Announcing: Slashdot Deals - Explore geek apps, games, gadgets and more. (what is this?)

Thank you!

We are sorry to see you leave - Beta is different and we value the time you took to try it out. Before you decide to go, please take a look at some value-adds for Beta and learn more about it. Thank you for reading Slashdot, and for making the site better!



Cosmologists Show Negative Mass Could Exist In Our Universe

bittmann Why do we keep trying, then? (214 comments)

Having built a number of gravitational wave observatories that have to see a single gravitational wave...

If they must see that same single gravitational wave over and over again, why do we need to keep building more of them? Why don't we build some to see OTHER gravitational waves?


about 6 months ago

Apple Has a Lot In Common With The Rolling Stones (Video)

bittmann Re:"People" vs "pundits" (147 comments)

the single thing a new iPhone will do that an Android won't is the fingerprint thing

My 2011-era Motorola Atrix has a fingerprint reader...

about a year ago

Germany Produces Record-Breaking 5.1 Terawatt Hours of Solar Energy In One Month

bittmann Re:Why migrate to Turkey? (687 comments)

In the "strange but true" category, you can actually convert Turkey into light petroleum. A few years back, a Thermal Depolymerization plant was built next to a ConAgra Butterball Turkey processing plant, intended to convert feathers and other waste into oil.

A Thermal Depolymerization demonstration plant was completed in 1999 in Philadelphia by Thermal Depolymerization, LLC, and the first full-scale commercial plant was constructed in Carthage, Missouri, about 100 yards (91 m) from ConAgra Foods' massive Butterball turkey plant, where it is expected to process about 200 tons of turkey waste into 500 barrels (79 m3) of oil per day.

So, while it may not be "cheaper" to burn Turkey than coal or natural gas, it is arguably "cleaner", at least from a net-CO2 perspective.

about a year and a half ago

Sprint May Have Unlimited Data Plans, But Not Unlimited Customers

bittmann Re:t-mobile is the best low cost carrier (207 comments)

StraightTalk is almost never ATT service anymore - you can't even order a StraightTalk AT&T SIM kit nowadays (unless you pay a premium on eBay, I suppose). Interestingly enough: You can get a couple of different SmartTalk smartphones that are activated on the Verizon network.

about a year and a half ago

FAA Wants All Aircraft Flying On Unleaded Fuel By 2018

bittmann Re:Thanks Slashdot. (366 comments)

The higher octane is required in higher altitudes. 87 at sea level will give more power than 91 in Denver (caveats apply)

The main caveat being: That first statement is completely incorrect. As altitude increases, ambient pressure decreases. As ambient pressure goes down, max pressure developed in the cylinder decreases as well. As max pressure decreases, the tenancy for pre-ignition (knocking) decreases. As the tenancy for pre-ignition decreases, octane requirements are lessened.

In other words, all things being equal, higher octane is required in lower altitudes.

(But you are right in your second assertion - assuming your engine will run on 87 octane at sea level, it will indeed make more power at sea level than it will in Denver - mainly because of the increased air density at sea level.)

about a year and a half ago

GOP Blocks Senate Debate On Dem Student Loan Bill

bittmann Re:Obama knows how to play politics if anything. (834 comments)

And thirdly, the Democrats have already argued for a 33% reduction in that very fund themselves:

A tentative deal struck late Tuesday between House Ways and Means Committee Chairman Dave Camp (R-Mich.) and Senate Finance Committee Chairman Max Baucus (D-Mont.) would cut federal healthcare spending by $21.1 billion.

The savings would be used to pay for a "doc fix" that would eliminate a scheduled 27.4 percent reduction in Medicare physician payment rates for 10 months.

Savings include:

- A $5 billion cut to the health law's $15 billion prevention trust fund;

- The elimination of $2.5 billion in enhanced Medicaid payments to Louisiana in the wake of Hurricane Katrina;

- A $4 billion reduction in so-called Medicaid Disproportionate Share Hospital (DSH) payments to hospitals that take care of people without insurance;

- A $6.8 billion reduction in federal payments to hospitals that collect "bad debt" from insolvent patients, down to 60 percent; and

- Cuts to how much Medicare pays for clinical laboratory tests.

How is it that reducing the funding by 33% *isn't* an attack on women if reducing the fund by a lesser amount *is*?

more than 2 years ago

Man-In-the-Middle Remote Attack On Diebold Voting Machines

bittmann Re:Not very correct (251 comments)

Most "void if broken" seals can be easily replicated. It's just a matter of getting a replacement seal in time. For the most part, people are dumb. If you do a good job of cleaning off the seal, they'd never notice it is missing.

I'll go you one further--I seriously doubt that "void if broken" seals would even be honored! If they were, any griefer with an axe to grind could quietly slice a "void if broken" seal and arguably void (nullify) any votes cast on that box up until the point that broken seal is noticed -- possibly all day. Unless (of course) the seals are visually checked in between each voter, right? So next time you go to the polls, watch how the lines move, and see if you think everything is visually inspected and verified between each voter.

And if this sort of vandalism did happen, what would you bet that the votes up till then wouldn't be nullified regardless of the state of the tamper seals? What makes you think that this sort of thing hasn't already happened? In past election, seals have been found missing/cut on machines, it's been reported, and it's been ignored and the votes counted regardless, e.g. as reported here. Nice.

more than 3 years ago

AT&T's Metered Billing Off By Up To 4,700%

bittmann Re:My router's traffic shows 10-15% lower than AT& (250 comments)

Agree about the lag. The current reporting period is (presumably) 3/1-3/29, but the only way I can get the reported traffic to add up is to include traffic back to 2/27 which (last I checked) came before 3/1. Plus, as of 3/29, AT&T has no detail from 3/25 onward. If that holds true at the end of the month, the reported usage won't help a whole lot.

Pretty tough to manage to "not exceed" 150GB of traffic when (on a 6 megabit connection) you could in theory consume 50GB of traffic in just those 4 missing days. (55+ GB, counting ATM and PPPoE overhead....)

Of course, AT&T doesn't *want* you to be able to manage your own Internet usage, what they want to do is to either scare you into non-utilization, or to charge you extra (note that overages are billed in chunks, as in $10 for each additional .0001 to 50 GB, instead of $1 for each additional .0001 to 5 GB). If you're a "heavy user", best to let you go over by a half-gig and ding you for another $10, eh?

One thing this whole experience has taught me -- no reason to pay for faster Internet access if all that will happen is that you'll hit your caps faster. You folks on 6mbit AT&T DSL -- why not use this change in contract to fall back to 3mbit and save quite a bit each month?

One thing this has done is to convince me to look at Cox again. I left Cox for AT&T when Cox pricing and plans went unrealistic, but now they are starting to offer tiers that compete with AT&T on price while still (at least on paper) beating them on performance. Just need to see what traffic caps are in place...

more than 3 years ago

AT&T's Metered Billing Off By Up To 4,700%

bittmann My router's traffic shows 10-15% lower than AT& (250 comments)

If I'm paying for PPPoE and ATM overhead, I'm gonna be pissed.

AT&T must be measuring bits at the DSLAM, if what they're reporting is anywhere close to being accurate. If a 150GB "cap" includes the approx. .5% PPPoE and 10+% ATM overhead, what I'm seeing means that my 150GB cap is in reality closer to 135GB.


more than 3 years ago

First Ever HIPAA Fine Is $4.3M

bittmann Don't diss Big Brother (197 comments)

Seriously -- is this fine about HIPAA, or is it about failing to snap to attention when the Big Government Agency came calling?

Also seriously: One of the HIPAA loopholes that patients aren't always told about is that HIPAA privacy rules don't necessarily apply when the government gets involved. One could easily argue that Cignet shouldn't have released those 4,500 unneeded records, you bet...but one could also argue that the release of those records didn't automatically trigger a HIPAA violation, as they were released in response to an oversight request, e.g. "Covered entities may usually disclose PHI to a health oversight agency for oversight activities authorized by law." (source: CDC.gov). If HITECH changed that, it'd be news to almost everyone -- when is the last time that the government willingly adopted rules restricting their own capabilities?

Regardless, IMO if they would've done exactly the same release of information BUT responded in a timely fashion to the Government's demands, there wouldn't have even been a $43 fine. Because that's the way that the Government seems to work.

more than 3 years ago

Smithsonian Celebrates 50 Years of COBOL

bittmann Re:I thought COBOL basically died after Y2K. (178 comments)

How many closet dinosaur-language slashdotters are there?

Here's another, although my COBOL programs run on an AS/400 (a.k.a iSeries, system I, whatever the heck IBM has decided to call it today) instead of Really Big Iron. Some of the code that I wrote on a CISC-based System 38 over 2 decades ago is still running on a Power6 RISC-based system today, with nary a recompile involved.

It's hard to argue with "works".

more than 4 years ago

Free Clock Democratizes Atomic Accuracy

bittmann Segal's Law (178 comments)

"A man with a watch knows what time it is. A man with two watches is never sure." -- Segal's Law.

more than 4 years ago

How To Build Roads To Control How Fast You Drive

bittmann Re:Other strategies... (801 comments)

The Portugese system goes further and makes other drivers angry with you for speeding.

I think the Portuguese system is the future. Note that it shames you in front of other drivers, but that it also slows you as a penalty for speeding. People will naturally adopt the behavior that gets them where they are going fastest. If you make 'speeding' the slower option, people will just naturally drive safer.

So what's the likelihood of someone who has no problem breaking one law (speeding) breaking yet another law (running a red light)?

I suppose the next response would have to be something like "red light cameras".

more than 4 years ago

Federal Deadline Hobbling eHealth IT Rollout

bittmann Re:The Flip Side (99 comments)

I'm not big on government interference with many parts of our lives, but they are addressing a very real problem and they're doing it with kid gloves. They did not pass regulations requiring hospitals to comply, they just tied federal funding to that compliance and gave the hospitals many years in which to get their shit together. If medical providers have not done so and are rushing about now, that is absolutely not the fault of the feds.

Actually...one of the dirty little secrets here is that the final rule for meeting "meaningful use" still isn't actually final. The "interim final rule" wasn't even available to view until Jan, 2010 (link), comments are accepted through March 15th, and we should have a final rule that we can (hopefully) comply with by the end of this month.

And: We don't have "many years" to do the install. We have a few years...very few, if we want to actually participate in the government incentives. Have to be installed and in production by late 2011 to qualify for the full incentive. Any delay, and the incentives go down drastically.

In our case, this whole thing really bites. We have an EMR, fully deployed, and we haven't maintained a paper chart in years. But, because of the definition of "Certified EMR" (which at this point basically means "Must be certified by CCHIT"), we can't qualify for "Meaningful Use" under these proposed rules. So, we have an EMR, we produce escripts, we do online order entry, we can even exchange imaging information (something that this round of certification doesn't require), but because we can't fill in all of the check-boxes in a CCHIT audit, we have to scrap our homegrown EMR and pay millions to replace it with a "certified" alternative. And the government will give us some of that money back if we cram it in fast enough *and* if we are able to show that we meet whatever standards the final rule eventually mandates...all within the next 18-30 months.


It may not be the fault of the Feds that some providers haven't transitioned to digital records, but the Feds certainly aren't making things very easy, either.

more than 3 years ago

Federal Deadline Hobbling eHealth IT Rollout

bittmann Re:Fast, Good, Cheap, pick 2... (99 comments)

We're talking about the US Federal Government here. In particular, the CMMS (Center for Medicare and Medicaid Security). You get all three.

"Ggovernment is bad" sock puppet, we're talking about private-sector insurance here.

"A federal deadline that begins next year and requires hospitals to prove they're meaningfully using electronic health records will lead to technical problems and data errors affecting patient care, say politicians and top IT professionals responsible for the deployments. Physicians and hospitals have until the end of 2011 to receive the maximum federal incentive monies to deploy the technology. If not deployed by 2015, they face penalties through cuts in Medicare reimbursements. 'I think we have nontechnology people making decisions about technology,' said Gregg Veltri, CIO at Denver Health. 'I wonder if anybody understands the reality of IT systems and how complex they are, especially when they're integrated together. You're going to sacrifice quality if you increase the speed [of the rollout].'"

"Private-sector" applies to this discussion ... exactly how?

more than 3 years ago

"Logan's Run" Syndrome In Programming

bittmann Re:Yes and No (599 comments)

I maintain C code written by a COBOL programmer. You can tell.

The code is written in a verbose, heavily-commented, yet easy-to-read style, and actually does what it appears that it should?

more than 4 years ago

How Do You Accurately Estimate Programming Time?

bittmann Use the "double-up" method: (483 comments)

Do a best-case estimate of the time required, assuming everything goes perfectly and that there are no surprises. Next, double your estimate. And finally, switch to the next highest unit of time measure.

A quarter-hour change will take half of a day. 2 hours becomes 4 days. 3 days becomes 6 weeks. A 6-month project will take 12 quarters, or 4 years.

It's eerie how often that rule of thumb seems to accurately depict the actual calendar time required -- eerily enough that when a so-called "realistic" estimate DOESN'T approach this metric, I find it's usually worth a second look.

Thankfully, at least at my place of work, this rule of thumb seems to break down once the unit of measure hits a year...

more than 4 years ago


bittmann hasn't submitted any stories.



How to add new events to new keys in RedHat 9.0

bittmann bittmann writes  |  more than 11 years ago

In my last entry (July 8), I noted that I needed to document how I added keybindings to the volume keys in RedHat 9.0.

First of all, it isn't as easy as it might be. There is documentation in RH9 under Preferences/Keyboard Shortcuts that indicates that there is an "Add a Custom Binding" function in the Gnome Keyboard Shortcuts tool. No, there isn't. So...it's harder than editing an "add" screen. But not TOO hard...

First of all, the keybindings must be put in place to map a keyboard scancode to an event. I added the following to /etc/X11/Xmodmap:

keycode 115 = F30
! Windows key
keycode 174 = F31
! Vol down
keycode 176 = F32
! Vol up
keycode 160 = F33
! Mute
keycode 137 = F34
! eject (busted!)
keycode 117 = F35
! Menu key

Note that the eject key doesn't seem to work well for me...need to get back to that.

So, now, when you press the Windows key, Xwindows will receive an F30.

Now, the obscure (but easy) part:

Go into System Tools/More System Tools/Configuration Editor (GConf)
Drill down to apps/metacity.

Step 1: Add a custom command to an unused entry.

Under keybinding_commands, I edited the following:

command_2 became aumix -v-5
command_3 became aumix -v+5
command_4 became aumix -v0
command_5 became gnome-terminal

I left out eject after I couldn't get the key to work. Will try again "later".

Step 2: Bind the new commands to the new keycodes

Under global_keybindings, I edited the following:

run_command_2 is set to F31
run_command_3 is set to F32
run_command_4 is set to F33
run_command_5 is set to F35

(I left the "windows" key unmapped...I keep bumping it! But I mapped gnome-terminal to the menu selector key (above F9), so pressing that key gives me a new terminal. Slick.)

Restart X, and now the volume keys work.

I have been cautioned that function keys above F35 may not work. I haven't tried it to see. YMMV.


Linux on a Dell Latitude D800

bittmann bittmann writes  |  more than 11 years ago

Going to list some of the triumphs and pitfalls of Linux on a Dell D800 "Centrino" system here. If it goes far enough, I'll update my own homepage with a Howto:...but want to get this down while I'm thinking of it.

Using RedHat 9.0.

RH9 installed "OK" out of the box. Needs a couple of things to get it going, and a little *more* work to get it actually usable.

1) Download Broadcom 5700 driver from Broadcom's site. Compile and get module ready. Set up /etc/modules to load driver when ethernet comes up. You can either use a (supported) USB or PCMCIA NIC, or do the floppy shuffle (it's small).

2) (optional) If you have access to high-speed Internet, go ahead and run up2date. Take the trouble to get the new RedHat kernel (it does IDE chipset DMA, and will make the rest of this less painful).

3) (optional) Get NVIDIA drivers from NVIDIA's site. Or, use the open-source 2d drivers...that'll work, too.

If you only get 1600x1200, you may need to add a modeline to XF86config to allow 1920x1200.

Section "Monitor"
                Identifier "Monitor0"
                VendorName "Monitor Vendor"
                ModelName "Dell 1920x1200 Laptop Display Panel"
                HorizSync 28-110
                VertRefresh 43-90
                Modeline "1920x1200" 162 1920 1984 2176 2480 1200 1201 1204 1250 +hsync +vsync
                Option "dpms"

The box should now be usable. There's no power management, however (and you'll have to "push the button" to turn off the computer at shutdown).

What I did to make it "better":

Downloaded modem drivers (they build and connect... but I haven't tried them after upgrading the kernel. To-Do).

Downloaded 2.4.22-pre2 kernel, which has ACPI working. Quick and dirty: use one of the Configs in one of the RedHat sources directories (I used the "686" config from 2.4.20-18.9). Enable ACPI (I did it modularized for better control). Turn off APM. Turn off whatever else you don't want (ATM? whatever). Build. Set up modules to load in rc.local (or wherever).

Now, I have ACPI working!

Note: processor.o won't set anything but the first 2 clocking states. Haven't had time to look at it yet. Downloaded developer binary from Intel...it works for all 7 processor states. Now I can underclock to the lowest setting, and still play a video at full speed without buzzing the fan.

Note: The linux AGPGART has trouble with the lid switch...if the linux AGPGART is loaded, the lid switch will hang the machine. For now, don't load it (it won't load unless you set the "try unsupported" flag), and it'll work fine. Or never touch the lid switch :).

Downloaded acpid, and have "borrowed" a few scripts for suspend, power off, etc.

Note: Suspend works, but mouse is hung upon return. Can Ctrl_Alt_F1, the Ctrl_Alt_F7 to go back, and it'll work again.

Have added keymappings for volume up, down, mute. Note: Need to tell how under RH9...it isn't obvious (at least to me). Also added a command to the "menu selector" button (above F9) to pop up a terminal. Slick...

More to come...


Kansas joke

bittmann bittmann writes  |  more than 11 years ago

Had a "week from hell", as my real-life workplace implemented a new Digital Transcription system. Many problems/long hours/Microsoft SQL server issues (ugh!)...but anyway, was finally able to do the standard post-install "thing" and get out for a vendor-paid-for bash at the Scotch and Sirloin (a local establishment of some repute...good food, and a waitstaff containing serious eye-candy). One of the vendor's techs, who had *quite* a bit to drink, motioned me over...

"Q: What does a Kansas divorce and a Kansas tornado have in common?"

"A: Either way, someone's losing a trailer."


Annoyances - RedHat 8.0 Dial-up networking

bittmann bittmann writes  |  more than 11 years ago

Note To Those Who Build System Configuration Scripts: Make Sure What You're Doing Carries Through To The Rest Of The System!

Yes, most of this is my fault for not completely auditing the system that I sent away with my parents. However, I do think that there is something in here that says *something* about the lack of coordination within the RedHat 8.0 install.

Bear with me, here...

Just set up RH8 for my parents. They dug it. Looked good to them, gave them a good, integrated "feel", and did away with most of their complaints about arcane commands/editing config files, etc.

They took it home and set up dial-up-networking themselves (the wizard actually runs well enough that they had the modem configured and DUN up and running on the first try!) However, my folks are hooked on the "Modem Lights" applet that shipped with RH7.x. No demand dial for them! No problem...they added the applet to the panel. Except that, in its default state, it doesn't work.

A phonecall and a couple of "ssh" sessions later, and here's what I found:

1) Modem Lights is (by default) pointing at "ifup" and "ifdown" as the commands to do the dirty deeds--commands which aren't in a base user's path. So for root this ran just fine, but the basic user couldn't use it. Click on Modem Lights, get no error message, no response, no nothing. No problem...I saw that one fairly quickly, and told them to change the path to the commands. It would now dial...

2) But now, it won't disconnect. Grrr...it's pointing at the wrong lockfile, so it doesn't know it's connected, so it doesn't try to disconnect (it wants to connect again!). The Internet Cofiguration Wizard sets up the device to generate a lockfile using the device name (/var/lock/LCK..ttyS0 or some such). Modem Lights by default was looking for /var/lock/LCK..modem). What gets to me is that the hardware detection seemingly created the correct link in the device directory (/dev/modem > /dev/ttyS0). Why the networking setup didn't use /dev/modem (or even /dev/modemn for that matter), along with the appropriate lockfile, I can't fathom.

OK, so we got THAT figured out and changed. But wait, it gets better! Being a parent ("Dad, why did you do that?" "I don't know. I just did."), the next thing that happened was that *somehow* the applet was removed from the panel. So it was put it back--And lo and behold! All of the changes that he made were reset to the default values, losing all of the work we just went through. Now, this may have taken place because he hadn't exited/saved, but then again, I didn't want to risk him losing this data later, so...

No problem, we'll just symlink ifup/down from /usr/local, and change the Networking setup to use /dev/modem. But the next time the networking wizard was run...grrr. Back to the tty device.

So: in a nutshell, here's what I'd like to see fixed:

1) RedHat's configuration of the basic tools/configurators/applets/etc. that they deliver in the base desktop install are not necessarily configured in a compatible fashion. This is frustrating to the adept, and confusing/maddening to the newbie, which doesn't help for user acceptance at all. Audit your system installs, and make certain that all tools/applets/etc. can agree on the basics. If something doesn't, then change it (to use a config file/command-line arguments/whatever) and submit the source back to the author. After all, that's what your value-add is supposed to be: Making things "just work". Otherwise, you're no better than (insert distrubution name here).

2) Some of the apps that RedHat ships don't give the user ANY indication of an error. Again...fix the problem! If all they did was give the user some easy/consistent way to view stdout/stderr, at least he'd get some feedback about what was happening. It's made even more tough to troubleshoot when GDM runs...there isn't even a pseudoterminal that's easy/quick to get to. And no, this didn't appear in the System Log, either. As far as my parents were concerned, they were flying blind.

So: Here's what I suggest.

1) If you're going to support abstract devices, make the configurators abstract the devices in a consistent fashion. Add a number if you have more than one of a given device (/dev/modem and /dev/modem1 instead of /dev/ttyS0 and /dev/ttyS1 or whatever).

2) Make all applications that use these devices default to the lowest-common-denominator abstract device (/dev/mouse or /dev/mice, /dev/modem, /dev/cdrom, etc.). Worst-case scenario: Tell the user that they can add a number if they need to differentiate between devices. But in most cases, the device should select "appropriately".

3) And for pity's sake: Test the system install as a user, not just as root. How RedHat managed to ship something that a user would run (Modem Lights) with a command path that is completely non-functional is beyond me. Imagine how much flack M$ would get if THEY pulled this. Hey--maybe their wizards make things run at the lowest common denominator...but at least they kinda-sorta run, or at least pop up an undecipherable error message or blue screen...they don't just sit there and do NOTHING!

OK. I feel better now. Got *that* off my chest!

Slashdot Login

Need an Account?

Forgot your password?