That recovery CD / OS reinstall CD is absolutely standard across all Dell systems.
There is a second recovery CD that includes drivers for the machine. That CD is updated and changed on a regular basis as Dell adds new hardware to their lineup.
So yes, it means you need two disks -- one for the OS, and one for new drivers for new hardware.
That's no different than any other OS. Even Linux.
Or does your 5 year old linux install CD work on the new video hardware? Probably not.
Will that new linux install CD correctly detect next year's new video hardware?
Seperate, updated driver disks are normal and expected.
Expecting the one OS disk to work with everything? Sure, in 1984 VGA mode.
The last time I tried running as an unprivileged user was windows 2000 terminal server. Uggh.
Windows has groups and users Really? Take "Administrator" and "Administrators" for a moment. I have both of these tagged to my user account. Are they treated as different types, or just two different names, with different security ID's, that are used interchangably?
ACL comment
Unix permissions can do things that Windows either cannot, or are horribly hard to code.
Example: Group cannot do X, owner and world can do X. Assign group name of "Students", and world is the teachers, admins, etc. Now you cannot examine the files of the other students, but the teachers can.
In windows, in order to state that a larger, generalized group can do something, and you cannot, requires listing EVERY other member explicitly. Why?
Because in unix, "Group" means "The rest of the group except me", and "world" means "The rest of the world except my group and except me".
Those exceptions require explicit listing of every user. What happens when the users change?
How can you state "Group X can do this, but specific member Y of group X cannot"?
Run-as administrator? Well, gee...
#1. You've lost all security. To use that, you have to give the admin password to anyone that wants to use that program. No more "This program is safe for anyone to use". You have to trust everyone with the privileged password, which means that "cmd.com" can be run as that secure account.
#2. Fails for group concept. Lets say I want to have the equivalent of "set-GROUP-ID games" on programs that can access a special shared directory, that is group writable by games.
I want these programs to run as user real-person, group games.
In other words, access ability of the user (so they can load personal save game, or save stuff in their personal folders) and group games (so the game can do things like securly update high score files).
Or, take the "standard" linux / redhat security model. Every person is a member by default of their own group. Files inherit group permissions of the directory they are in. Default permissions are group write.
If 20 people are working together on a project, files saved in the group directory are group writable, tagged to the group that the directory belongs to, and writable by any member of that group.
This model works fine even if a person is a member of 15 different groups. The file is tagged with the creator (which can be ignored), the group (generated by the dictory, checked based on the user that wants to access it).
How do I do that setup in windows? And Linux does this with 9 bits.
the fact is that at least since 2003 MS has kicked itself into shape and now has security as the top priority. We're actually seeing for the first time security concerns trumping 'user friendliness', which is great.
I'll believe it when I see some sort of restricted execution mode.
I'll be blunt: Windows needs the set-UID system. The only way to prevent abuses of a "hook" system -- registry (windows), input manager (Mac osX), etc -- is to disallow arbitrary access to the hooks, and require hook access to go through a trusted program.
The only way to make sure that the code that is accessing the hooks is the trusted code is to have some sort of "This app runs with special permissions". That's some form of "set permission on execution", either set-uid, or more likely, set-gid (Except that windows, as far as I know, doesn't seperate group and user concepts).
Is microsoft willing to break backwards compatibility to ensure security?
The bottom line: Any time a program is expected, or permitted, to change a hook, means that an untrusty program can change those hooks and install a virus/worm/trojan that can run as the current user, compromising at least that account, and any system that encourages or requires that a normal home user runs as "root" means that the same security flaw will expose the entire system.
The input manager flaw on OsX (described later down on this page)? Fix it by making the input manager directory only writable by a privileged app, that uses the root input manager instead of the user input manager. Now there is no way for a sneaky program to hook itself into arbitrary apps.
Hooking things into IE, the windows shell, etc: apparently trivial if you know what you're doing, a PITA to try to track down, determine what's what, etc.
Heck, microsoft would do a big service just to introduce a tool that looks through the registry, tells you "here is a list of programs that say they handle _x_, choose the one that you want to actually handle _x_". Except that this might not be possible with the current registry design, as there might only be a place to store the "Currently assigned choice is _y_" result, not the list of all possible choices.
Seriously, why should installing program X give that program the permission to say "I will now declare that extension.xyz is xylaphone yellow zero" as well as "I handle this type" as well as "This is the icon for this type"?
For that matter, who says that ".jpg" is "Microsoft office picture document" or whatever it is called after installing office?
Yes, this is security. That trusted program X can do this sort of thing without any double checks means that any untrusted code Y that can run because of a bug can make any changes that they want.
The point being: You can try to squash bugs and holes all you want; that won't stop someone from clicking on a link to killcomp.com, without realizing that it's an executable:-). It's equally important to prevent runaway code from bringing down the account, or the system, by putting in some sort of restriction.
And once my component is ready, the source is shipped to the security group for one final run through for vulnerabilities.
Ok, so your components are tested, and individually secure. Is the whole system -- how the components hook together, and are hooked together -- ever reviewed? Is the group that reviews all this willing to break compatibility to fix security?
Rather, buffer overflows are trivial to avoid in class assignments (and indeed, small projects).
I try to avoid posting on slashdot, really. But that...
One word: Libraries Two words: Object-Oriented.
All programming is small projects
That cannot be stressed enough. Everything is the interface between small components.
It's when the project grows larger, gets split into multiple program units and gets multiple authors that you really start scratching the surface of industrial strength development (something the armchair developers on/. have never really experienced).
And this is where documentation for these libraries really matters.
Specify what you can depend on. Specify what is guaranteed to be kept true from revision to revision Your language may not enforce programming by contract, but that should not stop you from writing documentation that spells out the contracts.
"Bitrot" refers to the change in function of a routine, a data structure, a database entry (think window's registry), that breaks the behavior of older programs because they assumed that "X" was the invariant when actually "Y" was the invariant. This happens because that invariant was not spelled out in the documentation.
As long as the assumptions of the callers match the invariant guarantees of the callees, you've eliminated at least 75% of all bugs.
As long as each little piece is free of buffer overruns, your whole thing will be put together without buffer overruns.
Now, in the old days, where the programming model was "This subroutine just copies all the way until the zero byte that terminates the string", you might have been able to excuse the problem of putting all the pieces together. It was easier to just write code to make one pass through data, process it, and assume no one was trying to crash you. Today, between length-limited copy routines, two passes (to count and check length first, process second), or libraries that manage length-tagged data structures, there is no reason -- NONE -- for overflows in modern programming.
After all, if you spend the extra time to make sure it is done right, well, it's not like you'll lose market share to people that just rush crap code out, right? I mean, people that rush crap, broken code out will get sued for defective products, just like if (for example) a car was released that broke down once a week, or if a toaster was released that could not eject the bread, or if an airplane's control hardware would deliberately override your control stick pushes...
Ultimately, that's the problem. It's not the software. We can write bug free software today. We might have to start over and rewrite all the libraries based around what we know now about software design, but we could do it.
But as long as people can say "This software has no warranty, may not work at all, and we cannot be sued for a defective product", as long as software can ignore all the product liability laws, people have no reason to write that way. After all, the competition is rushing bad software out and selling, we either ship or die. Right?
Sure. Freecell and 2 suit Spider both have the important factors: Playable, easy to learn, winnable, but neither too easy and boring, nor too hard and impossible. You've got an acheavable goal, and a reasonable (5-20 minutes) timeframe.
What have been some of my major favorites over the years?
Magic: The Gathering. Over a 10-30 minute time frame, you could play a duel, and win or not. At the toughest difficulty level, wins were not 100%, heck I suspect 50% overall, but still enough that you could gain and improve. You had a difficult time in assaulting the dungeons and castles, and that required being able to set aside 3-4 hours without having to reboot (difficult, on windows:-), and you'd need to make major deck improvements first, but it was doable.
Yes, you had additional challenges -- the deck you'd want to take into the mountain areas (big red monsters) wasn't the same you'd use in, say, black or green, so you'd have deck adjustments. [Side note: The only time I had a "one deck works everywhere" was a green/white denial deck. COP green and red, designed for delaying the game, massive life growth (essential in the dungeons), and run them out of cards.]
What else is a good game? XCom. 20-45 minutes to prepare for a battle. Then, at another time, 20-45 minutes to fight that battle. Later, 20-45 minutes of headquarters time until the next ship to fight shows up.
What else? 1830. Nethack. A lot of "-craft" games (Dune 2, Super Dune 2, Warcraft (1 and 2), Impossible creatures (beta), etc. As much as I *WANTED* to love dungeon craft, err, Dungeon Master, it STANK, at least partially because it was released as a -craft game.)
Heroes of Might and Magic? Yes, but those levels were too long to finish in one sitting. After a while, they stared to look alike -- I never got past the 4th of the 8 levels.
The bottom line? Games that you can acheive something useful in a relatively short setting (10-60 minutes), with an ongoing reason to come back.
Compare that to the games that have been on my back burner for too long:
NOLF -- I've played 2 or 3 levels. The current level I started so badly (protect the stupid senator) that I'm going to have to restart that level. The "Oh no, I've got to go through that all over again?" factor keeps me from playing the rest, even though my friend who gave me his copy tells me that the story alone is worth playing for.
Half Life and System Shock 2 -- Again, I'm told that they are very enjoyable, with a story that makes them worthwhile. But they are BIG. Maybe when I have time for such long games...
Titans of Steel 1.4: I'm a big fan of Mechforce/Amiga. Yes, I know that this is a "remake" for the PC. But aside from playing a quick game with random teams, if I'm going to play a campaign, I've got a lot of startup work. What is the recycle time of this weapon? How much damage does it do? What is the damage per time? What is the heat done to you? Damamge per hear? How much coolant do you need? Etc. Designing the mechs is a major factor. Just use stock mechs? Well, you've got to examine the 60 or so starting mechs, decide if you want a team of 3 larger, or 5 smaller, or 4 mixed.
The point? There's a huge startup expenditure of time studying the game specs before you can begin playing seriously. And, when I did finally examine the damage/cool factor of the weapons, I found that napalm missiles were far and away the best weapon, but none of the stock mechs were designed around a 3 napalm launcher design (I was able to make a larger starting mech that had it).
And what about the games where I was never able to get past a certain point? Starcraft, protoss 7, where the enemy has a major tech advantage? Heck, even War1's "bugged human" level didn't stop me -- I could complete that level without building my town at all.
The point of this? 1. Keep the time frame focused. IF you require the players to put in several hours of time at a time, then sure, you'll sell
Yes, I think of trusted computing. I think I'd like it.
Oh, you mean do I trust microsoft? Yes, I trust them to mess things up.
Oh, you mean do I trust microsoft to implement trusted computing? Yes, on the XBox. After all, only signed code can run there, guaranteeing that the code is exactly as intended by the code maker (that's the definition of trusted computing, right?)
After all, nobody has an XBox that permits unsigned code, except maybe developers.
It's like saying "My audio player will talk directly to the hardware, so I'll make sure that my 5.1 decoding only goes to the hardware. It can't possibly go to a software virtual device that then goes back into a file for someone to share".
Gee, and I used GCC to compile my code, so I know it can be trusted. After all, was it ever bootstrapped from pcc? Nah, couldn't be.
Trusted computing is provably impossible. You can never determine at any one layer if you can trust the layer beneath you to behave properly.
An application can never trust the OS. (compatibility emulation). The OS can never trust the hardware (emulation). CPU can never trust the boards No board can trust that another board isn't sending it fake signals.
Etc. You can always be put into a hostile environment.
Hmm... this started off going for funny, but now seems serious.
Prior to the world wide web, back on NeXTStep machines, in 1993? Oh yea, I remember that.
Lets see, lots and LOTS of data published in FTP, with an indexing service (the previous "google") able to tell you what was available on all the public FTP servers, and even an NFS to FTP gateway, so that all the ftp sites out there looked like a regular part of the file system. (That one worked real well on high speed connections, but not well on dialup. Primarily because it ignored any NFS request that wasn't available in the cache, but waiting until a re-request came later and the fetching was done.)
And, other people have mentioned Usenet. Or do you remember that prior to 1993 Usenet actually had enough signal to make it viable? (I've long since stopped using usenet because of the sheer size and low signal).
And Gopher, and WAIS, and.........
No, the real truth is this:
Prior to 1993, there was X-Trek, and the non-X-Trek.:-)
"We must absorb the non, until it cesses to be non" (or something like that.)
Now, with a slight alteration to what you said,
The World-Wide-Web is about making profit not some crazy hippie idea of freely exchanging knowledge for the benefit of mankind.
Want proof? Look at the ads here on slashdot:-). Try surfing with graphics turned off. Gee, why am I getting a lot more ads now than before? Why did it load that graphic? Oh -- it's a flash, not a graphic. Hmm... no option anywhere in IE to turn them off, is there?
First, you won't solve the problems by making the _CORPORATIONS_ criminally responsible.
Second, it's not the customers demanding "low quality now" that causes the problems.
The brief answers: You don't want to punish shareholders and employees for actions of the board and vice presidents. (Think AA after Enron). Customers cannot demand a certain quality unless that quality is actually out there -- right now it's all too often a choice of several different things that are bad in different ways.
The long answer: Yes, I agree that companies should be liable for bad products. I agree that product liability laws should extend to software. I think it's stupid to allow companies to say "This product is not guaranteed to work, and may have no actual use for what we market it for" (ok, that's not the actual words, but every software product has that disclaimer, even OS's.).
But you want to target the right people. If you go after the company, and put the company out of business, what happens? Answer: Shareholders lose. Bond holders get pennies. Banks collect what they are owed at the bankrupcy auctions. Employees are out of work. The number of companies in the business go down, so there is less competition => consumers lose.
But have you hurt the people responsible for this disaster? Not really. They might even get a job with another company:-). Or start another one.
Choosing good quality products? Well,...
I'll try to explain by two examples how hard this is. First, let look at a VCR.
What do you want a VCR to do when it reaches the end of the tape?
Choices: 1. Do nothing. 2. Rewind. 3. Eject 4. Rewind and eject 5. Rewind and replay.
Now, I don't know about you, but to me, the answer could be different depending on:
1. Viewing a regular tape. 2. Viewing a non-erasable (rental?) tape. 3. Recording in regular mode 4. Recording in timer mode 5. Whether "memory" mode is turned on.
Now, when was the last time you even checked this behavior before buying a VCR? Or met a sales person that even considered, let alone knew, the answer to this?
Consider telephone answering machines. Consider digital ones.
Can I play back new messages, stop after some of the new, and not have all of them marked "Old"?
Can I play back a bunch of new messages, skip some, and leave those skipped ones as new?
If two people share the machine, can I play the "tape", and say "leave that as new, mark that as old" to explicitly leave messages for my roommate?
Again, when was this ever even listed on that 4x6 card at Fry's?
Demand quality? Sure. Tell me who to talk to. Or do you actually expect me to track down ALL of the companies that make answering machines, and explain to all of them what I want, when most of them have no way for me to talk to a person? [ Side note: In economic theory, this is why middlemen exist -- if you have a lot of sellers, and a lot of buyers, there are much fewer transactions if every maker sells to a middleman, and every purchaser buys from a middleman. It would be much easier to have "product enhancements" if everyone that wanted to say "improve this" only had to go to one place, and product makers only had to check one "What do customers want" place. Now tell me how to make money running such a web site and it'll spring up.]
Haven't you noticed that most companies now only have automated help lines, with no ability to speak to a real person, unless (maybe) you've purchased a new product within 30 or 90 days and need help getting it set up. But product improvement?
I finally got so upset with bugs in the Dish Network receivers that I started calling, and reporting them. I also reported some "I'd like to see this" feature requests. I reported two such things, had two different people at the other end say "Hey, that sounds really good". But when I tried to report a 3rd, I was told "We do
Display Postscript and NeXT, and apparently the PDF support in OS/X, are all terrible in comparison. All of them rely on you creating a window and "context" with completely unrelated calls, before you can finally draw in Postscript. Making the window the same shape as an object you draw is incredibly difficult or impossible. Passing the same color you are drawing to some other interface is difficult or impossible. Synchronizing your drawing with changes to the windows is difficult or impossible. Getting a mouse event and figuring out what part of your graphic was clicked is often an enormously complex and obtuse API.
Reply: I played around with this back in the 3.2 days. Creating a windows WAS a postscript call. A simple one. It would give you a block of whitespace on the monitor that you could draw in. It was NOT the same as an appkit window, which was additional stuff based on this postscript window.
If you want to create an Appkit-level window in DPS, then you have to move all of appkit into DPS.
While I'm not an X programmer, it looks (based on using X, and a lot of different X apps and window managers) like X is more based on "Give the programmer the ability to tweak arbitrary pixels on the screen, let them limit their drawing into this clipping region (window) if they want, and give them (and only them) control over the content of this clipping region". All the "window" stuff comes from another program called the window manager. Well, that low level, "you control the contents of this clipping region" is exactly what DPS gives you.
Similarly, display contexts are manipulated with direct postscript. Oh sure, there's nice ObjC messages to make it fit the language. But it's in DPS at the low level.
Synchronizing your actions with the display? windowdeviceround if I remember correctly does that, as a zero argument postscript command.
Mouse and/or keyboard events?
There was a global postscript file that affected the system. All events went through this file before going to apps. With a relatively simple mod (and this was where I started learning postscript, learning how this worked) you could modify all the system level behaviors, such as keyboard commands for cycling windows, activating windows, etc. Yes, you could read the keyboard, and send commands anywhere; you could check what the mouse was doing, etc. The only problem with this? It triggered the nasty swap space bug in NeXTStep 3 (bug was still there in OpenStep 4, btw -- massive growth and non-shrinking of swapfile).
Window the same shape as your drawing? Best I could recommand is to draw a 100% transparent clear into your box first. Then draw normally. 100% clear will pass the mouse through, and will act like it's not there.
Where is the mouse compared to where my window is? I think you could ask the window manager, in postscript, to tell you either the 4 corners, or 1 corner and two lengths; you get the mouse position from the event (in postscript), and do some adds and subtracts.
Scaled windows? Not sure. Rotated windows? No. Disconnected parts? 100% transparent plus clipping path.
No, you just make a copyright of how the printing press works. Sure, patents expire, but corporate copyrights just keep going (label A:) and going and going and going and going and going (goto A)
Hang on... The day _WILL_ come when there are only a few ISPs?
Today we have:
1. AOL.
Distant #2: Earthlink.
Way way off in the distance, some tiny startups: juno, people pc, walmart.
Other than that? Regonal tiny companies, like phone companies or city-wide ISPs.
If you want a nation wide ISP that you can use while traveling with your laptop, you've got 5. And only 2 of them are any size at all.
Keep in mind: AOL spent 10-12 years purchasing market share at a major loss before finally starting to make a profit. AOL's investors and VC's spent their money buying a virtual dominance if not outright monopoly. Even today, what does Microsoft do to inflict its new email fiat on everyone? That's right. Gets AOL to sign up, and ignores earthlink, walmark, phone companies, etc. They don't matter. They're too small.
According to the supreme court, if I'm accused of 30 counts of drug dealing and sexual molestation, and 1 count of speeding, then when I'm convicted of speeding, I can be sentenced as though all 31 counts were guilty, because "not guilty is not the same as innocent".
According to the supreme court, the Menendez brothers (remember them?) were well served in the requirements of justice by having seperate juries, a prosecuting attourney that argued before both juries that their opponent was guilty of this crime that only one of the two could have done, getting guilty convictions against both of them, and having both serve life sentences. IF you didn't understand that: Here you have 2 nasty people. Both have committed crimes. But for this one crime, at most 1 of these two can be guilty of this crime, yet both are found guilty. You are guaranteed to have one innocent person in these two, yet both the guilty and the innocent are sent to life imprisonment, and the supreme court says "OK, both were found guilty by the due process of the law".
Jury nullification. The jury has the right and power to return "not guilty" even when the law is technically violated, when the act in question is done. That's been upheld by both appellate and supreme courts. The supreme court has said that it's not a mistrial if the jury is NOT informed of this right. Yet trial court judges will insist to juries that they must return guilty if the law is technically violated, and no appallate nor supreme court has said "trial courts cannot instruct the jury that way".
And don't get me started on either jurisdiction (carefully read the first sentence of the 14th amendment, and ask yourself why the federal government and federal courts gets involved in intRA-state matters) or tax law issues:-).
Keep in mind: The preamble of the constitution:
"We the people, of the united states of america, in order to do A, B, C, and D, do establish this constitution". That's a statement of intent and purpose if I ever saw one.
So what is the intent of the constitution? 1. Form a more perfect union. 2. Establish justice. 3. Ensure domestic tranquility. 4. Provide for the common defense. 5. Promote the general welfare 6. Secure the blessings of liberty for ourselves and future generations.
What else does the constitution say about the power of the courts? That it is in one supreme court, and any lesser courts established by law.
*** The constitution requires that our courts are courts of JUSTICE (#2). Any judge that says "This is a court of law, not a court of justice" is in violation of the constitution, yet the supreme court historically upholds that belief.
I haven't looked at this since Java 1.0 or 1.1, so this might be out of date.
Java "binaries" are restricted in two ways:
#1: It must fit the machine's CPU -- the byte ocde sequence which is legal for the JVM.
#2: It must fit the restirctions imposed by the class loader.
#2 is critical. Years ago, there was an Objective C compiler that compiled to Java byte code. The implementation managed to do everything except (if I remember correctly) poseAs, and I'm not sure about categories (I believe that did work, don't quote me).
The problem? It was legal JVM code, but not legal according to the class loader.
C++ and Java are NOT closely related. Years ago I worked for a company that had a data architechture that started from (and still needed to support) 8 bit device drivers on IO boards; at the other end, they ported a large C code system into Java, and made it instantiable -- specifically to permit multiple copies of this large system to run at the same time. The entire high end of the system needed to suport all of C++, Java, and Objective C (this was running on Mac Os X).
The design results? C++ was the limiter. Both Objective C and Java were effectively identical -- either of them could do what the other could do, they could talk to each other (apple's native bridge), etc. But C++ was highly limited. The inheritance system of C++ is vastly weaker and more limited, the bridging for C++ had to be completely manual, etc.
I'd be suprised if you could get a C++ compiler to compile to a Java backend. It seems to me that the basic assumptions of the languages is just too different.
Sure, you can't keep the URL (unless that's up for auction), but the basic product -- an alternative music distribution channel for people that want more money for their songs than they could get from the big 5 record companies -- is still viable.
Mp3.com didn't die because of a bad business model. It died because of a bad court decision from a bad law, and bad advice from lawyers. Don't forget, after VU took over mp3.com, the new mp3.com sued that law firm for bad advice (now that they had the money to sue the lawyers:-).
From my personal experience, the UI of mp3.com went downhill after VU took it over. It went from an easy to use site to an impossible to use site, so I stopped using it. So did a lot of people. Then, it was dropped and put out of business.
So lets reopen it. The guys that built the original mp3.com know how to do it, they could be the advisors even if they can't (because of lawsuit) run it themselves.
This time around, people know what they are doing. After buying the servers, you've got much of the system code. You should be able to restart it for a 1/10 of the cost this time around, especially with the lower programmer costs:-).
Hey, I'm unemployed, and willing to work with this company!
Lightweight question: Will Ziggo win? Or will someone with a robot specifically designed to take it out succede? (Answer: It was taken out by a specialty bot in its first TV appearance).
Heavyweights: Will "Son of Whyachi" make it to the end? Will it defeat BioHazard?
Will Torro manage the full season without a technical problem that kills its flipper? Even if it does, can it win the shoving match?
Will Taz-Bot manage to outdrive, and out-manuver BioHazard this year? Or will it be a perrenial #2-#4 bot?
Will the 6 wheeled pusher bot outshove the others? If it can outdrive and get to the side of the others, it has more pushing power than anyone else (sorry, I don't remember the name).
Or will some unknown new bot make it all the way to the quarter-finals, semi-finals, or even take the victory?
Who will win the heavyweight division? Do you know? Or do you want to bet?
Yes, you can gamble on BattleBots. I would never have guessed that Son of Whyachi would have abandoned that excellent walker base, which gave it both extra weight for the weapons and base (more stable, less flippable), and extra clearance/distance between the weapons and the base (better able to inflict damage without taking it), for a "standard" wheeled base and a faster weapon spin. My prediction: It will take too much internal damage from feedback when it hits, and not make it to the final just from wear and tear. Care to bet against me?
Is BattleBots too predictable? I don't know. For you football fans, do all the lineups fit into 1 of three basic categories (4 blockers and 3 receivers on front, QB and ball snapper, 2 ends as the basic lineup, a couple of variations on that). Do people trying new things, like a shotgun (QB starts at a distance, so that the time spent running backwards is saved) abandon those because this is the "tried and true" system that works?
Do all the teams on baseball agree on where the 9 outfielders should be placed (3 distant, 4 along the baseline, pitcher and catcher)?
Is the question of which teams will be in the top 4 pretty well known before the season begins?
BattleBots is MORE changable than any established sport, just because anyone can enter.
Yes, a few (1 or 2) bots have been sold on Ebay. Most are made by the user. Some are totaled. Some are rebuildable/repairable.
I'm sorry, but the average human mind CANNOT decrypt pig-latin in real time.
Even if you restrict it to the average english speaking person and english-pig-latin, it still doesn't hold.
Take it from someone that never learned pig latin as a child. I'm not joking.
You can learn any language as a child, and speak it in real time. Does that mean that translating something into a foreign language isn't encryption? There was one war where messages were encrypted by translating into an indian language (I think Cherokee, not certain), and it was the only code not broken in that war. (Vietnam? Again, not certain).
As someone who learned pig-latin only about 5 years ago, I can assure you that "real-time" for english is 600 words per minute reading, while enlgish-pig-latin is about 20-60 words per minute. That's slower than I speak, by a LONG shot.
Don't assume that just becasue you learned something in childhood that it's a trivial real time decryption. That just means that you're bi-lingual, nothing more.
I was one of the people that received an email saying "You can get red hat IPO through e-trade". I called e-trade, got a verbal quiz, and then told "You don't qualify". Never got around to opening an account because I was pre-disqualified.
Now I find that there are class action suits, and NASD arbitration reviews. How do I get involved in this?
The supreme court has ruled that paper gains on stocks is NOT taxable income. The basis for this was the realization that a person who had a controlling interest in a company might find that in order to pay the "paper gain" taxes, they had to sell their controlling interest in the company. This, according to the S.C., was unacceptable. Therefore, paper gains are not taxable, only the actual gain/loss when the actual sale occurs.
What you are proposing here actually is just as bad: I'm given stock options. All of a sudden, I owe money that I don't have for paper gains. I do, however, have the option of going into debt to the company (call it the "company loan store"), and at a later time, when I sell, if they are still valuable, paying them off, with interest. Gee, I didn't make enough to pay off my loan on my stock options? Oh well, gotta work more next month, and next month, and...
Companies paying in company script, redeemable only at company stores, forcing people into debt is illegal in this country (it is considered slavery). Your proposal brings it back under a different form.
To use this, you do not need to have a name server; you need to have access to someone else's name server. And, you need a central "modified" server somewhere that you can talk to.
I haven't seen a system yet that didn't have some way to look up names, even if it could not actually connect to them.
So, to use this you need:
1. A linux system.
2. A toll free limited account (quickbooks, MSN signup, your bank's "bank online" feature, etc).
3. The remote DNS that you dial into configured to check the root DNS servers (normal:-)
4. A special server set up somewhere (come on guys, post your server names for us to use).
Then, you start up your client by giving the number address of your limited DNS, and the name of the remote tunnelling DNS.
From the README:
::
Then start the server on one end:
./nstxd tun.yomama.com
and the client on the other end:
./nstxcd tun.yomama.com 125.23.53.12
125.23.53.12 has to be a DNS-server which can be reached by the client-side. The server *must* run on a server where an NS-record for tun.yomama.com points to. So if the server has the IP 1.2.3.4 there must exist an entry in the zonefile of yomama.com: tun IN NS 1.2.3.4
::
A server runs somewhere on the internet. I, the client, need the name of that server, and a DNS operated by my limited ISP whose number I know or can find out. PPPD supports modifying the/etc/resolv.conf file when the ISP provides DNS information, so I can read it from that.
Slashdot Mirror
I've seen a bunch of Dell systems. Guess what?
That recovery CD / OS reinstall CD is absolutely standard across all Dell systems.
There is a second recovery CD that includes drivers for the machine. That CD is updated and changed on a regular basis as Dell adds new hardware to their lineup.
So yes, it means you need two disks -- one for the OS, and one for new drivers for new hardware.
That's no different than any other OS. Even Linux.
Or does your 5 year old linux install CD work on the new video hardware? Probably not.
Will that new linux install CD correctly detect next year's new video hardware?
Seperate, updated driver disks are normal and expected.
Expecting the one OS disk to work with everything? Sure, in 1984 VGA mode.
I use XP on a daily basis.
...
The last time I tried running as an unprivileged user was windows 2000 terminal server. Uggh.
Windows has groups and users
Really? Take "Administrator" and "Administrators" for a moment. I have both of these tagged to my user account. Are they treated as different types, or just two different names, with different security ID's, that are used interchangably?
ACL comment
Unix permissions can do things that Windows either cannot, or are horribly hard to code.
Example: Group cannot do X, owner and world can do X. Assign group name of "Students", and world is the teachers, admins, etc. Now you cannot examine the files of the other students, but the teachers can.
In windows, in order to state that a larger, generalized group can do something, and you cannot, requires listing EVERY other member explicitly. Why?
Because in unix, "Group" means "The rest of the group except me", and "world" means "The rest of the world except my group and except me".
Those exceptions require explicit listing of every user. What happens when the users change?
How can you state "Group X can do this, but specific member Y of group X cannot"?
Run-as administrator? Well, gee
#1. You've lost all security. To use that, you have to give the admin password to anyone that wants to use that program. No more "This program is safe for anyone to use". You have to trust everyone with the privileged password, which means that "cmd.com" can be run as that secure account.
#2. Fails for group concept. Lets say I want to have the equivalent of "set-GROUP-ID games" on programs that can access a special shared directory, that is group writable by games.
I want these programs to run as user real-person, group games.
In other words, access ability of the user (so they can load personal save game, or save stuff in their personal folders) and group games (so the game can do things like securly update high score files).
Or, take the "standard" linux / redhat security model. Every person is a member by default of their own group. Files inherit group permissions of the directory they are in. Default permissions are group write.
If 20 people are working together on a project, files saved in the group directory are group writable, tagged to the group that the directory belongs to, and writable by any member of that group.
This model works fine even if a person is a member of 15 different groups. The file is tagged with the creator (which can be ignored), the group (generated by the dictory, checked based on the user that wants to access it).
How do I do that setup in windows?
And Linux does this with 9 bits.
the fact is that at least since 2003 MS has kicked itself into shape and now has security as the top priority. We're actually seeing for the first time security concerns trumping 'user friendliness', which is great.
.xyz is xylaphone yellow zero" as well as "I handle this type" as well as "This is the icon for this type"?
:-). It's equally important to prevent runaway code from bringing down the account, or the system, by putting in some sort of restriction.
I'll believe it when I see some sort of restricted execution mode.
I'll be blunt: Windows needs the set-UID system. The only way to prevent abuses of a "hook" system -- registry (windows), input manager (Mac osX), etc -- is to disallow arbitrary access to the hooks, and require hook access to go through a trusted program.
The only way to make sure that the code that is accessing the hooks is the trusted code is to have some sort of "This app runs with special permissions". That's some form of "set permission on execution", either set-uid, or more likely, set-gid (Except that windows, as far as I know, doesn't seperate group and user concepts).
Is microsoft willing to break backwards compatibility to ensure security?
The bottom line: Any time a program is expected, or permitted, to change a hook, means that an untrusty program can change those hooks and install a virus/worm/trojan that can run as the current user, compromising at least that account, and any system that encourages or requires that a normal home user runs as "root" means that the same security flaw will expose the entire system.
The input manager flaw on OsX (described later down on this page)? Fix it by making the input manager directory only writable by a privileged app, that uses the root input manager instead of the user input manager. Now there is no way for a sneaky program to hook itself into arbitrary apps.
Hooking things into IE, the windows shell, etc: apparently trivial if you know what you're doing, a PITA to try to track down, determine what's what, etc.
Heck, microsoft would do a big service just to introduce a tool that looks through the registry, tells you "here is a list of programs that say they handle _x_, choose the one that you want to actually handle _x_". Except that this might not be possible with the current registry design, as there might only be a place to store the "Currently assigned choice is _y_" result, not the list of all possible choices.
Seriously, why should installing program X give that program the permission to say "I will now declare that extension
For that matter, who says that ".jpg" is "Microsoft office picture document" or whatever it is called after installing office?
Yes, this is security. That trusted program X can do this sort of thing without any double checks means that any untrusted code Y that can run because of a bug can make any changes that they want.
The point being: You can try to squash bugs and holes all you want; that won't stop someone from clicking on a link to killcomp.com, without realizing that it's an executable
And once my component is ready, the source is shipped to the security group for one final run through for vulnerabilities.
Ok, so your components are tested, and individually secure. Is the whole system -- how the components hook together, and are hooked together -- ever reviewed? Is the group that reviews all this willing to break compatibility to fix security?
Rather, buffer overflows are trivial to avoid in class assignments (and indeed, small projects).
...
/. have never really experienced).
...
I try to avoid posting on slashdot, really. But that
One word: Libraries
Two words: Object-Oriented.
All programming is small projects
That cannot be stressed enough. Everything is the interface between small components.
It's when the project grows larger, gets split into multiple program units and gets multiple authors that you really start scratching the surface of industrial strength development (something the armchair developers on
And this is where documentation for these libraries really matters.
Specify what you can depend on.
Specify what is guaranteed to be kept true from revision to revision
Your language may not enforce programming by contract, but that should not stop you from writing documentation that spells out the contracts.
"Bitrot" refers to the change in function of a routine, a data structure, a database entry (think window's registry), that breaks the behavior of older programs because they assumed that "X" was the invariant when actually "Y" was the invariant. This happens because that invariant was not spelled out in the documentation.
As long as the assumptions of the callers match the invariant guarantees of the callees, you've eliminated at least 75% of all bugs.
As long as each little piece is free of buffer overruns, your whole thing will be put together without buffer overruns.
Now, in the old days, where the programming model was "This subroutine just copies all the way until the zero byte that terminates the string", you might have been able to excuse the problem of putting all the pieces together. It was easier to just write code to make one pass through data, process it, and assume no one was trying to crash you. Today, between length-limited copy routines, two passes (to count and check length first, process second), or libraries that manage length-tagged data structures, there is no reason -- NONE -- for overflows in modern programming.
After all, if you spend the extra time to make sure it is done right, well, it's not like you'll lose market share to people that just rush crap code out, right? I mean, people that rush crap, broken code out will get sued for defective products, just like if (for example) a car was released that broke down once a week, or if a toaster was released that could not eject the bread, or if an airplane's control hardware would deliberately override your control stick pushes
Ultimately, that's the problem. It's not the software. We can write bug free software today. We might have to start over and rewrite all the libraries based around what we know now about software design, but we could do it.
But as long as people can say "This software has no warranty, may not work at all, and we cannot be sued for a defective product", as long as software can ignore all the product liability laws, people have no reason to write that way. After all, the competition is rushing bad software out and selling, we either ship or die. Right?
Sure. Freecell and 2 suit Spider both have the important factors: Playable, easy to learn, winnable, but neither too easy and boring, nor too hard and impossible. You've got an acheavable goal, and a reasonable (5-20 minutes) timeframe.
:-), and you'd need to make major deck improvements first, but it was doable.
...
What have been some of my major favorites over the years?
Magic: The Gathering. Over a 10-30 minute time frame, you could play a duel, and win or not. At the toughest difficulty level, wins were not 100%, heck I suspect 50% overall, but still enough that you could gain and improve. You had a difficult time in assaulting the dungeons and castles, and that required being able to set aside 3-4 hours without having to reboot (difficult, on windows
Yes, you had additional challenges -- the deck you'd want to take into the mountain areas (big red monsters) wasn't the same you'd use in, say, black or green, so you'd have deck adjustments. [Side note: The only time I had a "one deck works everywhere" was a green/white denial deck. COP green and red, designed for delaying the game, massive life growth (essential in the dungeons), and run them out of cards.]
What else is a good game? XCom. 20-45 minutes to prepare for a battle. Then, at another time, 20-45 minutes to fight that battle. Later, 20-45 minutes of headquarters time until the next ship to fight shows up.
What else? 1830. Nethack. A lot of "-craft" games (Dune 2, Super Dune 2, Warcraft (1 and 2), Impossible creatures (beta), etc. As much as I *WANTED* to love dungeon craft, err, Dungeon Master, it STANK, at least partially because it was released as a -craft game.)
Heroes of Might and Magic? Yes, but those levels were too long to finish in one sitting. After a while, they stared to look alike -- I never got past the 4th of the 8 levels.
The bottom line? Games that you can acheive something useful in a relatively short setting (10-60 minutes), with an ongoing reason to come back.
Compare that to the games that have been on my back burner for too long:
NOLF -- I've played 2 or 3 levels. The current level I started so badly (protect the stupid senator) that I'm going to have to restart that level. The "Oh no, I've got to go through that all over again?" factor keeps me from playing the rest, even though my friend who gave me his copy tells me that the story alone is worth playing for.
Half Life and System Shock 2 -- Again, I'm told that they are very enjoyable, with a story that makes them worthwhile. But they are BIG. Maybe when I have time for such long games
Titans of Steel 1.4: I'm a big fan of Mechforce/Amiga. Yes, I know that this is a "remake" for the PC. But aside from playing a quick game with random teams, if I'm going to play a campaign, I've got a lot of startup work. What is the recycle time of this weapon? How much damage does it do? What is the damage per time? What is the heat done to you? Damamge per hear? How much coolant do you need? Etc. Designing the mechs is a major factor. Just use stock mechs? Well, you've got to examine the 60 or so starting mechs, decide if you want a team of 3 larger, or 5 smaller, or 4 mixed.
The point? There's a huge startup expenditure of time studying the game specs before you can begin playing seriously. And, when I did finally examine the damage/cool factor of the weapons, I found that napalm missiles were far and away the best weapon, but none of the stock mechs were designed around a 3 napalm launcher design (I was able to make a larger starting mech that had it).
And what about the games where I was never able to get past a certain point? Starcraft, protoss 7, where the enemy has a major tech advantage? Heck, even War1's "bugged human" level didn't stop me -- I could complete that level without building my town at all.
The point of this?
1. Keep the time frame focused. IF you require the players to put in several hours of time at a time, then sure, you'll sell
Grr. The internet ate my first response.
Yes, I think of trusted computing. I think I'd like it.
Oh, you mean do I trust microsoft? Yes, I trust them to mess things up.
Oh, you mean do I trust microsoft to implement trusted computing? Yes, on the XBox. After all, only signed code can run there, guaranteeing that the code is exactly as intended by the code maker (that's the definition of trusted computing, right?)
After all, nobody has an XBox that permits unsigned code, except maybe developers.
It's like saying "My audio player will talk directly to the hardware, so I'll make sure that my 5.1 decoding only goes to the hardware. It can't possibly go to a software virtual device that then goes back into a file for someone to share".
Gee, and I used GCC to compile my code, so I know it can be trusted. After all, was it ever bootstrapped from pcc? Nah, couldn't be.
Trusted computing is provably impossible. You can never determine at any one layer if you can trust the layer beneath you to behave properly.
An application can never trust the OS. (compatibility emulation).
The OS can never trust the hardware (emulation).
CPU can never trust the boards
No board can trust that another board isn't sending it fake signals.
Etc. You can always be put into a hostile environment.
Hmm... this started off going for funny, but now seems serious.
No problem, that was posted Feb 28th. So, even if they shut you down til next month, well, it's March as I post this.
:-).
'Sides, I wouldn't even have loaded your site if your post hadn't let me know that there was a link back in the article
Michael
Prior to the world wide web, back on NeXTStep machines, in 1993? Oh yea, I remember that.
... ... ...
:-)
:-). Try surfing with graphics turned off. Gee, why am I getting a lot more ads now than before? Why did it load that graphic? Oh -- it's a flash, not a graphic. Hmm... no option anywhere in IE to turn them off, is there?
Lets see, lots and LOTS of data published in FTP, with an indexing service (the previous "google") able to tell you what was available on all the public FTP servers, and even an NFS to FTP gateway, so that all the ftp sites out there looked like a regular part of the file system. (That one worked real well on high speed connections, but not well on dialup. Primarily because it ignored any NFS request that wasn't available in the cache, but waiting until a re-request came later and the fetching was done.)
And, other people have mentioned Usenet. Or do you remember that prior to 1993 Usenet actually had enough signal to make it viable? (I've long since stopped using usenet because of the sheer size and low signal).
And Gopher, and WAIS, and
No, the real truth is this:
Prior to 1993, there was X-Trek, and the non-X-Trek.
"We must absorb the non, until it cesses to be non" (or something like that.)
Now, with a slight alteration to what you said,
The World-Wide-Web is about making profit not some crazy hippie idea of freely exchanging knowledge for the benefit of mankind.
Want proof? Look at the ads here on slashdot
> essentially rewarding people who wanted to find ways to exploit the awkardness in the design to take on other players.
:-).
Hmm... The "Buddy" spell, no attacks allowed? Ahh, but destroy wall isn't attacking you.
(Sorry, I just couldn't resist. How many of you know this reference?)
Two problems.
:-). Or start another one.
...
First, you won't solve the problems by making the _CORPORATIONS_ criminally responsible.
Second, it's not the customers demanding "low quality now" that causes the problems.
The brief answers: You don't want to punish shareholders and employees for actions of the board and vice presidents. (Think AA after Enron). Customers cannot demand a certain quality unless that quality is actually out there -- right now it's all too often a choice of several different things that are bad in different ways.
The long answer:
Yes, I agree that companies should be liable for bad products. I agree that product liability laws should extend to software. I think it's stupid to allow companies to say "This product is not guaranteed to work, and may have no actual use for what we market it for" (ok, that's not the actual words, but every software product has that disclaimer, even OS's.).
But you want to target the right people. If you go after the company, and put the company out of business, what happens? Answer: Shareholders lose. Bond holders get pennies. Banks collect what they are owed at the bankrupcy auctions. Employees are out of work. The number of companies in the business go down, so there is less competition => consumers lose.
But have you hurt the people responsible for this disaster? Not really. They might even get a job with another company
Choosing good quality products? Well,
I'll try to explain by two examples how hard this is. First, let look at a VCR.
What do you want a VCR to do when it reaches the end of the tape?
Choices:
1. Do nothing.
2. Rewind.
3. Eject
4. Rewind and eject
5. Rewind and replay.
Now, I don't know about you, but to me, the answer could be different depending on:
1. Viewing a regular tape.
2. Viewing a non-erasable (rental?) tape.
3. Recording in regular mode
4. Recording in timer mode
5. Whether "memory" mode is turned on.
Now, when was the last time you even checked this behavior before buying a VCR? Or met a sales person that even considered, let alone knew, the answer to this?
Consider telephone answering machines. Consider digital ones.
Can I play back new messages, stop after some of the new, and not have all of them marked "Old"?
Can I play back a bunch of new messages, skip some, and leave those skipped ones as new?
If two people share the machine, can I play the "tape", and say "leave that as new, mark that as old" to explicitly leave messages for my roommate?
Again, when was this ever even listed on that 4x6 card at Fry's?
Demand quality? Sure. Tell me who to talk to. Or do you actually expect me to track down ALL of the companies that make answering machines, and explain to all of them what I want, when most of them have no way for me to talk to a person?
[ Side note: In economic theory, this is why middlemen exist -- if you have a lot of sellers, and a lot of buyers, there are much fewer transactions if every maker sells to a middleman, and every purchaser buys from a middleman. It would be much easier to have "product enhancements" if everyone that wanted to say "improve this" only had to go to one place, and product makers only had to check one "What do customers want" place. Now tell me how to make money running such a web site and it'll spring up.]
Haven't you noticed that most companies now only have automated help lines, with no ability to speak to a real person, unless (maybe) you've purchased a new product within 30 or 90 days and need help getting it set up. But product improvement?
I finally got so upset with bugs in the Dish Network receivers that I started calling, and reporting them. I also reported some "I'd like to see this" feature requests. I reported two such things, had two different people at the other end say "Hey, that sounds really good". But when I tried to report a 3rd, I was told "We do
If you are able to slashdot from work, are you really overworked?
[ I'd like to just be worked and paid. ]
Quote:
Display Postscript and NeXT, and apparently the PDF support in OS/X, are all terrible in comparison. All of them rely on you creating a window and "context" with completely unrelated calls, before you can finally draw in Postscript. Making the window the same shape as an object you draw is incredibly difficult or impossible. Passing the same color you are drawing to some other interface is difficult or impossible. Synchronizing your drawing with changes to the windows is difficult or impossible. Getting a mouse event and figuring out what part of your graphic was clicked is often an enormously complex and obtuse API.
Reply:
I played around with this back in the 3.2 days. Creating a windows WAS a postscript call. A simple one. It would give you a block of whitespace on the monitor that you could draw in. It was NOT the same as an appkit window, which was additional stuff based on this postscript window.
If you want to create an Appkit-level window in DPS, then you have to move all of appkit into DPS.
While I'm not an X programmer, it looks (based on using X, and a lot of different X apps and window managers) like X is more based on "Give the programmer the ability to tweak arbitrary pixels on the screen, let them limit their drawing into this clipping region (window) if they want, and give them (and only them) control over the content of this clipping region". All the "window" stuff comes from another program called the window manager. Well, that low level, "you control the contents of this clipping region" is exactly what DPS gives you.
Similarly, display contexts are manipulated with direct postscript. Oh sure, there's nice ObjC messages to make it fit the language. But it's in DPS at the low level.
Synchronizing your actions with the display?
windowdeviceround
if I remember correctly does that, as a zero argument postscript command.
Mouse and/or keyboard events?
There was a global postscript file that affected the system. All events went through this file before going to apps. With a relatively simple mod (and this was where I started learning postscript, learning how this worked) you could modify all the system level behaviors, such as keyboard commands for cycling windows, activating windows, etc. Yes, you could read the keyboard, and send commands anywhere; you could check what the mouse was doing, etc. The only problem with this? It triggered the nasty swap space bug in NeXTStep 3 (bug was still there in OpenStep 4, btw -- massive growth and non-shrinking of swapfile).
Window the same shape as your drawing? Best I could recommand is to draw a 100% transparent clear into your box first. Then draw normally. 100% clear will pass the mouse through, and will act like it's not there.
Where is the mouse compared to where my window is? I think you could ask the window manager, in postscript, to tell you either the 4 corners, or 1 corner and two lengths; you get the mouse position from the event (in postscript), and do some adds and subtracts.
Scaled windows? Not sure. Rotated windows? No. Disconnected parts? 100% transparent plus clipping path.
No, you just make a copyright of how the printing press works. Sure, patents expire, but corporate copyrights just keep going (label A:) and going and going and going and going and going (goto A)
Michael
Hang on... The day _WILL_ come when there are only a few ISPs?
Today we have:
1. AOL.
Distant #2: Earthlink.
Way way off in the distance, some tiny startups:
juno, people pc, walmart.
Other than that? Regonal tiny companies, like phone companies or city-wide ISPs.
If you want a nation wide ISP that you can use while traveling with your laptop, you've got 5. And only 2 of them are any size at all.
Keep in mind: AOL spent 10-12 years purchasing market share at a major loss before finally starting to make a profit. AOL's investors and VC's spent their money buying a virtual dominance if not outright monopoly. Even today, what does Microsoft do to inflict its new email fiat on everyone? That's right. Gets AOL to sign up, and ignores earthlink, walmark, phone companies, etc. They don't matter. They're too small.
You're kidding, right?
:-).
According to the supreme court, if I'm accused of 30 counts of drug dealing and sexual molestation, and 1 count of speeding, then when I'm convicted of speeding, I can be sentenced as though all 31 counts were guilty, because "not guilty is not the same as innocent".
According to the supreme court, the Menendez brothers (remember them?) were well served in the requirements of justice by having seperate juries, a prosecuting attourney that argued before both juries that their opponent was guilty of this crime that only one of the two could have done, getting guilty convictions against both of them, and having both serve life sentences. IF you didn't understand that: Here you have 2 nasty people. Both have committed crimes. But for this one crime, at most 1 of these two can be guilty of this crime, yet both are found guilty. You are guaranteed to have one innocent person in these two, yet both the guilty and the innocent are sent to life imprisonment, and the supreme court says "OK, both were found guilty by the due process of the law".
Jury nullification. The jury has the right and power to return "not guilty" even when the law is technically violated, when the act in question is done. That's been upheld by both appellate and supreme courts. The supreme court has said that it's not a mistrial if the jury is NOT informed of this right. Yet trial court judges will insist to juries that they must return guilty if the law is technically violated, and no appallate nor supreme court has said "trial courts cannot instruct the jury that way".
And don't get me started on either jurisdiction (carefully read the first sentence of the 14th amendment, and ask yourself why the federal government and federal courts gets involved in intRA-state matters) or tax law issues
Keep in mind: The preamble of the constitution:
"We the people, of the united states of america, in order to do A, B, C, and D, do establish this constitution". That's a statement of intent and purpose if I ever saw one.
So what is the intent of the constitution?
1. Form a more perfect union.
2. Establish justice.
3. Ensure domestic tranquility.
4. Provide for the common defense.
5. Promote the general welfare
6. Secure the blessings of liberty for ourselves and future generations.
What else does the constitution say about the power of the courts? That it is in one supreme court, and any lesser courts established by law.
*** The constitution requires that our courts are courts of JUSTICE (#2). Any judge that says "This is a court of law, not a court of justice" is in violation of the constitution, yet the supreme court historically upholds that belief.
I haven't looked at this since Java 1.0 or 1.1, so this might be out of date.
Java "binaries" are restricted in two ways:
#1: It must fit the machine's CPU -- the byte ocde sequence which is legal for the JVM.
#2: It must fit the restirctions imposed by the class loader.
#2 is critical. Years ago, there was an Objective C compiler that compiled to Java byte code. The implementation managed to do everything except (if I remember correctly) poseAs, and I'm not sure about categories (I believe that did work, don't quote me).
The problem? It was legal JVM code, but not legal according to the class loader.
C++ and Java are NOT closely related. Years ago I worked for a company that had a data architechture that started from (and still needed to support) 8 bit device drivers on IO boards; at the other end, they ported a large C code system into Java, and made it instantiable -- specifically to permit multiple copies of this large system to run at the same time. The entire high end of the system needed to suport all of C++, Java, and Objective C (this was running on Mac Os X).
The design results? C++ was the limiter. Both Objective C and Java were effectively identical -- either of them could do what the other could do, they could talk to each other (apple's native bridge), etc. But C++ was highly limited. The inheritance system of C++ is vastly weaker and more limited, the bridging for C++ had to be completely manual, etc.
I'd be suprised if you could get a C++ compiler to compile to a Java backend. It seems to me that the basic assumptions of the languages is just too different.
OK, why not open another mp3.com?
:-).
:-).
Sure, you can't keep the URL (unless that's up for auction), but the basic product -- an alternative music distribution channel for people that want more money for their songs than they could get from the big 5 record companies -- is still viable.
Mp3.com didn't die because of a bad business model. It died because of a bad court decision from a bad law, and bad advice from lawyers. Don't forget, after VU took over mp3.com, the new mp3.com sued that law firm for bad advice (now that they had the money to sue the lawyers
From my personal experience, the UI of mp3.com went downhill after VU took it over. It went from an easy to use site to an impossible to use site, so I stopped using it. So did a lot of people. Then, it was dropped and put out of business.
So lets reopen it. The guys that built the original mp3.com know how to do it, they could be the advisors even if they can't (because of lawsuit) run it themselves.
This time around, people know what they are doing. After buying the servers, you've got much of the system code. You should be able to restart it for a 1/10 of the cost this time around, especially with the lower programmer costs
Hey, I'm unemployed, and willing to work with this company!
Michael
Ok, why not open another mp3.com?
Can you gamble on it?
Lightweight question:
Will Ziggo win? Or will someone with a robot specifically designed to take it out succede? (Answer: It was taken out by a specialty bot in its first TV appearance).
Heavyweights:
Will "Son of Whyachi" make it to the end? Will it defeat BioHazard?
Will Torro manage the full season without a technical problem that kills its flipper? Even if it does, can it win the shoving match?
Will Taz-Bot manage to outdrive, and out-manuver BioHazard this year? Or will it be a perrenial #2-#4 bot?
Will the 6 wheeled pusher bot outshove the others? If it can outdrive and get to the side of the others, it has more pushing power than anyone else (sorry, I don't remember the name).
Or will some unknown new bot make it all the way to the quarter-finals, semi-finals, or even take the victory?
Who will win the heavyweight division? Do you know? Or do you want to bet?
Yes, you can gamble on BattleBots. I would never have guessed that Son of Whyachi would have abandoned that excellent walker base, which gave it both extra weight for the weapons and base (more stable, less flippable), and extra clearance/distance between the weapons and the base (better able to inflict damage without taking it), for a "standard" wheeled base and a faster weapon spin. My prediction: It will take too much internal damage from feedback when it hits, and not make it to the final just from wear and tear. Care to bet against me?
Is BattleBots too predictable? I don't know. For you football fans, do all the lineups fit into 1 of three basic categories (4 blockers and 3 receivers on front, QB and ball snapper, 2 ends as the basic lineup, a couple of variations on that). Do people trying new things, like a shotgun (QB starts at a distance, so that the time spent running backwards is saved) abandon those because this is the "tried and true" system that works?
Do all the teams on baseball agree on where the 9 outfielders should be placed (3 distant, 4 along the baseline, pitcher and catcher)?
Is the question of which teams will be in the top 4 pretty well known before the season begins?
BattleBots is MORE changable than any established sport, just because anyone can enter.
Yes, a few (1 or 2) bots have been sold on Ebay. Most are made by the user. Some are totaled. Some are rebuildable/repairable.
Michael
Can you even imagine baseball or football without prime time coverage?
I'm sorry, but the average human mind CANNOT decrypt pig-latin in real time.
Even if you restrict it to the average english speaking person and english-pig-latin, it still doesn't hold.
Take it from someone that never learned pig latin as a child. I'm not joking.
You can learn any language as a child, and speak it in real time. Does that mean that translating something into a foreign language isn't encryption? There was one war where messages were encrypted by translating into an indian language (I think Cherokee, not certain), and it was the only code not broken in that war. (Vietnam? Again, not certain).
As someone who learned pig-latin only about 5 years ago, I can assure you that "real-time" for english is 600 words per minute reading, while enlgish-pig-latin is about 20-60 words per minute. That's slower than I speak, by a LONG shot.
Don't assume that just becasue you learned something in childhood that it's a trivial real time decryption. That just means that you're bi-lingual, nothing more.
Keybounce
Ok, how do I get invovled in this now?
I was one of the people that received an email saying "You can get red hat IPO through e-trade". I called e-trade, got a verbal quiz, and then told "You don't qualify". Never got around to opening an account because I was pre-disqualified.
Now I find that there are class action suits, and NASD arbitration reviews. How do I get involved in this?
Michael.
No, no no.
...
The supreme court has ruled that paper gains on stocks is NOT taxable income. The basis for this was the realization that a person who had a controlling interest in a company might find that in order to pay the "paper gain" taxes, they had to sell their controlling interest in the company. This, according to the S.C., was unacceptable. Therefore, paper gains are not taxable, only the actual gain/loss when the actual sale occurs.
What you are proposing here actually is just as bad: I'm given stock options. All of a sudden, I owe money that I don't have for paper gains. I do, however, have the option of going into debt to the company (call it the "company loan store"), and at a later time, when I sell, if they are still valuable, paying them off, with interest. Gee, I didn't make enough to pay off my loan on my stock options? Oh well, gotta work more next month, and next month, and
Companies paying in company script, redeemable only at company stores, forcing people into debt is illegal in this country (it is considered slavery). Your proposal brings it back under a different form.
[the parent message to this message is insightful, and worth reading]
This is not rediculous.
:-)
/etc/resolv.conf file when the ISP provides DNS information, so I can read it from that.
To use this, you do not need to have a name server; you need to have access to someone else's name server. And, you need a central "modified" server somewhere that you can talk to.
I haven't seen a system yet that didn't have some way to look up names, even if it could not actually connect to them.
So, to use this you need:
1. A linux system.
2. A toll free limited account (quickbooks, MSN signup, your bank's "bank online" feature, etc).
3. The remote DNS that you dial into configured to check the root DNS servers (normal
4. A special server set up somewhere (come on guys, post your server names for us to use).
Then, you start up your client by giving the number address of your limited DNS, and the name of the remote tunnelling DNS.
From the README:
::
Then start the server on one end:
./nstxd tun.yomama.com
and the client on the other end:
./nstxcd tun.yomama.com 125.23.53.12
125.23.53.12 has to be a DNS-server which can be reached by the client-side. The server *must* run on a server where an NS-record for tun.yomama.com points to. So if the server has the IP 1.2.3.4 there must exist an entry in the zonefile of yomama.com: tun IN NS 1.2.3.4
::
A server runs somewhere on the internet. I, the client, need the name of that server, and a DNS operated by my limited ISP whose number I know or can find out. PPPD supports modifying the