I figure it took him about 10 years to make his way to the tails of the bell curve of all the incandescents in his house. Besides, if he thinks 15 W compact fluorescent is enough for a living room, 25 W for a kitchen, that dude is pretty parsimonious with light. Fluorescents are a big saving over incandescents, but the 4:1 watt ratio is pushing it a bit -- you are going to be a bit dim after the substitution.
Long time ago, I saw our Cocker Spaniel with a mouse in its mouth, and yes, I let it eat it. I understand that wolves feed off mice when they can't get anything else, so I cannot see the harm to a dog unless the mouse had got into some Warfarin or something.
On the other hand, I don't understand the "how" of this dog ever catching a mouse -- this guy would pursue any prey barking and with floppy ears flapping: he was not big in the stealth department. Cats, however, are real killing machines. They are quiet and the "pounce" on their prey.
The stock Classic Star Trek plot involving an intelligent machine is when Kirk asks it to divide by zero, calculate pi to all decimal places, or determine the meaning of life, at which point smoke come out and the machine melts into a pile of slag.
When MS were the good guys
on
Corel Goes Private
·
· Score: 2, Insightful
There was a time when MS were considered "good guys." Yeah, yeah, if you read "Accidental Empires" you learn Bill was always Bill from the beginning and was one step ahead in terms of sharp business practices than anyone else. But MS started out as the outsider/upstart.
In the 1980's you had VAXen, some ran VMS, some ran UNIX, and then you had workstations, SUN, Apollo, SGI (largely UNIX although Apollo was some kind of its own thing): expensive hardware, vendor lock in, only thin source-code compatibility between generations (SUN going from Motorola to Intel (briefly) to SPARC), expensive operations (your grant needed to hire or share one or more "bearded programmers" to operate them and write custom software for them).
Then you had PC's and DOS -- the rootstock of Windows. PC's were like PDP-11's and DOS was like RT-11: low level, cheap, lab computers that an engineer or scientist could hack up their own software to control an experiment. PC's were looked down upon by the CS department-academic computing department establishment as "not real computers", and engineer and scientist types looked at academic computing department types as arrogant so-and-so's who wanted to siphon money out of one's grant. PC's were the second wave of what the Apple-II started -- computers for the masses and liberation from the computer elites.
With PC's you rode the wave -- cheap clone computers, cycles of binary-compatible Intel processors, every imaginable type of hardware (A/D and D/A to control those lab experiments), shrink-wrap software, third-party development tools (MS compilers always blew chunks: Borland's Pascal and later C and C++ was in another league, and you had language choices while UNIX seemed to offer any language you wanted provided it was C or later C++).
And then came Windows. You could ignore Windows 3 and 3.1, but by Windows 95 and NT 4, it was pretty much rammed down your throat. But around that time, the choices were pretty much switching over to the Mac, and the closed-hardware Mac lacked all of the hardware add-ons from Data Translation, and expensive workstations were getting replaced with cheap PC's, so programming for X/Motif wasn't even on the table). If you could live with the performance limitations of Visual Basic, you had an easy way to develop apps; if you wanted performance you had to wade through volumes of the Windows API, but having done that there was kind of a reluctance to learn that much over again to do GUI's on other systems.
So this Linux thing comes along and on '86 hardware as well. And then Bill is Bill: what he had been doing to clone computer makers all along he starts to do to the "little guy." Ghosting over a copy of Windows to a clone you put together from parts for your lab was no big deal until all this fanfare of a crackdown in the form of activation. The University buys a mass MS site license so we all switch from Word Perfect to Word, and this summer the University drops the site license because MS tightens the screws on the terms, and we worry that the BSA is going to bust down the door to our lab.
I have been looking over my shoulder, and I have looked at Java, and I have looked at Linux/Gnome/KDE. Our grant funding is way pared back in these tough times, and we are down to 20 percent share of one UNIX programmer who could care less about setting up a Linux box for the lab, and our engineer guy who is a digital-logic design genius has been thrashing around for over a year trying to get a Linux box up and running but doesn't have the time to swap network cards and fool around with getting drivers.
And just as Java was supposed to make big inroads on the desktop by now, the same was true for Linux. The Engineering library had put in a raft of Linux boxes, but I guess none of the engineering students knew what to do with them that they have all been reverted to Windows.
My engineer logic-design pal and I have talked about how a Linux box would be a good as a Windows box if to the user the thing were just a GUI
In the Soviet accident, they were cutting corners because it was the Cold War, they were testing a new type of missile with storable liquid fuel (the storable fuels, nitric acid and hydrazine or their variants, are very toxic, corrosive, and burn on contact -- that is why Korolev tried to stick with LOX and kerosene, but not very practical for a military missile, and the U.S. switched over to all solid fuel). It was a military sense of urgency to keep their test on schedule. They lost their head military guy (Nedelin) who was holding the stick urging them to hurry up.
That is what I always thought that as you say, that when you fueled the rocket you took shelter in bunkers, pillboxes, and concrete block houses. Were they cutting corners? Was it a solid rocket that went off? Solids are of course always "fueled."
You don't need to flash a passport or visa stamp to go between European countries, but if you take a car from Germany into Austria, you had better have that Autobahn toll sticker called a "Vignette" that has the Hapsburg double eagle on it.
I think it was in 1996 when I drove my pappa from Munich through the Alps to Slovenia and back, and no one said anything to us through a half dozen mountain pass toll stations where you pay a cash toll until we were just short of Germany and we got waved over by the Gendarmerie. We handed over our blue passports to a pair of cops who knew little English, my German is the ungrammatical dialect learned from my grandma who came from the Serbian-Hungarian border, and my father started forgetting his rather fluent Croatian-dialect German real fast.
We were told the fine for not having the toll sticker was the equivalent of about $145, and I was sweating because I had only Deutche Marks on me and no Austrian cash, but after some hemming and hawing, the officer gave us our passports back and told us "Auf Wiederseien." My pop told me, "I think we need to drive over there and pay that fine," and I said "we are only 100 metres from the German border, I think I can make it!"
On later trips we bought the sticker and they had signs in English better explaining what it was, and the German rental car clerk finally told us about it (after asking, even though on the trip we got nicked, we told him we were taking the car through Austria to Slovenia). But can you picture two dumb Americans wasting away in an Austrian jail because they don't have any Austrian money on account of the toll sticker.
The way I figure it is that Austria is kind of like Wisconsin and Germany is kind of like Illinois -- the Germans think of Austrians as hicks and the Austrians think of Germans as those tourists with a lot of money and an attitude. I think the toll sticker is "lets put a tax on ourselves, but it stays in the country, but lets make those smart alec Germans pay it too." I think it is only a matter of time before Wisconsin comes up with a scheme like that.
The standard way to debug anything is to trace execution and to reveal the program state at points along that trace. A debugger with watches, breakpoints, and variable evaluation is perhaps a runtime counterpart to the edit-compile-execute cycles of printf() statements, so apart from the quickness of the debugger, I don't see what is so different with printf(). The debugger is so supposed to be more inobtrusive (printf() masking a bug by shifting stuff around in time or in memory space), but I am hardpressed that a debugger makes such bugs (your thread time-synch problem, stack overflow, variable overwrite) that much easier. You are going to have trouble with a debugger or with printf() in those cases.
Another problem with a debugger is if your app is split into DLL's: how do you debug into a DLL if that is where it is choking? How do you run the DLL stand-alone in that case?
An idea I have tried under Windows is to output the trace to an out-of-process (.EXE style) Automation server program. That Automation server can optionally show the trace in a console window or it can redirect the trace to a file. This is usable with a GUI or a non-GUI app. Also, if the process containing your program shuts down, crashes, or locks up, you still have the trace displayed or saved to disk by another process. Not only that, you can implement some kind of "error level" system or perhaps simply unregister the Automation server when you don't want to see the trace -- I put a singleton object in my app as the connection to the Automation server, and that object can check if the Automation server is available before forwarding the text display.
This Automation server seems to work in just about every case -- except for inside ActiveX components. That's my whole point: you can unit-test stuff all you want, but weirdness happens when you hang it all together. My Automation server linkup locks up on program termination if an ActiveX component is using the Automation server. Hey, its Windows, and who knows what kind of prolog-postlog is happening in wizard-generated ActiveX base object and skeleton code, so the only thing I can think of for tracing an ActiveX is writing to a file, MessageBox() (not always good -- sometimes pops up zillions of message boxes on account of thread issues). If anyone has a better idea of how to trace inside an ActiveX control while that control is used in a designer or an app, I would like to know.
It's simple -- they will feed off old persons' medications. And they will be made of metal and be strong, so the old people won't be able to fight them off.
Great post. If something (i.e. that data-turned-script) is Turing Complete it is also possibly Turing Undecidable.
That is also the nagging feeling I have had about extending applications with scripts. Not only are the scripts possibly violating the principle of least power, but you are handing over to your non-programmer users essentially a programming tool, no matter how sandboxed, hand-held, error-checked, and feature-restricted. At some level programming is programming with all the potential for bugs and bad design.
Is J# a joke? J# is a joke inasmuch as it is not really Java because for something to be really Java you have to have the whole class library. But as far as being a successful "skin" for Java syntax on the CLR, I can't see what is wrong with it. For people with Java syntax under their fingers and who find the minor changes to it in C# to be annoyances, I don't see what the deficiencies are in J# -- what does C# allow you to do that you can't in J#? What about J# would be foreign to a Java programmer?
As for VB# vs VB, I have only done "toy" programs in each and find each equally bizarre, only VB# seems to me less bizarre because it lacks the half-an-OO language restrictions of straight VB.
Perhaps a good analogy is Borland's Delphi and C++ Builder. They both share the same class library and GUI framework underneath, which is written in Object Pascal. The automatically-generated C++ code is a kind of line-by-line translation of Object Pascal. Not only that, C++ Builder can mix source code modules written in C++ or in Object Pascal.
Diesels are a real big deal in Europe right now while the combination of memories of the Olds and Rabbit Diesels together with environmental authorities more concerned about local particulate emissions than global CO2 admission combine to make Diesels a real minor player in the U.S.. I know, European low-sulfer Diesel fuel, blah, blah, but those are the facts on the ground.
European Diesel is a lot cheaper than gasoline at the pump, but they have a couple-hundred-a-year extra tax on Diesel cars (or at least in Austria and Germany). An Austrian was explaining to me how progressive this policy was -- people who drove few miles used gasoline cars while people like salesmen who were always on the road paid the yearly tax on a Diesel and then made up the difference at the pump -- the purpose of the policy was to put the most efficient cars in the hands of the high-mileage drivers.
Hah, don't start thinking that Germans and Austrians are ahead of the curve of dumb laws with unintended consequences. Watching TV, looking at newspapers, and using my limited grasp of German, I figured out that Diesel is much less heavily taxed at the pump on account of the trucking lobby. Since Diesel has less tax, Diesel car owners reap a windfall. To make up for this, the authorities slapped on a high annual license fee on Diesels. I learned about this because the politicians were trying to raise the Diesel pump tax and getting the truckers mad.
So here I thought German politics was so enlightened and that they were so progressive, but then I found they used just as much patchwork legislation as everyone else, and the end result of making Diesels cost effective for high-mileage drivers was just an accident of this process.
Yes, there were communists in the US, and for all I know there were so many communists in the State Department that you could shoot into a staff meeting and hit a Communist.
It seems, however, that when McCarthy held up his "list of Communists", he was "blowing it out his but" (i.e. bluffing) with the hope that a real list of Communists would be revealed to him.
Suppose you are a prosecutor and are to charge someone with a crime. There are tradeoffs to be made in terms of how much investigating do you do before you arrest someone, and once that person is arrested, that starts a clock that your case had better fall in place or you need to let that person go. On the other hand, if you let a murderer stay at large, more people will die, but if you bring the murderer in without a solid case, you may end up blowing the case. The idea is that you sometimes start out with hunches and suspicions and hope a murder weapon turns up from a search warrant.
The Nixon-Alger Hiss matter was investigated by Congress, although it resulted in an eventual perjury conviction for Alger Hiss on account of grand jury testimony he was compelled to give. Nixon admits that his investigation of Alger Hiss was one of the crisis-turning points in his career -- that he, Nixon went out on a limb and the Whittaker Chambers testimony and the "Pumpkin Papers" microfilm and the tracing of the documents to Hiss's type writer came in the course of the thing to save Nixon's butt and put Hiss's in a sling. I guess prosecutors are given more leeway, but in these public arenas a person like Nixon was taking chances -- there is libel law, political fallout, and so on.
I think in McCarthy's case, he started with a bluff based either on a prosecutorial hunch or perhaps a venal desire for fame and political gain, and once he got the ball rolling he couldn't back out only there was no Whittaker Chambers to come forward and make the charges stick. If you ever watch Cops (sort of real life) or Law and Order (fiction, but ripped-off from notorious current crime stories), cops and prosecutors lie all the time in order to induce confessions, admissions, or guilty pleas. If you confess, you don't get to take it back saying, "the police lied that I would get off easy." The difference is that we give police and prosecutors more lattitude, and their lying is done privately or in conference with attorneys. There is more risk with such tactics when someone is tried by Congressional investigation.
Knuth seemed to believe in writing in an interpreted "machine language" (which also had an assembler). He seemed to think his algorithms were more clear if they were expressed in atomistic operations.
Let us assume for the moment that the requirement is to program for Windows -- there are pros and cons that we don't need to get into because they have been hashed over, so lets move on here.
Visual Basic is Microsoft's flagship project for Windows application software development while Visual C++ has been a poor stepchild. Windows has two major features that are hard to do cleanly without some kind of support hard wired into the computer language. One is the one-size fits all WndProc() interface to the Windows API with its zillion messages. The other is the different-shoe-for-every-foot of the zillions of interfaces that make up COM along with the support for QueryInterface to interrogate objects for supported interfaces at run time.
VB has message dispatch and COM support built in as language extensions to Basic. C++ is not supposed to need language extensions because of macros, classes and templates, and Microsoft uses a mishmosh of macros, classes, and templates to support message dispatch (the infamous message map macros) and COM with QueryInterface(). We can have flame wars about languages all we want, but I think that Windows support in VB is pretty transparent because they adapted the language while the Windows support in VC++ is pretty clunky because they have stuck to C++ without extensions, and there may be a cleaner way to use C++ to support message dispatch and QueryInterface(), but Microsoft has not figured it out but come up with a hodge podge (actually several hodge podges: MFC, ATL, and perhaps something else).
Also, VB is truly visual with the Form Designer while I could never figure out how VC++ is visual -- you need to create Windows forms as dialog resources that you have to wizard your way into code somehow. In the Microsoft world, VB is the solution they offer for the cook who wants to develop a recipe program, for the race car team that wants to develop and engine analyser program, for the accountant who wants to develop a financial program. VC++ is the solution they offer for the hard-core programmer who wants to squeeze out efficiency that you can never get in VB, but is willing and able to put in the extra time required for a program that perhaps will have a large shrink-wrap market. VC++ is not meant for the rapid development of the application-specific one-of program.
While Visual Basic is Microsoft's favored child for application software development, its syntax is not modern (i.e. C-like: I am a Pascal person myself, but I admit that C-style syntax has won out), its object model is restricted (lacks class inheritance), and it is slow. A VB program is actually pretty responsive to loading (a dig at Java) and to menu clicks, but it can bog down big time if you have heavy numeric stuff.
To understand where.NET came from, you need to look at Visual J++, a product they developed by hiring Anders "Delphi" Hejlsberg away from Borland. Visual J++ had a C-style syntax (it was Java), but in terms of drag-and-drop form layout, WndProc() handling, and COM/QueryInterface() support, it was everything Visual Basic was and Visual C++ wasn't. It wasn't up to VC++ execution speed, but it was much faster than VB, and it was one of the fasted-executing Java's out there.
J++ was also a marketing failure and the focus of the infamous SUN/Microsoft Java lawsuit. It wasn't proper Java, so it didn't get any Java people to cross over, and it wasn't of interest to the great unwashed masses of VB programmers either. The lawsuit and associated bad publicity about J++ didn't help either.
Microsoft is not one to give up, and.NET is really J++ Version 2.0. Part of.NET and C# is Microsoft taking their football and going home over the Java lawsuit, part is trying a second time with J++. If you have tried Visual C#.NET, you really need to install a copy of Visual J++ 6.0, play with it, and experience the powerful sense of deja vu.
At one end you have complete utilization of biomass and at the other hand you have this business of corn ethanol with questions about whether you get more energy than you put in. Somewhere in the middle is bio-Diesel, where you make a Diesel fuel out of vegetable oil -- I think canola gives something like 100 gallons per acre.
There are Web sites telling how easy it is to make bio-Diesel. The process involves 10 parts vegetable oil plus 2 parts methanol plus some lye to make 10 parts Diesel-usable fatty esters plus 2 parts glycerine that you need to do something with. The process seems intermediate in complexity between soap making and running a meth lab, and these hippie types who say how easy it is to make bio-Diesel probably have some other mid-level process experience involving some mildly dangerous chemistry.
There is talk of running Diesels on straight vegetable oil, but there is caution that you can shellac up the rings and ruin an engine. Forget about Diesel engines -- the other big use of Diesel is in oil furnaces: apart from the road tax, #2 Diesel is the same as #2 home-heating oil.
What would it take to run an oil furnace on straight canola oil? An oil furnace repair requires a $100 service call, but it is nothing like rebuilding an engine, so could vegetable oil be burnt in an oil furnace if you could put up with more maintenance. I think the resale value of my house would increase if winter visitors were greated with the smell of french fries.
But Mr. Fusion only powers the time machine, but to use the time machine you have to get the DeLorean up to the critical speed, and if the Indians shoot an arrow in the tank of the DeLorean you are hosed.
There is a lot more to a chip than having a semiconductor. Using zone refining, it is possible to make silicon to insane levels of purity, allowing greater control when it is doped. As a compound semiconductor, GaS doesn't allow the same level of control of purity and of defects. As to diamond, it is an element semiconductor, but I can't see zone refining or drawing it from a melt as a way to purify it.
But the property of silicon that makes it chip worthy is not so much that of silicon, but that of silicon oxide (i.e. beach sand), that makes a perfect insulating end cap on the silicon crystal that makes the MOSFET possible. GaS has some other kind of FET they need to contend with, and as for diamond, oxides of carbon are gasses, so I don't know what they have in mind there.
The evidence seems rather suspicious to me, especially when one of the archealogists carried that vial of blood in his pocket when he visited the site rather than taking it to the museum lab right away . . .
We all know about the neutron bomb, that was supposed to kill the crews of Warsaw Pact tank columns crossing into West Germany while leaving the civilian infrastructure intact, and we all know the propaganda about the "Capitalist Bomb" that killed people but left their property standing. But was the neutron bomb itself a deception campaign or is their some "physics" for building such a thing?
The Ulam-Teller H-bomb is this Rube Goldberg contraption of a series of effects: a good-sized A-bomb to give off x-rays, a casing to channel the x-rays into Styrofoam, a blanket of deuterium (or lithium deuteride) to be compressed by the Styrofoad given oomph by the x-rays, a central "sparkplug" of plutonium to be compressed and give of yet more neutrons, and finally a U-238 jacket to take all those neutrons and convert them into explosive power.
Is a neutron bomb something as simple as an H-bomb with the U-238 jacket removed? With the plutonium sparkplug (essentially the second A-bomb imploded by the first A-bomb?) removed? Does this thing still work with those modification? If you take off the U-238 jacket and keep the sparkplug, you still have a very dirty bomb with a lot of fission effect. If you take out the fission secondary, are you even able to ignite deuterium to any effect? You can boost an A-bomb handsomly with lots of tritium, but that is an expensive, messy thing with a short half-life to use a lot of.
But the original H-bomb took essentially two, staged fission devices to get anything going with the deuterium (which fuses to produce a lot of energetic neutrons), and the original Teller idea of sticking an A-bomb at one end of a can of deuterium got nowhere -- the thing would have just fizzled. Is it possible that the "neutron bomb" was a fiction? I am thinking that a flood of fusion neutrons from a very low fission yield has to be, otherwise efforts to control fusion for power generation wouldn't be so difficult.
Getting the cosmic background may be an important test of any model. What I am curious about is whether plasma cosmology makes any sense at all on a galaxy to galaxy cluster scale. As I said, missing mass occurs on several scales: on the galaxy scale to make the 'burbs spin around like a phonograph record, on the galaxy cluster scale to hold them together from flying apart.
The "standard model" is that systems such as galaxies and galaxy clusters are in gravitational equalibrium, while the plasma model argues that they don't have to be because of magnetic fields and plasma currents. Does this plasma model make any sense at all even for the medium-scale structures, that those fields and currents are even remotely strong enough? Or is this all quack-science hand waving. The plasma model is not invoking any unknown forces, and there should be some EM dudes out there who can say if the fields, currents, and forces could be anywhere near what they would have to be.
I think this goes under the rubric "Plasma Cosmology." There is missing mass/dark matter in varying amounts, but the argument goes something like this. If you look at the "suburbs" of a spiral galaxy, instead of the the stuff following Keplerian orbits, the stuff appears to be going around like a phonograph record, which means there has to be (a lot) hidden mass in the halo.
There are a couple of alternative hypotheses. One is MOND (Modified Newtonian Dynamics) which says that inertia changes just a little at large distances, just enough to produce that effect.
Another is this "plasma universe" thingy. The idea is that no, gravity is not the force holding together, say, spiral galaxies or galaxy clusters. These things are not in gravitational equalibrium, and these large scale structures you see are plasma fluctuations giving rise to things like galaxy spiral arms or galaxy clusters. These are large scale counterparts to the arc in a Jacob's Ladder science toy, and as such they are not stable gravitationally-bound structures requiring missing mass, they are evanescent structures that come and go as the plasma arcs around.
I asked some of my ECE department colleagues who work on plasma (magnetic-confinement fusion) if they heard about this plasma cosmology and if this was for real or if it was quack science, and they never heard of it.
Is this plasma universe thing quack science? If it is sincere, what are the holes in that theory? Any good references?
Trouble is for us Northern Hemisphere dwellers that the close Mars opposition happens 1) with Mars in a low-declination constellation like Scorpius or Sagitarius, putting it down on the horizon and down in the murk, and 2) at the peak of mosquito season. How are you folks in Oz looking forward to Mars viewing?
Slashdot Mirror
I figure it took him about 10 years to make his way to the tails of the bell curve of all the incandescents in his house. Besides, if he thinks 15 W compact fluorescent is enough for a living room, 25 W for a kitchen, that dude is pretty parsimonious with light. Fluorescents are a big saving over incandescents, but the 4:1 watt ratio is pushing it a bit -- you are going to be a bit dim after the substitution.
What are you supposed to do if you had already replaced your lights with compact fluorescents?
On the other hand, I don't understand the "how" of this dog ever catching a mouse -- this guy would pursue any prey barking and with floppy ears flapping: he was not big in the stealth department. Cats, however, are real killing machines. They are quiet and the "pounce" on their prey.
The real test if the Cindy doll has Microsoft software is if the offending words include Jackson, Open Source, Star Office, or Netscape.
The stock Classic Star Trek plot involving an intelligent machine is when Kirk asks it to divide by zero, calculate pi to all decimal places, or determine the meaning of life, at which point smoke come out and the machine melts into a pile of slag.
In the 1980's you had VAXen, some ran VMS, some ran UNIX, and then you had workstations, SUN, Apollo, SGI (largely UNIX although Apollo was some kind of its own thing): expensive hardware, vendor lock in, only thin source-code compatibility between generations (SUN going from Motorola to Intel (briefly) to SPARC), expensive operations (your grant needed to hire or share one or more "bearded programmers" to operate them and write custom software for them).
Then you had PC's and DOS -- the rootstock of Windows. PC's were like PDP-11's and DOS was like RT-11: low level, cheap, lab computers that an engineer or scientist could hack up their own software to control an experiment. PC's were looked down upon by the CS department-academic computing department establishment as "not real computers", and engineer and scientist types looked at academic computing department types as arrogant so-and-so's who wanted to siphon money out of one's grant. PC's were the second wave of what the Apple-II started -- computers for the masses and liberation from the computer elites.
With PC's you rode the wave -- cheap clone computers, cycles of binary-compatible Intel processors, every imaginable type of hardware (A/D and D/A to control those lab experiments), shrink-wrap software, third-party development tools (MS compilers always blew chunks: Borland's Pascal and later C and C++ was in another league, and you had language choices while UNIX seemed to offer any language you wanted provided it was C or later C++).
And then came Windows. You could ignore Windows 3 and 3.1, but by Windows 95 and NT 4, it was pretty much rammed down your throat. But around that time, the choices were pretty much switching over to the Mac, and the closed-hardware Mac lacked all of the hardware add-ons from Data Translation, and expensive workstations were getting replaced with cheap PC's, so programming for X/Motif wasn't even on the table). If you could live with the performance limitations of Visual Basic, you had an easy way to develop apps; if you wanted performance you had to wade through volumes of the Windows API, but having done that there was kind of a reluctance to learn that much over again to do GUI's on other systems.
So this Linux thing comes along and on '86 hardware as well. And then Bill is Bill: what he had been doing to clone computer makers all along he starts to do to the "little guy." Ghosting over a copy of Windows to a clone you put together from parts for your lab was no big deal until all this fanfare of a crackdown in the form of activation. The University buys a mass MS site license so we all switch from Word Perfect to Word, and this summer the University drops the site license because MS tightens the screws on the terms, and we worry that the BSA is going to bust down the door to our lab.
I have been looking over my shoulder, and I have looked at Java, and I have looked at Linux/Gnome/KDE. Our grant funding is way pared back in these tough times, and we are down to 20 percent share of one UNIX programmer who could care less about setting up a Linux box for the lab, and our engineer guy who is a digital-logic design genius has been thrashing around for over a year trying to get a Linux box up and running but doesn't have the time to swap network cards and fool around with getting drivers.
And just as Java was supposed to make big inroads on the desktop by now, the same was true for Linux. The Engineering library had put in a raft of Linux boxes, but I guess none of the engineering students knew what to do with them that they have all been reverted to Windows.
My engineer logic-design pal and I have talked about how a Linux box would be a good as a Windows box if to the user the thing were just a GUI
That is what I always thought that as you say, that when you fueled the rocket you took shelter in bunkers, pillboxes, and concrete block houses. Were they cutting corners? Was it a solid rocket that went off? Solids are of course always "fueled."
I think it was in 1996 when I drove my pappa from Munich through the Alps to Slovenia and back, and no one said anything to us through a half dozen mountain pass toll stations where you pay a cash toll until we were just short of Germany and we got waved over by the Gendarmerie. We handed over our blue passports to a pair of cops who knew little English, my German is the ungrammatical dialect learned from my grandma who came from the Serbian-Hungarian border, and my father started forgetting his rather fluent Croatian-dialect German real fast.
We were told the fine for not having the toll sticker was the equivalent of about $145, and I was sweating because I had only Deutche Marks on me and no Austrian cash, but after some hemming and hawing, the officer gave us our passports back and told us "Auf Wiederseien." My pop told me, "I think we need to drive over there and pay that fine," and I said "we are only 100 metres from the German border, I think I can make it!"
On later trips we bought the sticker and they had signs in English better explaining what it was, and the German rental car clerk finally told us about it (after asking, even though on the trip we got nicked, we told him we were taking the car through Austria to Slovenia). But can you picture two dumb Americans wasting away in an Austrian jail because they don't have any Austrian money on account of the toll sticker.
The way I figure it is that Austria is kind of like Wisconsin and Germany is kind of like Illinois -- the Germans think of Austrians as hicks and the Austrians think of Germans as those tourists with a lot of money and an attitude. I think the toll sticker is "lets put a tax on ourselves, but it stays in the country, but lets make those smart alec Germans pay it too." I think it is only a matter of time before Wisconsin comes up with a scheme like that.
Another problem with a debugger is if your app is split into DLL's: how do you debug into a DLL if that is where it is choking? How do you run the DLL stand-alone in that case?
An idea I have tried under Windows is to output the trace to an out-of-process (.EXE style) Automation server program. That Automation server can optionally show the trace in a console window or it can redirect the trace to a file. This is usable with a GUI or a non-GUI app. Also, if the process containing your program shuts down, crashes, or locks up, you still have the trace displayed or saved to disk by another process. Not only that, you can implement some kind of "error level" system or perhaps simply unregister the Automation server when you don't want to see the trace -- I put a singleton object in my app as the connection to the Automation server, and that object can check if the Automation server is available before forwarding the text display.
This Automation server seems to work in just about every case -- except for inside ActiveX components. That's my whole point: you can unit-test stuff all you want, but weirdness happens when you hang it all together. My Automation server linkup locks up on program termination if an ActiveX component is using the Automation server. Hey, its Windows, and who knows what kind of prolog-postlog is happening in wizard-generated ActiveX base object and skeleton code, so the only thing I can think of for tracing an ActiveX is writing to a file, MessageBox() (not always good -- sometimes pops up zillions of message boxes on account of thread issues). If anyone has a better idea of how to trace inside an ActiveX control while that control is used in a designer or an app, I would like to know.
It's simple -- they will feed off old persons' medications. And they will be made of metal and be strong, so the old people won't be able to fight them off.
That is also the nagging feeling I have had about extending applications with scripts. Not only are the scripts possibly violating the principle of least power, but you are handing over to your non-programmer users essentially a programming tool, no matter how sandboxed, hand-held, error-checked, and feature-restricted. At some level programming is programming with all the potential for bugs and bad design.
As for VB# vs VB, I have only done "toy" programs in each and find each equally bizarre, only VB# seems to me less bizarre because it lacks the half-an-OO language restrictions of straight VB.
Perhaps a good analogy is Borland's Delphi and C++ Builder. They both share the same class library and GUI framework underneath, which is written in Object Pascal. The automatically-generated C++ code is a kind of line-by-line translation of Object Pascal. Not only that, C++ Builder can mix source code modules written in C++ or in Object Pascal.
European Diesel is a lot cheaper than gasoline at the pump, but they have a couple-hundred-a-year extra tax on Diesel cars (or at least in Austria and Germany). An Austrian was explaining to me how progressive this policy was -- people who drove few miles used gasoline cars while people like salesmen who were always on the road paid the yearly tax on a Diesel and then made up the difference at the pump -- the purpose of the policy was to put the most efficient cars in the hands of the high-mileage drivers.
Hah, don't start thinking that Germans and Austrians are ahead of the curve of dumb laws with unintended consequences. Watching TV, looking at newspapers, and using my limited grasp of German, I figured out that Diesel is much less heavily taxed at the pump on account of the trucking lobby. Since Diesel has less tax, Diesel car owners reap a windfall. To make up for this, the authorities slapped on a high annual license fee on Diesels. I learned about this because the politicians were trying to raise the Diesel pump tax and getting the truckers mad.
So here I thought German politics was so enlightened and that they were so progressive, but then I found they used just as much patchwork legislation as everyone else, and the end result of making Diesels cost effective for high-mileage drivers was just an accident of this process.
It seems, however, that when McCarthy held up his "list of Communists", he was "blowing it out his but" (i.e. bluffing) with the hope that a real list of Communists would be revealed to him.
Suppose you are a prosecutor and are to charge someone with a crime. There are tradeoffs to be made in terms of how much investigating do you do before you arrest someone, and once that person is arrested, that starts a clock that your case had better fall in place or you need to let that person go. On the other hand, if you let a murderer stay at large, more people will die, but if you bring the murderer in without a solid case, you may end up blowing the case. The idea is that you sometimes start out with hunches and suspicions and hope a murder weapon turns up from a search warrant.
The Nixon-Alger Hiss matter was investigated by Congress, although it resulted in an eventual perjury conviction for Alger Hiss on account of grand jury testimony he was compelled to give. Nixon admits that his investigation of Alger Hiss was one of the crisis-turning points in his career -- that he, Nixon went out on a limb and the Whittaker Chambers testimony and the "Pumpkin Papers" microfilm and the tracing of the documents to Hiss's type writer came in the course of the thing to save Nixon's butt and put Hiss's in a sling. I guess prosecutors are given more leeway, but in these public arenas a person like Nixon was taking chances -- there is libel law, political fallout, and so on.
I think in McCarthy's case, he started with a bluff based either on a prosecutorial hunch or perhaps a venal desire for fame and political gain, and once he got the ball rolling he couldn't back out only there was no Whittaker Chambers to come forward and make the charges stick. If you ever watch Cops (sort of real life) or Law and Order (fiction, but ripped-off from notorious current crime stories), cops and prosecutors lie all the time in order to induce confessions, admissions, or guilty pleas. If you confess, you don't get to take it back saying, "the police lied that I would get off easy." The difference is that we give police and prosecutors more lattitude, and their lying is done privately or in conference with attorneys. There is more risk with such tactics when someone is tried by Congressional investigation.
Knuth seemed to believe in writing in an interpreted "machine language" (which also had an assembler). He seemed to think his algorithms were more clear if they were expressed in atomistic operations.
Visual Basic is Microsoft's flagship project for Windows application software development while Visual C++ has been a poor stepchild. Windows has two major features that are hard to do cleanly without some kind of support hard wired into the computer language. One is the one-size fits all WndProc() interface to the Windows API with its zillion messages. The other is the different-shoe-for-every-foot of the zillions of interfaces that make up COM along with the support for QueryInterface to interrogate objects for supported interfaces at run time.
VB has message dispatch and COM support built in as language extensions to Basic. C++ is not supposed to need language extensions because of macros, classes and templates, and Microsoft uses a mishmosh of macros, classes, and templates to support message dispatch (the infamous message map macros) and COM with QueryInterface(). We can have flame wars about languages all we want, but I think that Windows support in VB is pretty transparent because they adapted the language while the Windows support in VC++ is pretty clunky because they have stuck to C++ without extensions, and there may be a cleaner way to use C++ to support message dispatch and QueryInterface(), but Microsoft has not figured it out but come up with a hodge podge (actually several hodge podges: MFC, ATL, and perhaps something else).
Also, VB is truly visual with the Form Designer while I could never figure out how VC++ is visual -- you need to create Windows forms as dialog resources that you have to wizard your way into code somehow. In the Microsoft world, VB is the solution they offer for the cook who wants to develop a recipe program, for the race car team that wants to develop and engine analyser program, for the accountant who wants to develop a financial program. VC++ is the solution they offer for the hard-core programmer who wants to squeeze out efficiency that you can never get in VB, but is willing and able to put in the extra time required for a program that perhaps will have a large shrink-wrap market. VC++ is not meant for the rapid development of the application-specific one-of program.
While Visual Basic is Microsoft's favored child for application software development, its syntax is not modern (i.e. C-like: I am a Pascal person myself, but I admit that C-style syntax has won out), its object model is restricted (lacks class inheritance), and it is slow. A VB program is actually pretty responsive to loading (a dig at Java) and to menu clicks, but it can bog down big time if you have heavy numeric stuff.
To understand where .NET came from, you need to look at Visual J++, a product they developed by hiring Anders "Delphi" Hejlsberg away from Borland. Visual J++ had a C-style syntax (it was Java), but in terms of drag-and-drop form layout, WndProc() handling, and COM/QueryInterface() support, it was everything Visual Basic was and Visual C++ wasn't. It wasn't up to VC++ execution speed, but it was much faster than VB, and it was one of the fasted-executing Java's out there.
J++ was also a marketing failure and the focus of the infamous SUN/Microsoft Java lawsuit. It wasn't proper Java, so it didn't get any Java people to cross over, and it wasn't of interest to the great unwashed masses of VB programmers either. The lawsuit and associated bad publicity about J++ didn't help either.
Microsoft is not one to give up, and .NET is really J++ Version 2.0. Part of .NET and C# is Microsoft taking their football and going home over the Java lawsuit, part is trying a second time with J++. If you have tried Visual C# .NET, you really need to install a copy of Visual J++ 6.0, play with it, and experience the powerful sense of deja vu.
Lets face it, the .NET langu
There are Web sites telling how easy it is to make bio-Diesel. The process involves 10 parts vegetable oil plus 2 parts methanol plus some lye to make 10 parts Diesel-usable fatty esters plus 2 parts glycerine that you need to do something with. The process seems intermediate in complexity between soap making and running a meth lab, and these hippie types who say how easy it is to make bio-Diesel probably have some other mid-level process experience involving some mildly dangerous chemistry.
There is talk of running Diesels on straight vegetable oil, but there is caution that you can shellac up the rings and ruin an engine. Forget about Diesel engines -- the other big use of Diesel is in oil furnaces: apart from the road tax, #2 Diesel is the same as #2 home-heating oil.
What would it take to run an oil furnace on straight canola oil? An oil furnace repair requires a $100 service call, but it is nothing like rebuilding an engine, so could vegetable oil be burnt in an oil furnace if you could put up with more maintenance. I think the resale value of my house would increase if winter visitors were greated with the smell of french fries.
But Mr. Fusion only powers the time machine, but to use the time machine you have to get the DeLorean up to the critical speed, and if the Indians shoot an arrow in the tank of the DeLorean you are hosed.
And there will still be a visible hesitation in Word when you scroll into a page into which a bitmap image has been copied . . .
But the property of silicon that makes it chip worthy is not so much that of silicon, but that of silicon oxide (i.e. beach sand), that makes a perfect insulating end cap on the silicon crystal that makes the MOSFET possible. GaS has some other kind of FET they need to contend with, and as for diamond, oxides of carbon are gasses, so I don't know what they have in mind there.
The evidence seems rather suspicious to me, especially when one of the archealogists carried that vial of blood in his pocket when he visited the site rather than taking it to the museum lab right away . . .
The Ulam-Teller H-bomb is this Rube Goldberg contraption of a series of effects: a good-sized A-bomb to give off x-rays, a casing to channel the x-rays into Styrofoam, a blanket of deuterium (or lithium deuteride) to be compressed by the Styrofoad given oomph by the x-rays, a central "sparkplug" of plutonium to be compressed and give of yet more neutrons, and finally a U-238 jacket to take all those neutrons and convert them into explosive power.
Is a neutron bomb something as simple as an H-bomb with the U-238 jacket removed? With the plutonium sparkplug (essentially the second A-bomb imploded by the first A-bomb?) removed? Does this thing still work with those modification? If you take off the U-238 jacket and keep the sparkplug, you still have a very dirty bomb with a lot of fission effect. If you take out the fission secondary, are you even able to ignite deuterium to any effect? You can boost an A-bomb handsomly with lots of tritium, but that is an expensive, messy thing with a short half-life to use a lot of.
But the original H-bomb took essentially two, staged fission devices to get anything going with the deuterium (which fuses to produce a lot of energetic neutrons), and the original Teller idea of sticking an A-bomb at one end of a can of deuterium got nowhere -- the thing would have just fizzled. Is it possible that the "neutron bomb" was a fiction? I am thinking that a flood of fusion neutrons from a very low fission yield has to be, otherwise efforts to control fusion for power generation wouldn't be so difficult.
The "standard model" is that systems such as galaxies and galaxy clusters are in gravitational equalibrium, while the plasma model argues that they don't have to be because of magnetic fields and plasma currents. Does this plasma model make any sense at all even for the medium-scale structures, that those fields and currents are even remotely strong enough? Or is this all quack-science hand waving. The plasma model is not invoking any unknown forces, and there should be some EM dudes out there who can say if the fields, currents, and forces could be anywhere near what they would have to be.
There are a couple of alternative hypotheses. One is MOND (Modified Newtonian Dynamics) which says that inertia changes just a little at large distances, just enough to produce that effect.
Another is this "plasma universe" thingy. The idea is that no, gravity is not the force holding together, say, spiral galaxies or galaxy clusters. These things are not in gravitational equalibrium, and these large scale structures you see are plasma fluctuations giving rise to things like galaxy spiral arms or galaxy clusters. These are large scale counterparts to the arc in a Jacob's Ladder science toy, and as such they are not stable gravitationally-bound structures requiring missing mass, they are evanescent structures that come and go as the plasma arcs around.
I asked some of my ECE department colleagues who work on plasma (magnetic-confinement fusion) if they heard about this plasma cosmology and if this was for real or if it was quack science, and they never heard of it.
Is this plasma universe thing quack science? If it is sincere, what are the holes in that theory? Any good references?
Trouble is for us Northern Hemisphere dwellers that the close Mars opposition happens 1) with Mars in a low-declination constellation like Scorpius or Sagitarius, putting it down on the horizon and down in the murk, and 2) at the peak of mosquito season. How are you folks in Oz looking forward to Mars viewing?