever was one) they also based the beginnings of their embedded offerings on it. Lineo is one of the better embedded companies right now, gunning for Wind River's marketshare, they are not going away.
Further DR-DOS history links
Eh, you're probably right. For the first couple of years of my enslistment I was working out of an Army base in Germany and that sucked. This was circa 1997 and we were still using Unisys i386 machines (souped up!) for the management's desktops. Unisys made a great machine, built like a tank and very expandable but it was still a 386-25 (running Win95, eek!). My next assignment was to a real Air Force Base but the IT shop was as I described, understaffed and clueless, spending all their time trying to piss out 4 alarm fires. It was total chaos. We used some Navy products in our job and they were pretty spiffy, we definately preferred them to our AF equivilants. Maybe I just have a "pasture is greener" syndrome 8^)
I'm an OpenSRS reseller as well, we have about 300 domains hosted for our various clients. The few times that I've had to talk to their tech support they have been friendly, helpful and actually know their product. I think that OpenSRS is still a small shop, and their tech support people actually know the developers personally. I can't think of the last time that they have had an unscheduled service outage and their web CGI's seem to be well written and feature-complete.
I highly recommend finding an OpenSRS reseller and get rid of the useless dirtbags over at netwanksolutions.com
While I was in the Air Force and liked many things about the experience (I think that they are the least obnoxiously uptight of the services) what little of the Navy IT system impressed me. I know everyone is going to point out the NT-on-a-ship thing but that story is mostly bogus anyway.
Anyway, what little I had to deal with Navy-built systems I found several examples of them being better designed (from an IT perspective) than our own Air Force stuff. My experience with the AF IT system as an ADPE manager was pretty horrible. The people I met weren't very bright or well trained, constantly understaffed (they only had 12 people, including networking, server, desktop and manager people trying vainly to support over 1500 desktop systems all over base) and had very poor tools (who's bright idea was basing the entire IT infrastructure on MS Outlook and NT (even for secure messaging) anyway!?!!!). The worst part was that they were completely incapable of supporting desktop Windows users, they didn't even have a helpdesk, and everyone had to fend for themselves (including software and hardware purchasing.) Unless things have changed you would find the constant disaster very frustrating.
Oh, and I know everyone already told you this but Don't Trust Recruiters. When they aren't lying to you because they honestly don't know something they are lying to hide the ugly truth. If they tell you that they can't get your name into a slot for a certain job, implore you to select "Open/General" as a carreer and hope for the best don't sign! Get it in writing that your name is in a slot for the job you want before you sign a commitment. Don't cry too hard if the slot closes after you sign, shit happens, but at least they should try. Unfortunately the military isn't going to have you sitting on your butt waiting for annother slot to open up so you will probably be reclassed into something that they need at the time.
CSS1 has been a standard since December 1996, CSS2 since May 1998. Don't give any crap about NS4 being older than the spec because it isn't really true. CSS didn't spring forth from the head of Berners-Lee, NS and MS both helped design the standard. At the time Netscape was pushing their proprietary HTML tags and CSS implementation and didn't bother to implement the standards. NS4 is intentionally broken.
I do blame Netscape for failing to implement the standards that they helped to create. It's very, very sad that browser implementations are 5 years behind their own standards, keeping web design in the dark ages. Blinking purple text on a blinking starfield background (with flames!!) here I come !!
Thanks, I'm using PINE 4.40 right now. I still don't find the filtering as useful as in Evolution, Mozilla, Netscape, Kmail, etc. Maybe I'm missing something but I don't see any way of specifying more than one match rule ( this _and_ that _but_not_ this other thing ) and/or more than one target rule ( autogenerate reply _and_ copy to folder foo ).
Luckilly I no longer require such precise filtering so PINE is working out fine.
I also was holding on to Netscape mail because I couldn't find a decent IMAP client for Linux. Several months ago I switched from NS Mail to PINE and have been pretty happy. PINE doesn't have the mail filtering support that NS does but it is by far the fastest and most efficient IMAP client I have ever used. Opening my Inbox folder with ~10,000 messages takes only a few seconds (server is a PII-300 IIRC) and the message finding/selecting/zooming feature is second to none. PINE seems to be highly optimized so that it issues the most efficient IMAP commands possible.
Even with current wiretapping laws capturing conversations of non-suspects is a problem. I seem to vaguely remember a scandal in the LAPD where they were illegally passing wiretap data that fell outside their warrent to Detectives who might find it interesting. This piece of legislation would make that problem worse.
Annother point is that to enact "roving" wiretaps you have to be able to rove. This implies that they will have automated and probably unmonitorred (in practice, maybe not according to the letter of the law) access to the phone system.
The legislation also seems to allow the ability to preemptively tap lines that a terrorist might use. Why don't you try proving to me that it would be impossible for a terrorist to use/borrow your phone. Remember that a terrorist is defined as anything the FBI wants it to be defined as. Basically this gives them the ability to tap any phone they want at any time with much less hassle.
Don't think about how this legislation might work in a perfect world, think about how it will fail in the real one.
Many deb packages, prolly including the ones on security.debian.org are GPG signed. You can get the Debian project's keyring, verify it and then know that the update you are getting is legit.
That doesn't make the slightest bit of sense. Apache has 2 or 3 times the marketshare that IIS (including derivitives like PWS) has. An Apache exploit comperable to the IIS one would rip through the network like fire and could easily take down the majority of servers. One of the reasons that this hasn't happened is that Apache was coded far more carefully than IIS. Annother important reason is that Apache servers are not a monoculture. Apache runs on many different CPU arch's and many different OS's preventing the "One True Shellcode" from working (not that a worm coulnd't have a library of shellcode for many different platforms) Permissions on UNIX hosts tend to be slightly more sane out of the box as well, not great but better than the competition.
There have been several RedHat (not Linux or Unix in general) worms recently, but they just weren't that obnoxious (not that there weren't quite a few fire-and-forget RH6.x boxen around). RedHat isn't making the same mistakes again, RH7.x doens't turn on every installed service by default and optionally can setup firewalling rules that protect your machine from attack. Mandrake has a "make secure" button that does a pretty good job of locking a machine down and distro's like Debian try to err on the side of security whenever possible.
I'm rambling but Unix servser are generally more hostile to attacking worms than other environments. Any monkey can setup an IIS server but the results tend to be slipshod. Any monkey can setup an Apache server as well but the results tend to fair better when exposed to the open cesspool of the Internet.
Blargh
Ok, I understand now. You are just not on the same plane of reality as the rest of the universe. I know it might hurt your head but you really should try thinking clearly and rationally some time. It is really much nicer when you don't believe that knowledge is out to "control" you.
Ok, that makes more sense. Right now we have huge problems with dynamic IP assignment, filtering of incoming traffic and absurd ToS agreements. This effectively, and maybe not intentionally, makes home computers "clients" and requires dedicated hosting to be a "server". This creates a situation where it is more difficult for a person to get web space that they fully control, if the data is stored on someone else's machine it is effectively under their control.
After reading the IPv6 article from the other day it seems that there is some hope. IPv6 seems to allow "roaming", much like cell phones you can transparently move your netblock between providers in around 60sec without losing any existing TCP connections. If your ISP doesn't work the way your want it is pretty easy to move to one who does. I just hope we can build it out the way we want to, one more step to Utopia 8^)
I use my IBM Model 8595 PS/2 for many things including Firewall, HTTP Proxy, NFS & Samba, DNS, VPN, DHCP, SMTP & IMAP, SSH, etc. It has ~27 GB of SCSI disks inside and outside of the frame. Not bad for a 486-50 with 64MB RAM.
I'm sorry, could you say that again??!! You aren't going to blame the people but you are going to blame an inanimate object instead. How convenient. Like it has been said, the Internet will not be what you want it to be but what you make of it.
I'm not sure I know what you are talking about with your bicycles and 18-wheelers. The metaphor is not the argument, maybe you could explain using concrete terms.
While I wouldn't buy Trident for performance but they are traditionally very reliable. The people I know have built OEM whitebox computers in the past know which vendors are reliable and which aren't, they prefer Trident based boards because they are cheap and people don't have to return them. Other, whizzier, video chipsets tend to have more wierd issues with particular software or they overheat and die, but the Trident based boards didn't have these problems.
Hmm, I've been reading Slashdot for 3-4 years and I don't remember a time when it was anything other than CmdrTaco's personal plaything. You're right about one thing, stories don't often break here anymore, at least not before I am aware of them through other channels. That might just be different reading habits on my part, though.
I hear ya, brother. I had to let my PC Gamer subscription lapse 2-3 years ago because I just couldn't stand their crap anymore. In that timeframe they've had almost 100% turnover (everyone except for the RPG and Wargamer reviewers, last I checked.) They tried to keep an upbeat tone but it was obvious that it was forced. At one time their ads were handled seperately but as time went on it became obvious that they were editing the mag so that they could run adverts on the opposite page as the review.
I used to read boot as well but they started going south at around the same time. They changed their name to MaximumPC and were never quite the same again. As boot they would review nifty things like the BeBox, but as MaximumPC they would just pump out annother review for the latest whitebox Voodoo3. *gack*