OSS is the way. Well the proper licensed stuff is anyway.
(Parent...)
Oh, now that was just plain funny. It seems that/. alone posts a "Version 0.2.3.1.2 of GNUXKApp is out today" several times daily. RedHat has, what, at least a new major release every year for the past few years?
(And now me... *smile*)
The point is that no one can stop you from
using Free software in whatever manner you want to. If you love Linux 1.2.13 you can keep right on using it for as long as you want. You can contract out for firms to add drivers or fix bugs as much as you want. If you want to pay somebody to backport IPtables or Usermode Linux to 1.2.13, go for it!
Remember USB devices that used WDM drivers that say "Requires Windows 98". There's no reason why WDM drivers couldn't be made to work under Windows 95-- except that Microsoft didn't want to do that work when you could just pay them more and get Windows 98. That's just fine, too! It's their code, so that's their right. Want to add that support yourself? Too bad-- you don't have the code, and it probably violates a license anyway. Think "Group Policies" and Windows NT 4.0 (if you've never noticed, Group Policies are implemented mainly by a tweaked-up USERINIT.EXE), or perhaps FAT32 and Windows 95 OSR2. I'm picking on Microsoft a bit unfairly, 'cuz there are other manufacturers that are more flagrant about it-- but it's their code, so it's their right, and you're stuck "on the treadmill" because you chose to use their software.
Free software isn't anybody's code, though. You can add whatever you want-- or hire somebody else to do it for a fair and equitable rate.
"Upgrades" don't "have to" happen. These "forced upgrade" cycles are a symptom of the idiotic "commerical software industry" believing that they are somehow both manufacturing and service companies-- all at the same time! Use and contribute to Free software, and get yourself off the treadmill if you don't like it.
The unwavering constant in my world of IT consultation work is the assured shittiness of the Customer's line-of-business application. Either it's an off-the-shelf app. that they're pushing beyond its functional limitations, or it's some home-grown bag of dung that coddles their entrenched antique business processes and reeks of inconsistent user interface, poor or completely lacking forethought in design, and lamentable "technologies" (everybody say "toolbars of icons with no tooltips and no menus", "two digit decimal date fields" and "shared file database"). In the end, it really doesn't matter how they've chosen their IT fate, it always ends in everybody bitching about how bad it all is!
The idea of defining requirements and selecting off-the-shelf packages based on those requirements seems to be completely foreign to non-technical users ("I have three (3) kids and a dog-- I think that two-seater little red convertible sportscar will make a good family car!"). Of course, software marketing would have you believe that their products will allow you to travel backwards in time and transmute gold from pocket lint, as long as you keep up on your "maintenance agreement".
On the "internally developed" side, the failings I see almost always involve an inability for a development group to shut the fuck up about their fucking "technology" and learn about the users' BUSINESS REQUIREMENTS! They users aren't going to get any benefit from your buzzword-fortified J2EE-compliant mobile wireless XML fibre-channel attached pneumatic Bluetooth ass-rampager if they never USE the damned thing because it doesn't satisfy any of their business requirements.
We don't have the fucking computers in the business because we just want to have computers! We are here to make fucking money, and the computers are tools to help us.
Doesn't really matter much to me, though... They'll all still need switches, routers, and infrastructure gear, whether they ever get it together or not... *smile*
The "boring router switchy things" pic appears to show two (2) Cisco Catalyst 6513 chassis with dual-redundant supervisor modules. Yeesh... Depending on the options, there's another $200K in gear right there.
How can this company be doing well enough to afford this gear, yet be dumb enough to let their people "case mod" the E15K's?
I'm sorry that I don't have anything positive to say. I just didn't feel right about letting this kind of misinformation get out there unchallenged.
You're being a little too harsh. I agree that the mentality of "Oh, throw some PHP and MySQL at it" is unrealistic-- even idiotic, at best-- but medical billing is not "rocket science", from a programming standpoint.
(I've worked "integrating" medical, dental, and eyecare billing software into many practices. It's just software. If anything, it's software that actually DOES very little.)
The hardest part I've seen, from the standpoint of watching Medisoft and others modify their software (over and over and over and over) appears to be getting the software to work flexibly enough so that offices "entrenched" in their antique "paper-based processes" are able to function with some degree of efficiency.
Have you ever looked at the drivel that passes for billing software for small offices? All of Medisoft's DOS versions were just flat-file ASCII! Their new "client server" 'doze versions are based on the Extended Systems "Advantage" database engine-- but it's still not that complicated. I haven't seen a package for the dental or optometric field that's much more complicated (albeit most of those apps bring charting into the mix, too).
For a small office, the "technology" isn't the issue. I would agree with your statement that medical billing is a "science in and of itself"-- if only because it's been made that way by idiotic regulation and the insurance industry-- but those are problems that need to be solved by a human operator-- not by a software application. The software part of the job just isn't that big of a deal in a small office. (Hospitals, very large practices-- there's a difference there).
Some of this is probably my angst built-up from actually working with people who administer and operate billing and coding operations. I've met people that were, effectively, trained chimps with a certificate from a community college that could just assure me that I couldn't possibly grasp the science that is medical billing-- as though it's simply beyond the capability of mortal men to understand. It doesn't, however, seem to be beyond the capability of middle-aged fat women in flowery gowns and white shoes.
(It's times like these that a good ol' "salt the earth" Free software project sounds like such fun!)
You cannot boot Windows NT 3.51, NT 4.0, W2K, or XP w/o a video driver. They just won't do it, "out of the box".
The "embedded" versions of Windows NT (NT Embedded 4.0, the W2K SAK, or XP Embedded) contain a "Null VGA" driver that allows them to boot w/o a video card (also contain keyboard and mouse drivers that do the same). Take a look here for a marketroid overview, or here for a FAQ on NT Embedded 4.0 (Search for "Null" if you want to see the "headless operation" part-- I find it particularly interesting that they have an option for a serial console). (These "null" drivers were acquired from a 3rd party by Microsoft. I've looked at the DDK, and actually, it wouldn't be all that hard to implement a "null" driver for those services... maybe I ought to think about that.)
For those who are development minded, it's worth it to take a look around the NT "embedded" pages at Microsoft. I really like the NT kernel (not the bag-of-crap that is Win32, mind you), and I like that Microsoft is letting you strip out the crap "features" and run "bare metal" on the native-mode kernel if you want.
(I've always been convinced there's a Unix inside NT trying to get out... *grin* I'd love to roll an "NT distribution" that stripped out the GUI, and booted the Interix subsystem w/o Win32... Oh, well... I guess I'll just stick to tinkering w/ Linux.)
Even in our own office, I wish we could kill multicasting. We make games here, and in the evening, a lot of groups of guys are playing games that spew enough multicast packets to bring our 100mbit network to its knees. (Yes, we're using switches, not hubs.)
You need switches that support IGMP spoofing, then. If the traffic really is multicast, and not broadcast, IGMP spoofing is what you're looking for.
<KARMA-WHORING>
Take a look here-- it's a pretty decent article on IP multicast (not really technical-- just buzz-laden). Some rudimentary technical information is in here.
</KARMA-WHORING>
This isn't a good idea. I really wish I could get my technicians to stop doing it. There are items there that will never "re-appear" magically. Take a look under HKEY_LOCAL_MACHINE\Enum\SW. Those items won't "re-detect".
You would do better to boot in safe mode, and hand remove everything from the "Device Mangler" except for the software media items ("Microsoft Streaming Clock Proxy", "Microsoft Streaming Service Proxy"), the "IO read data port for ISA Plug and Play enumerator", and "Plug and Play Software Device Enumerator". Leave those alone, and reboot. Let it detect (it may complain that your video adapter is mal-configured-- that's a good sign-- let it detect). It might take a couple boots. Load your INF updates, drivers, etc, and you should be all set.
I've done a reasonable quantity of backup-solution deployments, from the simple "tape drive in a server" to multi-element DLT libraries. I've had customers "invent" a version of this idea on many occasions. Typically, the customer's "invention" takes the form of one of several similar ideas.
What it comes down to, though, is that the idea behind having multiple medias, stored _away_ from the production copy of the data, is a good thing. Until recently, this has only been really convenient with tape media. With the advent of very convenient hot-swappable hard drive carriages and support for hot swapping of hard disk media in nearly every commonly used operating system, I don't see why hard drives could not be used-- but they would need to be treated with a little more physical care than tapes.
The "problem" seems to come when the (typically small-business) customer "invents" this idea, buys one of those cruddy "centronics connector on the back" sub-consumer-grade plastic "drive bays", slaps a hard drive in it, and starts doing backups to one hard drive from another. The cycle is something like: (1) insert 2nd hard drive, (2) wipe 2nd hard drive, (3) copy contents of production hard drive(s) to 2nd hard drive, (4) remove 2nd hard drive. They don't think about what would happen if, say, between steps 2 and 3 the production hard drive(s) failed.
If you're going to use hard disks as "tapes", I don't think there's anything fundamentally wrong-- but buy the same number of hard disks as you'd buy tapes-- and rotate them in the same manner. Treat them as large, mechanical tapes. Keep them away from the production data except when in use.
Looks like there are SOME interesting uses for these old Metricom modems in Linux. STRIP looks pretty interesting. Still, I'm seeing articles like this that make me think that the newer, faster modems are going to be useless.
Anybody have any more info relating to using these new GS models in peer-to-peer? I just bought a GS model, and I'm looking to get another. I'm seeing people posting on here that the modems are trash now-- but I'm not so sure that's the case.
Get a motherboard with 4 IDE channels (most "raid versions" have this) and plug 12 drives into the Hotrods, 4 into the motherboard's raid channels, and 2 into the secondary ide channel. The boot hard drive goes on the primary ide channel.
Ouch! Spend a few hundred more, and get an Escalade Storage Switch. They perform very well and aren't wildly expensive (you should be able to have an 8-channel 32-bit, 33Mhz version for under $500.00). You also have your motherboard IDE channels free for things like DVD-ROM drives... heh heh... Lots of DVD-ROM drives... Heh heh...
Ahh, yes-- and there are Linux drivers available for the Escalade controllers. If you're looking for wild amounts of performance, they do have a 66Mhz, 64-bit PCI version available, too. Wowza.
Promise has their SuperTrak controller, which looks very interesting, but based on some messages I saw flying around on the Kernel List, apparently it's not as straightforward as just compiling in I2O support to use it under Linux. Grrr...
Check out this review and this review if you want to see how the Escalade stacked up to other "high end" IDE RAID controllers.
A Cisco 6509 is not a cheap piece of hardware (yours was at least $30K w/ what you've described in it). Especially if you've got redundant supervisor engines, and doubly so if you're doing layer-3 in the box. That having been said, why couldn't you have hired a competent technician to install the box properly when you installed it in the first place, rather than having a half-assed configuration loaded?
I doubt your setup would've been more than three or four hours of configuration, based on what you've described, and you'd have gotten decent documentation out of all of it, if you'd hired a good technician. It's quite obvious that one visit by a technician at $250.00/hr would've more than made up for the cost of the downtime and headaches you incurred as a result of having a poor configuration in the first place.
Putting a man into an amateur space vehicle is, IMHO, probably a bad idea. I think it's highly do-able for amateurs to put vehicles into orbit. I'm a "high power model rocketry" fanatic, and I have a lot of respect and appreciation for what can be accomplished with simple solid-fuel boosters. There are orders of magnitude difference, however, between the complexity of a small solid or liquid fueled booster that can achieve tens of thousands of feet of altitude, and a booster capable of throwing a payload into a stable orbit.
I believe the technology necessary to put men into space is not beyond the grasp of private individuals or small corporations. Look at the advances in technology since the 1960's and the "golden age" of the U.S. space program-- it should be possible to do the same things the Mercury or Gemini programs did for a fraction of the research and development cost.
Having civilians participate in the space program is the best way to build enthusiasm for the space program-- but I don't think it wise for private individuals or corporations to go trying to loft men into space just yet... There has not been enough unmanned testing of vehicles yet to assure that launching these things will be safe for the men onboard, or those of us on the ground!
Maybe your working an 80 hour week becasue you are using the wrong tools or trying to implement the wrong solutions, etc...
I would agree that a measurable number of 80-hour weeks are caused by just what you say. I would also make the statement, though, that in many companies, it is expected that IT employees are expected to do far more work than employees in other capacities. That's not to say that there aren't a significant quantity of do-nothing IT personnel in the world-- but there are definitely a significant quantity of IT personnel that are expected to support (with no regard for the unreasonable time commitments, often) systems that they might not have implemented.
Ultimately, if the 80-hour a week job is a problem, quit. If the company's stupid decision-making process caused the "make work", quit. If being an "all Windows shop" causes the work, quit... or just kill yourself. Pick one.
When I started with this firm, I was a ~45-50 hour a week guy... That grew into the ~70-75 hour a week when I got placed on salary and had my "responsibilities increased commensurate to my salary".
Now that my wife has left me, and now that my responsibilities have been increased further, I work ~85-90 hours a week at this job (generally 6 or 7 days a week), and around ~12 a week at my "for fun" job... and I still have time to sleep, sometimes.
My wife leaving was a big bonus for the company, I guess-- I loathe going home and seeing her things the she has yet to come pick up, so I stay at the office and sleep on a couch in the back some nights.
Yeah-- I'd rather have my wife back than have this job-- but that's not gonna happen. Yeah, I made stupid fucking choices, and now I trapped in a shitty job w/ no future, alone. I know it's my fault, and I'm not really bitching. Hell, I like my ~80 hours in a slow week.
*grin* Sucks to be me, doesn't it... Being 23 is supposed to mean working for a dishonest slavedriver bastard over 80% of your waking hours. Of course, I still have time to read Slashdot.
Re:Optimism was based on over-simplistic model
on
A Map to Nowhere?
·
· Score: 1
The key is that the cell is an emergent system.
Moreover, the genome is wholy the product of evolution. We've seen, in a very simplistic way, the types of "optimizations" that evolution creates by using the genetic algorithm and such ilk to "evolve" programs.
It's silly to think that the genome is going to be anything but highly "optimized", incredibly "interdependent", and something that I can only express as "multidimensional". I'd liken reading and understanding the genome to disassembling highly optimized machine language back to assembler source code except I think that it's orders of magnitude more complex than that.
We are aware, in the case of a processor and machine language, exactly how all the interactions and processing occur. Though we have a lot of knowledge about protein folding and the spatial interaction of proteins, I daresay that we're not aware of 100% of the "rules", and will miss nuances in the "code" as a result.
This makes me think about the article posted a few days ago concerning the "evolved" FPGA systems, and the comment that the resulting systems were "strange", and were filled with structures that appear to "do nothing" and were not wired to surrounding parts of the systems. When these "do nothing" portions are removed, though, the system does not function properly-- showing the evolution worked to take advantage of all possible types of "interaction" within the substrate of the FPGA (quantum effects, possibly?).
I've often heard the observation that a large portion of the genome is "garbage" or is not used specifically for coding proteins. I would submit that we cannot dismiss any of the genome as "garbage", given that we aren't 100% sure of all the "rules" of the genetic system. Not at all unlike trying to understand the strange systems evolved on FPGA's.
When I wired my house last year, I decided to go ahead terminate every wallplate with dual RJ-45 jacks. All the cable is Cat-5, and terminates into a small patch panel in a closet. I brough my telephone lines into that closet, and bought a couple "1 to 5 way" phone splitters at a local discount store. When I need to route phone lines somewhere in the house, I patch from the phone splitter to the patch panel w/ RJ-11 patch cables. RJ-11's fit into RJ-45's fine-- albeit a little sloppily-- but they work.
I went the extra step of using gang outlets that support CATV terminations, as well. I brought my CATV feeds into the wiring closet, since I needed it there for the cable modem and firewall anyway-- so I have an amplifier there and panel to distribute CATV to the rest of the house. Someday I'll add an RF modulator and patch an external camera in on some channel... *grin*
I would agree w/ all the posters that say that cheap metal shelving makes for a nice rack. Paint it black and fill it with computer gear-- it looks pretty boss and doesn't need to be bolted to the floor to be stable!
If you're shopping around for a house, look for phone jacks in areas where you might want data jacks. Phone wire makes GREAT pull-wire for your new cables.
I HIGHLY recommend a patch panel. I've seen a lot of small offices and a couple homes that were just "crimped on RJ-45 ends" at the "wiring closet" end. This limits your flexiblity in regards to patching phone across the wires, and will make you very unhappy when you jerk on a wire, pull off a termination, and have to reterminate the wire. Patch panels are comparitively cheap, given that you want this wiring in place and functional for a long time.
Sounds like you want to build a mini "head end" for a CATV system and stick some PC's decoding MPEG streams on the end of it. You'll send commands to the video servers via Ethernet or some-such, and send the streams into RF modulators and onto the CATV lines.
I'm fuzzy on the RF modulator part (I've worked with using ONE RF modulator on a line-- but I know that there exist solutions that will allow you to pile multiple signals from several RF modulators onto the same wire)-- but that would really be the hardest part. We did a "closed circut" system in a school where I interned a few years ago-- but that was done over CAT 5 w/ some type of balun on each end and only one RF modulator in the mix.
I'd recommend foregoing hardware MPEG decoders and get some el-cheapo PC's that have enough horsepower to decode the streams you're looking to decode. Boot 'em diskless w/ a server hosting all the encoded files. Stick a video card w/ reasonable quality TV out on each PC, put 'em on an Ethernet switch w/ an uplink to your building LAN, plug 'em into the RF modulators, and you should be set to go.
I don't think you're going to find a cheap way to do this, unless you can come up w/ some cast-off RF modulators. You're talking about hardware that doesn't have a tremendous non-professional market (not many people run their own CATV system) so it's not cheap. Sounds like a fun project, though.
Have you used a sniffer to see what this "broadcast storm" traffic is? That would do for a good start. A description of the issue other than "kill the network" would do more good as well. What kind of byte/sec and packet/sec counts are we seeing on the media?
Assuming it is broadcast related: Your 'doze boxes need to be using "H-Node" name resolution for their cruddy NetBIOS name resolution. You need WINS servers. You need to disuse protocols that are broadcast intensive.
To the guys that are saying "install a switch"-- apparently there's an understanding issue w/ regard to what a switch does. A layer-2 switch won't help a bit in this case-- just like the poster said (referring to bridging routers). An analysis of what the traffic on the wire is would be a great first step. Then, intelligent decisions can be made to address the problem. Layer-3 switching might be a potential solution, depending on what the traffic is.
Unless you're using layer-3 entities inside of switches, your router based solution is going to do more than mess up NetBIOS name service-- it'll slow everything down to a crawl. Most low-end routers don't even come close to wire-speed.
I had a customer swearing to me that they were having "broadcast storms" because they were getting massive numbers of collisions on a shared-media LAN. We took a look at it w/ a sniffer and discovered that broadcasts played no part in it. They were doing large file transfers to a machine that was dual-homed on the same physical NIC, and the machine was thrashing packets on and off the wire, "routing" the packets to two hosts that were in different subnets, but on the same media. Duh.
Don't assume you know what your problem is unless you know what your problem is...
The previous run of this story is ref'd, but the discussion did center around use of codes on insecure channels during WWII. A good book, for the interested party, is Between Silk and Cyanide, by Leo Marks (ISBN 0684864223).
I'd have to echo the pessimistic sentiments of others in this discussion, though, and state that there's probably not a lot of hope in 'cracking' these transmissions, given that we have no knowledge of their origin or purpose.
This works well. We rolled out a butttload of Linux and Win95/98 machines like this. I've done 'recovery CD's' for Win95/98 w/ this method, too. (Gee-- isn't that misuse of a GPL'd operating system and tools... *grin*) If you can swing it, a bootable CD makes it even handier-- albeit if you have to get into the BIOS to make the machine boot off of CD, the time-savings is negligable over just booting off floppy.
As a more ambitious project, we made a little bootdisk w/ dhcpcd and support for the NIC's in the machines we were using, and used a little program called 'netpipe' to broadcast out.tar's to PC's hard disks over the wire. It _rocked_ to see 150+ machines pulling an image simultaneously at 800KB+ a second on garden variety 10Base-T.
First Neuromancer, then Snow Crash! NO!!! Books that I love, being put on the chopping block like this. Snow Crash has been an influence on my lifestyle. I can't even count the times I've stomped my brakes and cursed at 'fucking Bimbo boxes' on the highway...
Snow Crash (and Neuromancer, for that matter) was one of those books that FELT like a movie while I read it. Reflecting on it now, I feel like I've seen the perfect movie version already, because Stephenson's characters were so vivid. Anything they do with this story will disappoint me, and many others, I'm sure.
The one genre that I've thought would be able to handle films like Snow Crash (or Neuromancer) would be anime. Busting Snow Crash into 3 2-hour animated episodes would be great. I'd prefer it was voiced in English, but it'd look awesome as anime.
Oh, well... Who'd listen to me. I'd write a screenplay for an animated Show Crash, but having no artistic skills, I couldn't storyboard it. W/o storyboards, I don't think it's the kind of idea that would sell. Besides, the novel really doesn't need a lot of adaptation. (I can think of a few things that might need to be tamed a little for U.S. audiences... Y.T. and Raven's sex scene, potentially some violence.)
Picture it in your mind-- the stereotypical quickly-scrolling-background-art of many an anime production, w/ the magna-poon whipping thru the foreground and sticking on the back of a minivan... Picture it. Ahhh...
Maybe I'm just stuffy, but I'm concerned that the GNOME community is missing the boat on features that could legitimately make the dream of 'desktop domination' for GNU/Linux a reality. I'd certainly agree that user interface friendliness and a full-featured array of applications are the major factors in getting GNU/Linux onto corporate desktops, and I think competing with Microsoft OS's and Win32 apps on these points isn't the place when GNU/Linux can show the biggest win.
The 'centralized management' aspects of Microsoft operating systems and most win32 apps are horrendous. Expensive software 'hacks' like Tivoli, LANDesk, Novell ZenWorks, and Microsoft SMS are all ugly kludges to attempt to provide management features for large fleets of PC's. (I just can't say how much these hacks SUCK enough). The fact is, the entire 'single-computer, single-user' metaphor that win32 OS's and apps are based on is completely wrong for centralized management of fleets of PC's. Let's face it-- shit with icing on top is still shit.
As I look at the state of GNOME now, I'm not seeing the move toward management features that would bring serious money-saving features to the desktops of the corporate world. GNU/Linux being free is not enough for most IS managers to give up their MS-based ways-- the real investment is in management labor (all that 'Total Cost of Ownership' crap). Sure-- today a competent hacker could use scripts and some slight-of-hand to provide many remote management features, but what I'm talking about is providing these features to corporate IS managers in an easy-to-use system. Features like:
remote hardware inventory
remote application deployment
centralized configuration/settings management
software metering/use statistics
The only way I see to provide this correctly is to design it in from the start. Look at the MESS that is win32, and look at the leaps and bounds of hacks SMS, ZenWorks, and all the other 'management' frameworks go through. If we don't get to designing in these types of technologies NOW, we're going to end up with a hodge-podge of configuration repositories, labor-intensive application rollouts, and TCO similar to win32.
For GNU/Linux to offer serious labor savings to corporate IS managers, I'm talking about things like:
An abstracted configuration repository for applications (ODBC style connectivity to a back-end, either decentralized or centralized, much like Microsoft's Registry, but w/o all the brain-damaged stupidity and a GOOD SET OF DOCUMENTED RULES to follow for developers)
Standardized friendly application installation methods (which we have today in a large extent w/ RPM or Debian packages-but today's package management systems leave a bit to be desired, IMHO, when it comes to reporting installation error conditions to the user)
Some type of hardware inventory initiative that includes interfaces to back-end database systems for report generation (shouldn't really be very difficult-- again, I'm thinking ODBC-style connectivity)
Maybe we need a Linux Management Initiative or something of that ilk. I'd love to do it, if I was a competent enough coder and had enough time. Instead, though, I'm just a GNU/Linux user stuck working in a win32 based world and dying to get out of it. This stuff isn't glamorous, but I think it's a place that GNU/Linux can kick some serious Microsoft ass.
A coworker purchase one of the PC Power and Cooling "ultraquiet" PS's for his PC, so I've gotten to play w/ one a bit. Side-by-side a Sparkle Power ATX supply, it was definitely quieter-- but I wasn't terribly moved by the difference. Overall, I didn't think the extra cost was justified by the result.
The idea of moving the PC itself away and using extension cables is probably the best idea. The previous comment w/ the 100 degree closet scares me a bit-- I don't think I'd want to overheat the PC. Put some small vent holes in wherever you're placing the PC.
At a client site, a secretary put her PC in a cubby-hole in her new modular office furniture. She then placed several hell-desk calls in the following week complaining of random lockups. Teching thru it on the phone, we found nothing to attribute the problem to. An on-site visit, though, revealed the cubby-hole w/ the closed cabinet door. When we opened it, superheated air flooded out and, after a couple hours of 'cooling off' and the removal of the cubby-hole door, the PC functioned fine again. She said she shut it because she didn't like the 'grinding sound' it made when she started apps.
Slashdot Mirror
(Parent of parent...)
OSS is the way. Well the proper licensed stuff is anyway.
(Parent...)
Oh, now that was just plain funny. It seems that /. alone posts a "Version 0.2.3.1.2 of GNUXKApp is out today" several times daily. RedHat has, what, at least a new major release every year for the past few years?
(And now me... *smile*)
The point is that no one can stop you from using Free software in whatever manner you want to. If you love Linux 1.2.13 you can keep right on using it for as long as you want. You can contract out for firms to add drivers or fix bugs as much as you want. If you want to pay somebody to backport IPtables or Usermode Linux to 1.2.13, go for it!
Remember USB devices that used WDM drivers that say "Requires Windows 98". There's no reason why WDM drivers couldn't be made to work under Windows 95-- except that Microsoft didn't want to do that work when you could just pay them more and get Windows 98. That's just fine, too! It's their code, so that's their right. Want to add that support yourself? Too bad-- you don't have the code, and it probably violates a license anyway. Think "Group Policies" and Windows NT 4.0 (if you've never noticed, Group Policies are implemented mainly by a tweaked-up USERINIT.EXE), or perhaps FAT32 and Windows 95 OSR2. I'm picking on Microsoft a bit unfairly, 'cuz there are other manufacturers that are more flagrant about it-- but it's their code, so it's their right, and you're stuck "on the treadmill" because you chose to use their software.
Free software isn't anybody's code, though. You can add whatever you want-- or hire somebody else to do it for a fair and equitable rate.
"Upgrades" don't "have to" happen. These "forced upgrade" cycles are a symptom of the idiotic "commerical software industry" believing that they are somehow both manufacturing and service companies-- all at the same time! Use and contribute to Free software, and get yourself off the treadmill if you don't like it.
The unwavering constant in my world of IT consultation work is the assured shittiness of the Customer's line-of-business application. Either it's an off-the-shelf app. that they're pushing beyond its functional limitations, or it's some home-grown bag of dung that coddles their entrenched antique business processes and reeks of inconsistent user interface, poor or completely lacking forethought in design, and lamentable "technologies" (everybody say "toolbars of icons with no tooltips and no menus", "two digit decimal date fields" and "shared file database"). In the end, it really doesn't matter how they've chosen their IT fate, it always ends in everybody bitching about how bad it all is!
The idea of defining requirements and selecting off-the-shelf packages based on those requirements seems to be completely foreign to non-technical users ("I have three (3) kids and a dog-- I think that two-seater little red convertible sportscar will make a good family car!"). Of course, software marketing would have you believe that their products will allow you to travel backwards in time and transmute gold from pocket lint, as long as you keep up on your "maintenance agreement".
On the "internally developed" side, the failings I see almost always involve an inability for a development group to shut the fuck up about their fucking "technology" and learn about the users' BUSINESS REQUIREMENTS! They users aren't going to get any benefit from your buzzword-fortified J2EE-compliant mobile wireless XML fibre-channel attached pneumatic Bluetooth ass-rampager if they never USE the damned thing because it doesn't satisfy any of their business requirements.
We don't have the fucking computers in the business because we just want to have computers! We are here to make fucking money, and the computers are tools to help us.
Doesn't really matter much to me, though... They'll all still need switches, routers, and infrastructure gear, whether they ever get it together or not... *smile*
The "boring router switchy things" pic appears to show two (2) Cisco Catalyst 6513 chassis with dual-redundant supervisor modules. Yeesh... Depending on the options, there's another $200K in gear right there.
How can this company be doing well enough to afford this gear, yet be dumb enough to let their people "case mod" the E15K's?
You're being a little too harsh. I agree that the mentality of "Oh, throw some PHP and MySQL at it" is unrealistic-- even idiotic, at best-- but medical billing is not "rocket science", from a programming standpoint.
(I've worked "integrating" medical, dental, and eyecare billing software into many practices. It's just software. If anything, it's software that actually DOES very little.)
The hardest part I've seen, from the standpoint of watching Medisoft and others modify their software (over and over and over and over) appears to be getting the software to work flexibly enough so that offices "entrenched" in their antique "paper-based processes" are able to function with some degree of efficiency.
Have you ever looked at the drivel that passes for billing software for small offices? All of Medisoft's DOS versions were just flat-file ASCII! Their new "client server" 'doze versions are based on the Extended Systems "Advantage" database engine-- but it's still not that complicated. I haven't seen a package for the dental or optometric field that's much more complicated (albeit most of those apps bring charting into the mix, too).
For a small office, the "technology" isn't the issue. I would agree with your statement that medical billing is a "science in and of itself"-- if only because it's been made that way by idiotic regulation and the insurance industry-- but those are problems that need to be solved by a human operator-- not by a software application. The software part of the job just isn't that big of a deal in a small office. (Hospitals, very large practices-- there's a difference there).
Some of this is probably my angst built-up from actually working with people who administer and operate billing and coding operations. I've met people that were, effectively, trained chimps with a certificate from a community college that could just assure me that I couldn't possibly grasp the science that is medical billing-- as though it's simply beyond the capability of mortal men to understand. It doesn't, however, seem to be beyond the capability of middle-aged fat women in flowery gowns and white shoes.
(It's times like these that a good ol' "salt the earth" Free software project sounds like such fun!)
You cannot boot Windows NT 3.51, NT 4.0, W2K, or XP w/o a video driver. They just won't do it, "out of the box".
The "embedded" versions of Windows NT (NT Embedded 4.0, the W2K SAK, or XP Embedded) contain a "Null VGA" driver that allows them to boot w/o a video card (also contain keyboard and mouse drivers that do the same). Take a look here for a marketroid overview, or here for a FAQ on NT Embedded 4.0 (Search for "Null" if you want to see the "headless operation" part-- I find it particularly interesting that they have an option for a serial console). (These "null" drivers were acquired from a 3rd party by Microsoft. I've looked at the DDK, and actually, it wouldn't be all that hard to implement a "null" driver for those services... maybe I ought to think about that.)
For those who are development minded, it's worth it to take a look around the NT "embedded" pages at Microsoft. I really like the NT kernel (not the bag-of-crap that is Win32, mind you), and I like that Microsoft is letting you strip out the crap "features" and run "bare metal" on the native-mode kernel if you want.
(I've always been convinced there's a Unix inside NT trying to get out... *grin* I'd love to roll an "NT distribution" that stripped out the GUI, and booted the Interix subsystem w/o Win32... Oh, well... I guess I'll just stick to tinkering w/ Linux.)
Even in our own office, I wish we could kill multicasting. We make games here, and in the evening, a lot of groups of guys are playing games that spew enough multicast packets to bring our 100mbit network to its knees. (Yes, we're using switches, not hubs.)
You need switches that support IGMP spoofing, then. If the traffic really is multicast, and not broadcast, IGMP spoofing is what you're looking for.
<KARMA-WHORING>
Take a look here-- it's a pretty decent article on IP multicast (not really technical-- just buzz-laden). Some rudimentary technical information is in here.
</KARMA-WHORING>
(Yes-- it's offtopic, but I don't care.)
This isn't a good idea. I really wish I could get my technicians to stop doing it. There are items there that will never "re-appear" magically. Take a look under HKEY_LOCAL_MACHINE\Enum\SW. Those items won't "re-detect".
You would do better to boot in safe mode, and hand remove everything from the "Device Mangler" except for the software media items ("Microsoft Streaming Clock Proxy", "Microsoft Streaming Service Proxy"), the "IO read data port for ISA Plug and Play enumerator", and "Plug and Play Software Device Enumerator". Leave those alone, and reboot. Let it detect (it may complain that your video adapter is mal-configured-- that's a good sign-- let it detect). It might take a couple boots. Load your INF updates, drivers, etc, and you should be all set.
I've done a reasonable quantity of backup-solution deployments, from the simple "tape drive in a server" to multi-element DLT libraries. I've had customers "invent" a version of this idea on many occasions. Typically, the customer's "invention" takes the form of one of several similar ideas.
What it comes down to, though, is that the idea behind having multiple medias, stored _away_ from the production copy of the data, is a good thing. Until recently, this has only been really convenient with tape media. With the advent of very convenient hot-swappable hard drive carriages and support for hot swapping of hard disk media in nearly every commonly used operating system, I don't see why hard drives could not be used-- but they would need to be treated with a little more physical care than tapes.
The "problem" seems to come when the (typically small-business) customer "invents" this idea, buys one of those cruddy "centronics connector on the back" sub-consumer-grade plastic "drive bays", slaps a hard drive in it, and starts doing backups to one hard drive from another. The cycle is something like: (1) insert 2nd hard drive, (2) wipe 2nd hard drive, (3) copy contents of production hard drive(s) to 2nd hard drive, (4) remove 2nd hard drive. They don't think about what would happen if, say, between steps 2 and 3 the production hard drive(s) failed.
If you're going to use hard disks as "tapes", I don't think there's anything fundamentally wrong-- but buy the same number of hard disks as you'd buy tapes-- and rotate them in the same manner. Treat them as large, mechanical tapes. Keep them away from the production data except when in use.
I remembered this from the dead-tree edition, and luckily it's one of the articles that has full text available online.
Check it out here...
This shoulder-holster looks interesting. The 80's cop-show goodness is almost too much to pass up-- too bad I don't wear a suit jacket at work.
Looks like there are SOME interesting uses for these old Metricom modems in Linux. STRIP looks pretty interesting. Still, I'm seeing articles like this that make me think that the newer, faster modems are going to be useless.
Anybody have any more info relating to using these new GS models in peer-to-peer? I just bought a GS model, and I'm looking to get another. I'm seeing people posting on here that the modems are trash now-- but I'm not so sure that's the case.
Get a motherboard with 4 IDE channels (most "raid versions" have this) and plug 12 drives into the Hotrods, 4 into the motherboard's raid channels, and 2 into the secondary ide channel. The boot hard drive goes on the primary ide channel.
Ouch! Spend a few hundred more, and get an Escalade Storage Switch. They perform very well and aren't wildly expensive (you should be able to have an 8-channel 32-bit, 33Mhz version for under $500.00). You also have your motherboard IDE channels free for things like DVD-ROM drives... heh heh... Lots of DVD-ROM drives... Heh heh...
Ahh, yes-- and there are Linux drivers available for the Escalade controllers. If you're looking for wild amounts of performance, they do have a 66Mhz, 64-bit PCI version available, too. Wowza.
Promise has their SuperTrak controller, which looks very interesting, but based on some messages I saw flying around on the Kernel List, apparently it's not as straightforward as just compiling in I2O support to use it under Linux. Grrr...
Check out this review and this review if you want to see how the Escalade stacked up to other "high end" IDE RAID controllers.
A Cisco 6509 is not a cheap piece of hardware (yours was at least $30K w/ what you've described in it). Especially if you've got redundant supervisor engines, and doubly so if you're doing layer-3 in the box. That having been said, why couldn't you have hired a competent technician to install the box properly when you installed it in the first place, rather than having a half-assed configuration loaded?
I doubt your setup would've been more than three or four hours of configuration, based on what you've described, and you'd have gotten decent documentation out of all of it, if you'd hired a good technician. It's quite obvious that one visit by a technician at $250.00/hr would've more than made up for the cost of the downtime and headaches you incurred as a result of having a poor configuration in the first place.
Putting a man into an amateur space vehicle is, IMHO, probably a bad idea. I think it's highly do-able for amateurs to put vehicles into orbit. I'm a "high power model rocketry" fanatic, and I have a lot of respect and appreciation for what can be accomplished with simple solid-fuel boosters. There are orders of magnitude difference, however, between the complexity of a small solid or liquid fueled booster that can achieve tens of thousands of feet of altitude, and a booster capable of throwing a payload into a stable orbit.
I believe the technology necessary to put men into space is not beyond the grasp of private individuals or small corporations. Look at the advances in technology since the 1960's and the "golden age" of the U.S. space program-- it should be possible to do the same things the Mercury or Gemini programs did for a fraction of the research and development cost.
Having civilians participate in the space program is the best way to build enthusiasm for the space program-- but I don't think it wise for private individuals or corporations to go trying to loft men into space just yet... There has not been enough unmanned testing of vehicles yet to assure that launching these things will be safe for the men onboard, or those of us on the ground!
Maybe your working an 80 hour week becasue you are using the wrong tools or trying to implement the wrong solutions, etc...
I would agree that a measurable number of 80-hour weeks are caused by just what you say. I would also make the statement, though, that in many companies, it is expected that IT employees are expected to do far more work than employees in other capacities. That's not to say that there aren't a significant quantity of do-nothing IT personnel in the world-- but there are definitely a significant quantity of IT personnel that are expected to support (with no regard for the unreasonable time commitments, often) systems that they might not have implemented.
Ultimately, if the 80-hour a week job is a problem, quit. If the company's stupid decision-making process caused the "make work", quit. If being an "all Windows shop" causes the work, quit... or just kill yourself. Pick one.
When I started with this firm, I was a ~45-50 hour a week guy... That grew into the ~70-75 hour a week when I got placed on salary and had my "responsibilities increased commensurate to my salary".
Now that my wife has left me, and now that my responsibilities have been increased further, I work ~85-90 hours a week at this job (generally 6 or 7 days a week), and around ~12 a week at my "for fun" job... and I still have time to sleep, sometimes.
My wife leaving was a big bonus for the company, I guess-- I loathe going home and seeing her things the she has yet to come pick up, so I stay at the office and sleep on a couch in the back some nights.
Yeah-- I'd rather have my wife back than have this job-- but that's not gonna happen. Yeah, I made stupid fucking choices, and now I trapped in a shitty job w/ no future, alone. I know it's my fault, and I'm not really bitching. Hell, I like my ~80 hours in a slow week.
*grin* Sucks to be me, doesn't it... Being 23 is supposed to mean working for a dishonest slavedriver bastard over 80% of your waking hours. Of course, I still have time to read Slashdot.
The key is that the cell is an emergent system.
Moreover, the genome is wholy the product of evolution. We've seen, in a very simplistic way, the types of "optimizations" that evolution creates by using the genetic algorithm and such ilk to "evolve" programs.
It's silly to think that the genome is going to be anything but highly "optimized", incredibly "interdependent", and something that I can only express as "multidimensional". I'd liken reading and understanding the genome to disassembling highly optimized machine language back to assembler source code except I think that it's orders of magnitude more complex than that.
We are aware, in the case of a processor and machine language, exactly how all the interactions and processing occur. Though we have a lot of knowledge about protein folding and the spatial interaction of proteins, I daresay that we're not aware of 100% of the "rules", and will miss nuances in the "code" as a result.
This makes me think about the article posted a few days ago concerning the "evolved" FPGA systems, and the comment that the resulting systems were "strange", and were filled with structures that appear to "do nothing" and were not wired to surrounding parts of the systems. When these "do nothing" portions are removed, though, the system does not function properly-- showing the evolution worked to take advantage of all possible types of "interaction" within the substrate of the FPGA (quantum effects, possibly?).
I've often heard the observation that a large portion of the genome is "garbage" or is not used specifically for coding proteins. I would submit that we cannot dismiss any of the genome as "garbage", given that we aren't 100% sure of all the "rules" of the genetic system. Not at all unlike trying to understand the strange systems evolved on FPGA's.
When I wired my house last year, I decided to go ahead terminate every wallplate with dual RJ-45 jacks. All the cable is Cat-5, and terminates into a small patch panel in a closet. I brough my telephone lines into that closet, and bought a couple "1 to 5 way" phone splitters at a local discount store. When I need to route phone lines somewhere in the house, I patch from the phone splitter to the patch panel w/ RJ-11 patch cables. RJ-11's fit into RJ-45's fine-- albeit a little sloppily-- but they work.
I went the extra step of using gang outlets that support CATV terminations, as well. I brought my CATV feeds into the wiring closet, since I needed it there for the cable modem and firewall anyway-- so I have an amplifier there and panel to distribute CATV to the rest of the house. Someday I'll add an RF modulator and patch an external camera in on some channel... *grin*
I would agree w/ all the posters that say that cheap metal shelving makes for a nice rack. Paint it black and fill it with computer gear-- it looks pretty boss and doesn't need to be bolted to the floor to be stable!
If you're shopping around for a house, look for phone jacks in areas where you might want data jacks. Phone wire makes GREAT pull-wire for your new cables.
I HIGHLY recommend a patch panel. I've seen a lot of small offices and a couple homes that were just "crimped on RJ-45 ends" at the "wiring closet" end. This limits your flexiblity in regards to patching phone across the wires, and will make you very unhappy when you jerk on a wire, pull off a termination, and have to reterminate the wire. Patch panels are comparitively cheap, given that you want this wiring in place and functional for a long time.
Sounds like you want to build a mini "head end" for a CATV system and stick some PC's decoding MPEG streams on the end of it. You'll send commands to the video servers via Ethernet or some-such, and send the streams into RF modulators and onto the CATV lines.
I'm fuzzy on the RF modulator part (I've worked with using ONE RF modulator on a line-- but I know that there exist solutions that will allow you to pile multiple signals from several RF modulators onto the same wire)-- but that would really be the hardest part. We did a "closed circut" system in a school where I interned a few years ago-- but that was done over CAT 5 w/ some type of balun on each end and only one RF modulator in the mix.
I'd recommend foregoing hardware MPEG decoders and get some el-cheapo PC's that have enough horsepower to decode the streams you're looking to decode. Boot 'em diskless w/ a server hosting all the encoded files. Stick a video card w/ reasonable quality TV out on each PC, put 'em on an Ethernet switch w/ an uplink to your building LAN, plug 'em into the RF modulators, and you should be set to go.
I don't think you're going to find a cheap way to do this, unless you can come up w/ some cast-off RF modulators. You're talking about hardware that doesn't have a tremendous non-professional market (not many people run their own CATV system) so it's not cheap. Sounds like a fun project, though.
Have you used a sniffer to see what this "broadcast storm" traffic is? That would do for a good start. A description of the issue other than "kill the network" would do more good as well. What kind of byte/sec and packet/sec counts are we seeing on the media?
Assuming it is broadcast related: Your 'doze boxes need to be using "H-Node" name resolution for their cruddy NetBIOS name resolution. You need WINS servers. You need to disuse protocols that are broadcast intensive.
To the guys that are saying "install a switch"-- apparently there's an understanding issue w/ regard to what a switch does. A layer-2 switch won't help a bit in this case-- just like the poster said (referring to bridging routers). An analysis of what the traffic on the wire is would be a great first step. Then, intelligent decisions can be made to address the problem. Layer-3 switching might be a potential solution, depending on what the traffic is.
Unless you're using layer-3 entities inside of switches, your router based solution is going to do more than mess up NetBIOS name service-- it'll slow everything down to a crawl. Most low-end routers don't even come close to wire-speed.
I had a customer swearing to me that they were having "broadcast storms" because they were getting massive numbers of collisions on a shared-media LAN. We took a look at it w/ a sniffer and discovered that broadcasts played no part in it. They were doing large file transfers to a machine that was dual-homed on the same physical NIC, and the machine was thrashing packets on and off the wire, "routing" the packets to two hosts that were in different subnets, but on the same media. Duh.
Don't assume you know what your problem is unless you know what your problem is...
The previous run of this story is ref'd, but the discussion did center around use of codes on insecure channels during WWII. A good book, for the interested party, is Between Silk and Cyanide, by Leo Marks (ISBN 0684864223).
I'd have to echo the pessimistic sentiments of others in this discussion, though, and state that there's probably not a lot of hope in 'cracking' these transmissions, given that we have no knowledge of their origin or purpose.
This works well. We rolled out a butttload of Linux and Win95/98 machines like this. I've done 'recovery CD's' for Win95/98 w/ this method, too. (Gee-- isn't that misuse of a GPL'd operating system and tools... *grin*) If you can swing it, a bootable CD makes it even handier-- albeit if you have to get into the BIOS to make the machine boot off of CD, the time-savings is negligable over just booting off floppy.
As a more ambitious project, we made a little bootdisk w/ dhcpcd and support for the NIC's in the machines we were using, and used a little program called 'netpipe' to broadcast out .tar's to PC's hard disks over the wire. It _rocked_ to see 150+ machines pulling an image simultaneously at 800KB+ a second on garden variety 10Base-T.
First Neuromancer, then Snow Crash! NO!!! Books that I love, being put on the chopping block like this. Snow Crash has been an influence on my lifestyle. I can't even count the times I've stomped my brakes and cursed at 'fucking Bimbo boxes' on the highway...
Snow Crash (and Neuromancer, for that matter) was one of those books that FELT like a movie while I read it. Reflecting on it now, I feel like I've seen the perfect movie version already, because Stephenson's characters were so vivid. Anything they do with this story will disappoint me, and many others, I'm sure.
The one genre that I've thought would be able to handle films like Snow Crash (or Neuromancer) would be anime. Busting Snow Crash into 3 2-hour animated episodes would be great. I'd prefer it was voiced in English, but it'd look awesome as anime.
Oh, well... Who'd listen to me. I'd write a screenplay for an animated Show Crash, but having no artistic skills, I couldn't storyboard it. W/o storyboards, I don't think it's the kind of idea that would sell. Besides, the novel really doesn't need a lot of adaptation. (I can think of a few things that might need to be tamed a little for U.S. audiences... Y.T. and Raven's sex scene, potentially some violence.)
Picture it in your mind-- the stereotypical quickly-scrolling-background-art of many an anime production, w/ the magna-poon whipping thru the foreground and sticking on the back of a minivan... Picture it. Ahhh...
Maybe I'm just stuffy, but I'm concerned that the GNOME community is missing the boat on features that could legitimately make the dream of 'desktop domination' for GNU/Linux a reality. I'd certainly agree that user interface friendliness and a full-featured array of applications are the major factors in getting GNU/Linux onto corporate desktops, and I think competing with Microsoft OS's and Win32 apps on these points isn't the place when GNU/Linux can show the biggest win.
The 'centralized management' aspects of Microsoft operating systems and most win32 apps are horrendous. Expensive software 'hacks' like Tivoli, LANDesk, Novell ZenWorks, and Microsoft SMS are all ugly kludges to attempt to provide management features for large fleets of PC's. (I just can't say how much these hacks SUCK enough). The fact is, the entire 'single-computer, single-user' metaphor that win32 OS's and apps are based on is completely wrong for centralized management of fleets of PC's. Let's face it-- shit with icing on top is still shit.
As I look at the state of GNOME now, I'm not seeing the move toward management features that would bring serious money-saving features to the desktops of the corporate world. GNU/Linux being free is not enough for most IS managers to give up their MS-based ways-- the real investment is in management labor (all that 'Total Cost of Ownership' crap). Sure-- today a competent hacker could use scripts and some slight-of-hand to provide many remote management features, but what I'm talking about is providing these features to corporate IS managers in an easy-to-use system. Features like:
The only way I see to provide this correctly is to design it in from the start. Look at the MESS that is win32, and look at the leaps and bounds of hacks SMS, ZenWorks, and all the other 'management' frameworks go through. If we don't get to designing in these types of technologies NOW, we're going to end up with a hodge-podge of configuration repositories, labor-intensive application rollouts, and TCO similar to win32.
For GNU/Linux to offer serious labor savings to corporate IS managers, I'm talking about things like:
Maybe we need a Linux Management Initiative or something of that ilk. I'd love to do it, if I was a competent enough coder and had enough time. Instead, though, I'm just a GNU/Linux user stuck working in a win32 based world and dying to get out of it. This stuff isn't glamorous, but I think it's a place that GNU/Linux can kick some serious Microsoft ass.
A coworker purchase one of the PC Power and Cooling "ultraquiet" PS's for his PC, so I've gotten to play w/ one a bit. Side-by-side a Sparkle Power ATX supply, it was definitely quieter-- but I wasn't terribly moved by the difference. Overall, I didn't think the extra cost was justified by the result.
The idea of moving the PC itself away and using extension cables is probably the best idea. The previous comment w/ the 100 degree closet scares me a bit-- I don't think I'd want to overheat the PC. Put some small vent holes in wherever you're placing the PC.
At a client site, a secretary put her PC in a cubby-hole in her new modular office furniture. She then placed several hell-desk calls in the following week complaining of random lockups. Teching thru it on the phone, we found nothing to attribute the problem to. An on-site visit, though, revealed the cubby-hole w/ the closed cabinet door. When we opened it, superheated air flooded out and, after a couple hours of 'cooling off' and the removal of the cubby-hole door, the PC functioned fine again. She said she shut it because she didn't like the 'grinding sound' it made when she started apps.