Of COURSE is must be unsafe if its native code. Unless you mean native code on a new architecture that doesn't support pointers to arbitrary memory.
Think about it -- unless the code can be fully tested, how do you know that the COMPILER generated the correct code? And, the compiler is just another program -- a lemma of the stored program computing we have.
So, the only way to guarantee a native code program on current machines is... equivalent to solving the halting problem. Because it has to be COMPLETELY simulated.
Which brings us back to the architecture. If the architecture had either no pointers, or constrained typed memory, we could do it.
Since the x86 has deprecated segmentation (which could be used for this), we are stuck with an "arbitrary memory pointer" model.
The JVM and other intermediate machines attempt to use another "non pointer" model, and then convert the code into native code safely (if you trust the intermediate machine coding). This is doable without being stuck trying to solve the unsolvable.
Smalltalk was an experiment -- can everything be done with message passing?
The answer is "yes". (but, of course, this may not be desirable).
Conditionals ("if") fall out of sending messages to boolean objects.
boolean is sent message "if" with argument "a block of code".
IF the boolean is the "true" instance, return the value of evaluating the code block. IF the boolean is the "false" instance, return NULL.
If tail recursion is added, WHILE becomes just as easy.
Which means that the syntax degenerates to "send message to object with arguments", and "define a code block".
A pleasant result -- there is no special syntax to IF/WHILE or anything else.
The Smalltalk compiler may, of course, optimize this stuff (and usually does).
But, the NICE effect of this is that "boolean" and control structures are NOT CONCEPTUALLY PRIMITIVE -- they are defined in terms of "OOP" or message passing, giving the possibility of defining your own control structures. If the C do/while is not available in a Smalltalk implementation, it is easy to add, and (sharing a trait with LISP) the implementation looks the same as the "default" control structures.
doWhile method for "false" - execute code block and return.
doWhile method for "true" -- execute code block, evaluate condition code block, if "true", recurse, else return (and, please note that the conditional is a message sent to the boolean result of the conditional code block).
Of course, you got me on "boolean instance". Since code can only be put into a class, the "true" and "false" instances are actually subclasses of a boolean class.
Of course LISP is yet simpler... But I agree, Python gives almost the ideal balance for most programmers (personally, I like the LISP approach though -- I think Gambit-C SCHEME with "six" syntax is better than Python, but that's just me).
As you can see, China is bigger with manufacturing, US is bigger with services; China doesn't have the debt, or the excesses in the stock market. China has cash, and gold reserves. China doesn't have external debt. China exports more than it imports.
About the ONLY metric that China falls behind in is GDP -- which, from the other indicators is simply being propped up by... China.
(Oh, and the US has more energy reserves).
Have fun with the easy graphics. The GP was right -- China pretty much owns the US.
The "internet" started with the PSTN. Using modems, and the "internet" inherited the reliability of the the PSTN. Of course, having multiple paths is more reliable, and, unlike the PSTN, the "internet" is more tolerant of less reliable connections.
As more data needed to be shipped, things like fiber were implemented. Packet switched, just like the "internet". Now, the PSTN uses that same infrastructure. Instead of the "internet" being supported by the PSTN, the PSTN is supported by the "internet".
Which means that the "best effort paradigm of the Internet" DOES work for all applications, including telephone type services.
Now, local telephone loops supply power separately from the grid. If the grid doesn't function, the telephone may still be powered. An accident of history.
PS. There was no guaranteed quality of service from the phone company either. In fact, it was a running joke while it was still a monopoly.
Lily Tomlin: "We don't care; we don't have to. We're the phone company."
Wrong on a few points. Copyright doesn't mean I can't make money off a work that I don't have Copyright for. In many cases (not all, there are explicit laws preventing this universally) I can rent the Copyrighted material out.
For example, I could create a "pay-for" library. No copying takes place, and I make money "showing books to the public".
I can remove the signature from a painting, and then sell it. I can even claim I painted it; that would be plagiarism, but not a copyright violation.
What about removal of the big kernel lock? What about plug-in resource schedulers? What about fast ip locking? What about kernel video mode switching? What about systemtap? What about cgroups? And much more...
When taken in combination, the growth of the Linux kernel since 2003 definitely warrants a major jump.
The issue is whether it should be 2.8 or 3.0. I would side with 3.0.
Because Linux is now ready for serious MP, both on a local and a cluster level. And these features are not "backwards portable".
"And it's not on Linux? Only when you limit yourself to the "distro-specific app store, a.k.a. apt" do you get system-wide updates. Just like I do if I limit myself to MS software. I choose not to."
I just pointed out that those application vendors offer repositories. These repositories are "outside" of the Linux application sphere. Once added (and keys confirmed) the various software components are updated by the same updater that takes care of the OS.
Which means Linux users don't have to limit themselves either. And, the benefits of having controlled repositories can extend to third party software vendors.
What I really don't understand is why Microsoft Windows doesn't offer something comparable. Actually, Microsoft does, in the hardware space. Once a driver is qualified, Microsoft will distribute that. But, there appears to be no mechanism for extending the Microsoft updater to include third parties.
Maybe I am wrong (I am not a Windows user, really -- I just use Virtualbox/Microsoft Windows 7/Office/Project for compatibility reasons), and I really don't use (much) non-Microsoft Windows software (Adobe and putty). I do know that Adobe and putty don't use the Microsoft Updater, but I don't really understand why.
Data recovery "bit level"? I just use the standard POSIX command line tools on the primary disk (done on Linux, although I could have done this on another of our systems):
Really, this stuff works on pretty much all OSs (AIX, Solaris, Linux, HP/UX) out there. Not so sure about Windows, but I would HOPE it works (Microsoft does claim POSIX, don't they?).
With additional knowledge about the disk layout, a systems engineer can recover more information. Yes, I took a course on UFS, and Solaris Internals, and I can recover data from Solaris, Linux, BSD, AIX, and HP/UX. What is important here is that a client can call, and I can do this -- I don't need a "custom toolkit" for it either -- just OS level commands. But this is predicated on knowledge of the file system. I presume that this service can be supplied in the Windows world?
I put Linux on my laptop. Fedora 14. It resized the Windows partition, and installed as "dual boot".
No need to reactivate on starting Windows. Next, I installed VirtualBox, created a virtual machine that would use the Windows partition directly. That worked, but I had to put a call in to IT to have Windows reactivated.
Now, switching from hardware to virtual (P2V) is like changing the mainboard.
So, I don't understand your experience. Maybe you are using a "bootleg" Windows? I sure don't. Mine is 100% legitimate. Because the BSA can audit us any time.
On to repositories. Fedora repositories include:
adobe dropbox google google-chrome skype virtualbox
(and, I imagine, more -- these are the "third party" application repositories I use).
As to "reverse engineering" NTFS -- as long as you use Windows, it's not needed. Sure, why the hell would you? It's not like you would need reliable data recovery tools, or ever need to repair a filesystem for someone. After all, everyone does backups. (sarcasm ends)
B-Trees are certainly not the fastest way to store and retrieve data. You may want to investigate those strange things called "hashes". Fortunately for you, most database systems know how.
As well, "assembler" isn't the fastest coding system. It is so very slow and tedious to use that algorithm tends to get overlooked. Back to your "B-Tree" again -- have you tried coding a B-Tree in assembler?
To give a hard example -- back in the days of OS/2, the filesystem was coded in assembler. I imagine that it was considered the right way. After being recoded in C, the thing ran faster -- mostly because the developers could concentrate on better ways of doing things, rather than the drudge work of getting the assembler correct.
Another example -- replacing a "C" coded inner loop in a console game with a scheme interpreted version, that ended up being faster because it could fit into cache.
Sometimes upstream is simply better, even for performance.
Database systems are in the same category. It is rather difficult to optimize specifically, and, by the time you did this, typically the data store would be "out of touch" with the actual business requirements. So, most of us use a good database, with a query optimizer, and let it take care of that part of the problem. Yes, it may need some tweaking, but those tweaks will tend to be independent of the business logic. This can give rise to specialists who can be called on if needed.
Using your model, the "assembler tuned B-Tree" custom coded, there would be no paladins available to assist in the tuning. Indeed, moving to another architecture would be difficult (or impossible). It would be ridiculously expensive, and would likely never be upgraded. Hey, CICS and IMS are still in use, aren't they? And THEY are more portable than your suggestion. (CICS and IMS just turned 40).
PS. There are CICS and IMS paladin around. There wouldn't be any for your "solution". Also, IMS can use direct storage links, and, because data in IMS is hierarchical, even IMS can outperform "B-Tree".
Feedback -- sure, why not. But, it has to be in the form of questions.
1 - does your powerpivot appliance work with IBM ISAS appliances? 2 - does your powerpivot appliance work with IBM BCU? 3 - how does your appliance interoperate with Ab Initio? McSource? COGNOS? (8.4 and 10). If a customer has, say 6000 or more ETL jobs running with Ab Initio, and wants to start adding powerpivot functionality, is your appliance represented by a new GUI box to simply drag-and-drop in that environment? 4 - your appliance is configured to provide "backup" -- now, I don't have time (right away) to dive into your offering (week after next is the earliest, I have to present a Data Warehouse/ETL BI datacenter design this week). How do you intend on competing with TSM? 5 - please provide performance figures I can use (simply 24 core/96 GB doesn't tell me much). SPECint or nPerfs to allow me to compare against IBM P5 or P7; I don't really care about comparing against Oracle in this space, but T3 may be useful as well.
Thanks in advance. I presume that you have the answers -- Give a reply here, or simply email me at fred (dot) weigel (at) treklogic (dot) com
But sharing the information here may increase your exposure. Honestly, excel is used to deliver information views for BI, but I have never considered Microsoft as being a "core element" in a DW/ETL environment. And, given licensing fees for Ab Initio and the IBM appliances (a BCU is just south of a million), there certainly is room for competition.
This solution bundles the graphics adaptor with the sending of screens.
Yes, that's wrong. But its wrong because it sends too much data. The WORST amount of data is the screen itself. However, the graphics card itself takes information and generates that screen. For example, texture mapping. Why send all that texture data? Instead, send it once!
The idea would be to send higher level graphics commands via wireless, and then generate the actual screens at the device. An external, wireless graphics card. Now, this model fits the older "obsolete" OpenGL better than the newer OpenGL, which may be its downside.
But, the model should be similar to the "X" protocol (on which OpenGL rides). Commands at this (transport) level to be sent to the device. The software vendor optimizes to this level, and the card vendor implements, either using local transport, network transport or wireless transport.
This fits well into the Unix GUI approach -- maybe someone wants to carry this further into a Windows(tm) DX10/11 analysis?
I thought I had covered your points in my submission. I am going to quote (ellided) myself:
"In the case of my neighbor, that isn't a problem. But my best friend has his own network; but is located across the country. Data exchange with him would be impaired without my ISP (and, my use of/. would be curtailed)...In previous years...I shared this data with my distant friends via transport of CD, or hard disk drive...So, the mail and airplane system was part of my "Internet". It can be again, if those physical pipes are regulated to the point where the data will be inspected."
Note that I "mesh network" with my neighbor; does that make us ISPs?
How about an implementation that uses (say) 4 bytes for an object reference. But directly encodes small integers and characters up to +/- 2 ** 30, if the upper bit is zeroed (31 bit integer). If the upper bit is one, the lower bits can represent an index into a pointer table, which can, in turn be a 64 bit pointer.
This would allow efficient representation of up to 2 ** 64 bytes of object space, and a unified efficient representation of self-representing integers and characters, and allow up to 2 ** 30 objects.
In the memory space used by the OOP programming system (running 64 bit), map the entire first 4GB to "not accessible".
And, the object space can be completely separate, and not directly reachable or corruptible AND an "object reference" would not be usable as a pointer to subvert the type system.
Just saying... I would prefer that sort of design. As a nice bonus, memory could be completely reorganized without invalidating any object references.
And, you still have the property that only 4 bytes needs to be copied instead of an entire object. Note that Windows uses this sort of design -- the object reference is called a HANDLE.
Do you imagine that you have the right to copy a song from a CD to an iPod? In some jurisdictions, you do; in others, you don't. Indeed, in MAI v. Peak, the Ninth Circuit Court found that copying Software into RAM infringed Copyright. Which was not the intent of the law (US); so changes were made allowing for this, and archival copies. If it were not for those changes (remember they are quite recent, MAI v. Peak was decided in 1993), it would be a Copyright violation to transfer a CD to an iPod in the US as well. But I don't know the law in France (and my French is not good enough to do French legal research).
I think that, perhaps, you are confused by what the "Internet" is.
The Internet is the network of networks. When I connect my network, which, for the record, is under my sole control, and not under any other regulation, to another network, we have a piece of the "Internet".
The Internet is facilitated, but not defined, by the "backbone" -- fiber optic connections (currently). The Internet does not mandate the form of these backbone interconnects. They exist because, simply, there is a need or desire to send that much data between the interconnected systems and networks.
So, what are you proposing should be regulated?
Nothing in the description of the "Internet" mentioned content, or physical interconnection requirements, or physical systems.
Read the article -- the point is brought home when the author mentions "Bob" and "minitel" as being REQUIRED to connect... not to the "Internet" but to regulated information providers.
As an example - my neighbor has a network of around 5 systems, and I have a comparable network. We connect these networks using two different ISPs (a cable provider and a DSL provider). We also occasionally directly connect with 802.11g wireless. The connections are made to exchange data (photos, videos, Linux distributions, it really is none of anyones business). What regulations would you want to put into place, and where?
Because 802.11g, ethernet, etc. are just convenient standards that permit us interoperability. We could come up with our own local protocol if we had to. The ISP provides a higher degree of interoperability (an IPv4 address that allows my network to be reached from other ISPs that co-operate, packet services that allow data to be exchanged, and access to DNS services to provide a convenient naming service). There really isn't any service my ISP offers that I couldn't easily replace, except for access to the high-speed backbone.
In the case of my neighbor, that isn't a problem. But my best friend has his own network; but is located across the country. Data exchange with him would be impaired without my ISP (and, my use of/. would be curtailed).
I thought the article was actually very well thought out. The point is that the "Internet", being a simple connection between systems and networks, only exists because it is beneficial to the connecting parties, and the method of connection is irrelevant.
Now, as soon as enough systems and networks participate, there is a "network effect", and some of the "Internet" may degenerate into a pure producer/consumer relationship. I imagine that this is what you want to regulate. Be careful though, because the end game of regulation is simply that some parties may decide to pack up their dolls and leave. We are not TIED to the infrastructure. We have always had a bit of balkanization in the Internet. Certainly, not all of my content is publicly available. Some is simply not directly addressable, or findable via Google (darknet, an example of this would be my music collection, family photos and videos, which I share to friends but no one else). Some is SSL only, and password controlled (my TV watching schedule, my calendar. And some is only locally available (banking statements, financial records).
In previous years (when high-speed meant 1Mbps down/128kbps up), I shared this data with my distant friends via transport of CD, or hard disk drive. 5 Gbytes of music was a lot to jam down the limited communication pipe I had. So, the mail and airplane system was part of my "Internet". It can be again, if those physical pipes are regulated to the point where the data will be inspected.
Why would an object reference be a pointer? A more useful construct would be an index to a pointer; then the object can be moved and all pointers would still be valid. Or, possibly a hash.
Anyway, I would definitely shy away from making an object reference even APPEAR as a valid pointer. The benefit is that "cracking" a system, and supplying an object to an OS level routine just wouldn't work. (eg. Put all object references in one 32 bit space, and map that as read-only). Characters and smallints can live in object reference space "as-is", everything else would need the double indirection. Yes, performance may suffer a bit, but wouldn't the system be more robust?
And, as a side-benefit, any low-level code introduced would be forced to use the correct access mechanisms, making it more likely to be portable.
Basically, the idea is that bits of the key leak. And how is this accomplished?
For example - if a key bit is 0, you take one code path, if 1, another. Make the two paths different lengths. It may be possible to affect packet timing. Or... A function may end with "x - y" and then return "z". No leak? Not so clear, the carry/borrow may be leaking information to the caller (on x86 style hardware).
Anyway, it probably isn't a "back door", just some means of determining enough key bits to make brute force practical is enough. And this sort of thing can be subtle. It can even be based on the machine code generated for certain sequences by a particular compiler (the "x-y; return z" sequence above, for example).
In 1977/78, I worked on a system called PEAC. It was developed for CTW (Childrens Television Workshop), and used for Sesame Street production (among other programs).
A number of clicker boxes were plugged into a charging/syncing station. All of the clickers could be synchronized by a single button press. They were then disseminated into an audience. Both demographic questions (are you male/female, what is your age, etc.) were collected, and then polling data for the program was collected (like/dislike on a scale of 1 to 5, for example).
Afterwards, the clickers were put back into the syncing station, and the data was transferred to an Apple ][. If the material was sourced from a video tape, a secondary audio channel was written to the tape, with time marks.
The time marks were fed into the Apple ][ audio controller, and the floppy based demographic and polling data was then set up to produce interactive cross-tabulations, color swash (called "mondrian" analysis), and other investigative approaches. The bulk of the analysis software was written in Apple ][ Integer BASIC, along with machine code routines for actually performing the cross tabulation.
The Apple ][ was chosen because it supported "color" graphics, audio input, and floppy discs.
The system was delivered in 1977/78.
I don't know when it was decommissioned. But, I imagine it wasn't unique (well, the time synchronization made the "user interface" simple -- just fast forward, rewind and play on the video tape deck! The audio channel was composed of many short independent "files", each recording a time-stamp; and, I have never seen the "mondrian" style display for analysis again, but maybe it was a bust). But the patent is deliberatly vague about that end of things, anyway.
Now, honestly, we would have liked to use RF, infrared, or "mobile device" clickers. Instead, we had to develop our own stand-alone clicker device (and, honestly, I don't know if you can run 30 IR clickers in a classroom at the same time!). RF would have been an option -- but we put storage into each clicker, and synchronized the lot in the charging station. But, the need for clicker synchronization was a weak point, and we recognized that (back then). We could have used low frequency FM (47Mhz required a license, and cell and "wifi" didn't exist). So, low power 88Mhz was an option. Still, not as reliable as the hard-sync approach.
Another problem (back then) was that available AFFORDABLE computers would have been hard-pressed to reliably support 30 concurrent clickers. We would also have needed to add a processing layer in to support that. With a sync box, each clicker could be read out and converted to floppy in turn.
We did think of it. It was documented. Question is, does the paper from back then even exist? We never considered it patentable material.
Insanity: doing the same thing over and over again and expecting different results. Albert Einstein
I am (finally) forced to use Microsoft Office(tm) -- the one without the menu bar (2007). With Microsoft Windows(tm) (7? not sure). I have to have it plugged into a network in order to use "domain login".
It runs anti-virus software, some kind of remote access software, and Office (because heavily formatted documents from others don't work well with OpenOffice.org). I don't have "administrative access".
Since it's plugged in, I decided to portscan it. 10 open ports (AFAIR). Compare and contrast to 1 open port when I open my (considered non-secure) laptop (22, for ssh).
Yes, SSH has had two exploitable vulnerabilities in the past 15 years. (well... 1, unless you use debian/ubuntu in which case 2). But, client computers really don't need any other incoming ports. I guess it's a good thing my Windows laptop doesn't have port 22 open!
As to JScript? I have been using NoScript. Blanket denial of JScript unless I specifically enable it. We know that pattern-matching heuristics don't work (even though I am forced to put up with it on my MS Office laptop).
Still, I guess that the universal access to computers that has been enabled by Microsoft Windows and cheap Intel hardware has been a net positive to society.
The solution would be to produce two operating environments -- a "starter" and a "normal". We already have this, in the form of Windows vs. Everything Else. But, due to the commercial nature of Windows, there is limited compatibility across the tiers.
What I find fascinating is your distinction between "toy" and "real" functionality. Canvas, to be useful to me, must remain simple. I only use it every 6 months or so, and I don't (typically) have time to read up on it.
I "hack" up a quick & dirty thing, run it a few times and then throw it away. Really, the original question was about "personal" programming.
Typically, I try to generate the graphic (if static) with a spreadsheet first; or gnuplot. If interactive, I try to feed gnuplot instructions, or use OLE and DDE (if I remember how, normally, my attempts in that space don't work out quickly enough). If that doesn't do it, or I need mousey input, I turn to tcl/tk.
Qt is way overkill. The only way I could do that is by building a bunch of templates, and cargo-culting it from there.
If I want Qt, I'll throw the prototype over the fence, and get some folks in development to redo it.
Imagine an artist creates a piece of performance art. A photocopier, with a pushbutton that is simply marked "for content, press me". The input to the photocopier is sealed by being screwed shut, and a single page is loaded, ready to be copied.
You press the button, and a copy is made. Out roles a sheet that says "This work is copyrighted, and All copying rights are reserved". Behind the curtain is a government official, hiding.
You press the button, and find yourself arrested, and have to pay a $750 fine.
(If I were building this, I would simply write that on the paper)
What is interesting is the concept of "automatic copyright" introduced in 1976. If copyright had to be registered, a person MAY be expected (it would be difficult, but may be possible) to search copyright status and allowed rights prior to taking an action.
With the "automatic copyright", the only legally correct action is NOT TO PUSH THE BUTTON. But, what if the button said "For content, push me. I give you permission to push me." I argue that a transitive copyright has now been granted, especially since there is no other way to determine copyright status.
Downloading is NOT copyright infringement. Everything is "copyrighted" on the web. Even the google home page!
So, if downloading copyrighted material WERE infringing, we could make it really simple and just close the internet.
As a corollary, there is nothing wrong with searching for material, and including terms like "torrent", or "bittorrent".
Not the slightest hint of illegal. NOT THE SLIGHTEST.
Downloading is... just downloading. Searching is... just searching. Copyright infringement is copying a work without permission of the copyright holder.
Let's bring this back to web pages -- it is perfectly legal to DOWNLOAD that web page, but may not be legal to copy the bits and store them on your hard drive. It is perfectly legal to DOWNLOAD that web page, but may not be legal to print out a copy.
Interestingly, if the page has "Copyright (c) year, some person / All Rights Reserved", or NOTHING on it, a strict interpretation of copyright would lead to the above. Since browsers make a COPY in the ordinary course of operation, and the web site content provider is (should be) aware of this behaviour when preparing content, it can be argued that, at least, that copyright was granted. In other words "All Rights Reserved" makes no sense in the context of the web. This applies further to other downloaded content.
It is perfectly legal to download, as long as further copying is not engaged in (actually, in Canada, you can download music, and then further distribute as well, under certain circumstances).
Slashdot Mirror
Of COURSE is must be unsafe if its native code. Unless you mean native code on a new architecture that doesn't support pointers to arbitrary memory.
Think about it -- unless the code can be fully tested, how do you know that the COMPILER generated the correct code? And, the compiler is just another program -- a lemma of the stored program computing we have.
So, the only way to guarantee a native code program on current machines is... equivalent to solving the halting problem. Because it has to be COMPLETELY simulated.
Which brings us back to the architecture. If the architecture had either no pointers, or constrained typed memory, we could do it.
Since the x86 has deprecated segmentation (which could be used for this), we are stuck with an "arbitrary memory pointer" model.
The JVM and other intermediate machines attempt to use another "non pointer" model, and then convert the code into native code safely (if you trust the intermediate machine coding). This is doable without being stuck trying to solve the unsolvable.
Smalltalk was an experiment -- can everything be done with message passing?
The answer is "yes". (but, of course, this may not be desirable).
Conditionals ("if") fall out of sending messages to boolean objects.
boolean is sent message "if" with argument "a block of code".
IF the boolean is the "true" instance, return the value of evaluating the code block. IF the boolean is the "false" instance, return NULL.
If tail recursion is added, WHILE becomes just as easy.
Which means that the syntax degenerates to "send message to object with arguments", and "define a code block".
A pleasant result -- there is no special syntax to IF/WHILE or anything else.
The Smalltalk compiler may, of course, optimize this stuff (and usually does).
But, the NICE effect of this is that "boolean" and control structures are NOT CONCEPTUALLY PRIMITIVE -- they are defined in terms of "OOP" or message passing, giving the possibility of defining your own control structures. If the C do/while is not available in a Smalltalk implementation, it is easy to add, and (sharing a trait with LISP) the implementation looks the same as the "default" control structures.
doWhile method for "false" - execute code block and return.
doWhile method for "true" -- execute code block, evaluate condition code block, if "true", recurse, else return (and, please note that the conditional is a message sent to the boolean result of the conditional code block).
Of course, you got me on "boolean instance". Since code can only be put into a class, the "true" and "false" instances are actually subclasses of a boolean class.
Of course LISP is yet simpler... But I agree, Python gives almost the ideal balance for most programmers (personally, I like the LISP approach though -- I think Gambit-C SCHEME with "six" syntax is better than Python, but that's just me).
For a nice visual, try
http://www.mint.com/blog/trends/china-vs-united-states-a-visual-comparison/
As you can see, China is bigger with manufacturing, US is bigger with services; China doesn't have the debt, or the excesses in the stock market. China has cash, and gold reserves. China doesn't have external debt. China exports more than it imports.
About the ONLY metric that China falls behind in is GDP -- which, from the other indicators is simply being propped up by... China.
(Oh, and the US has more energy reserves).
Have fun with the easy graphics. The GP was right -- China pretty much owns the US.
The "internet" started with the PSTN. Using modems, and the "internet" inherited the reliability of the the PSTN. Of course, having multiple paths is more reliable, and, unlike the PSTN, the "internet" is more tolerant of less reliable connections.
As more data needed to be shipped, things like fiber were implemented. Packet switched, just like the "internet". Now, the PSTN uses that same infrastructure. Instead of the "internet" being supported by the PSTN, the PSTN is supported by the "internet".
Which means that the "best effort paradigm of the Internet" DOES work for all applications, including telephone type services.
Now, local telephone loops supply power separately from the grid. If the grid doesn't function, the telephone may still be powered. An accident of history.
PS. There was no guaranteed quality of service from the phone company either. In fact, it was a running joke while it was still a monopoly.
Lily Tomlin: "We don't care; we don't have to. We're the phone company."
Um...
Wouldn't sleep/hibernate functionality be the job of the hosting OS?
Please read the article:
"The respective proprietary graphics driver for each operating system was installed"
Wrong on a few points. Copyright doesn't mean I can't make money off a work that I don't have Copyright for. In many cases (not all, there are explicit laws preventing this universally) I can rent the Copyrighted material out.
For example, I could create a "pay-for" library. No copying takes place, and I make money "showing books to the public".
I can remove the signature from a painting, and then sell it. I can even claim I painted it; that would be plagiarism, but not a copyright violation.
Actually, Fonts are NOT copyrighted on the basis of their letter forms. It is perfectly legal to scan a font, and sell the reproduction.
Font NAMES can be copyrighted, and digital data can be (and, by default, is).
Not the actual letter forms.
Oh really...
What about removal of the big kernel lock?
What about plug-in resource schedulers?
What about fast ip locking?
What about kernel video mode switching?
What about systemtap?
What about cgroups?
And much more...
When taken in combination, the growth of the Linux kernel since 2003 definitely warrants a major jump.
The issue is whether it should be 2.8 or 3.0. I would side with 3.0.
Because Linux is now ready for serious MP, both on a local and a cluster level. And these features are not "backwards portable".
In the original message you said:
"And it's not on Linux? Only when you limit yourself to the "distro-specific app store, a.k.a. apt" do you get system-wide updates. Just like I do if I limit myself to MS software. I choose not to."
I just pointed out that those application vendors offer repositories. These repositories are "outside" of the Linux application sphere. Once added (and keys confirmed) the various software components are updated by the same updater that takes care of the OS.
Which means Linux users don't have to limit themselves either. And, the benefits of having controlled repositories can extend to third party software vendors.
What I really don't understand is why Microsoft Windows doesn't offer something comparable. Actually, Microsoft does, in the hardware space. Once a driver is qualified, Microsoft will distribute that. But, there appears to be no mechanism for extending the Microsoft updater to include third parties.
Maybe I am wrong (I am not a Windows user, really -- I just use Virtualbox/Microsoft Windows 7/Office/Project for compatibility reasons), and I really don't use (much) non-Microsoft Windows software (Adobe and putty). I do know that Adobe and putty don't use the Microsoft Updater, but I don't really understand why.
Data recovery "bit level"? I just use the standard POSIX command line tools on the primary disk (done on Linux, although I could have done this on another of our systems):
sudo cat /dev/sda | grep -ab black
128092793: COLOR: black;
128093889: border:1px solid black;
147496572: var strHiliteSearchSetting = "enable,yellow,black";
147496626: strHiliteSearchSetting = "enable,#b2b4bf,black";
Now, I can use dd to extract the bit of data I want:
sudo dd bs=1 skip=$((128092793-200)) count=400 if=/dev/sda
cal-align: middle;
height : 8px;
text-decoration: none;
display: table-cell;
font-weight: bolder;
font-size:65%
}
A.tableheaderlink:link, A.tableheaderlink:visited {
FONT-WEIGHT: bold;
COLOR: black;
vertical-align : middle;
text-decoration:none;
}
A.tableheaderlink:hover {
FONT-WEIGHT: bold;
COLOR: red;
vertical-align : middle;
text-decoration:none;
}
Really, this stuff works on pretty much all OSs (AIX, Solaris, Linux, HP/UX) out there. Not so sure about Windows, but I would HOPE it works (Microsoft does claim POSIX, don't they?).
With additional knowledge about the disk layout, a systems engineer can recover more information. Yes, I took a course on UFS, and Solaris Internals, and I can recover data from Solaris, Linux, BSD, AIX, and HP/UX. What is important here is that a client can call, and I can do this -- I don't need a "custom toolkit" for it either -- just OS level commands. But this is predicated on knowledge of the file system. I presume that this service can be supplied in the Windows world?
Never reactivating Windows -- good for you!
I put Linux on my laptop. Fedora 14. It resized the Windows partition, and installed as "dual boot".
No need to reactivate on starting Windows. Next, I installed VirtualBox, created a virtual machine that would use the Windows partition directly. That worked, but I had to put a call in to IT to have Windows reactivated.
Now, switching from hardware to virtual (P2V) is like changing the mainboard.
So, I don't understand your experience. Maybe you are using a "bootleg" Windows? I sure don't. Mine is 100% legitimate. Because the BSA can audit us any time.
On to repositories. Fedora repositories include:
adobe
dropbox
google
google-chrome
skype
virtualbox
(and, I imagine, more -- these are the "third party" application repositories I use).
As to "reverse engineering" NTFS -- as long as you use Windows, it's not needed. Sure, why the hell would you? It's not like you would need reliable data recovery tools, or ever need to repair a filesystem for someone. After all, everyone does backups. (sarcasm ends)
Um...
I considered modding, but "wrong" is not a mod.
B-Trees are certainly not the fastest way to store and retrieve data. You may want to investigate those strange things called "hashes". Fortunately for you, most database systems know how.
As well, "assembler" isn't the fastest coding system. It is so very slow and tedious to use that algorithm tends to get overlooked. Back to your "B-Tree" again -- have you tried coding a B-Tree in assembler?
To give a hard example -- back in the days of OS/2, the filesystem was coded in assembler. I imagine that it was considered the right way. After being recoded in C, the thing ran faster -- mostly because the developers could concentrate on better ways of doing things, rather than the drudge work of getting the assembler correct.
Another example -- replacing a "C" coded inner loop in a console game with a scheme interpreted version, that ended up being faster because it could fit into cache.
Sometimes upstream is simply better, even for performance.
Database systems are in the same category. It is rather difficult to optimize specifically, and, by the time you did this, typically the data store would be "out of touch" with the actual business requirements. So, most of us use a good database, with a query optimizer, and let it take care of that part of the problem. Yes, it may need some tweaking, but those tweaks will tend to be independent of the business logic. This can give rise to specialists who can be called on if needed.
Using your model, the "assembler tuned B-Tree" custom coded, there would be no paladins available to assist in the tuning. Indeed, moving to another architecture would be difficult (or impossible). It would be ridiculously expensive, and would likely never be upgraded. Hey, CICS and IMS are still in use, aren't they? And THEY are more portable than your suggestion. (CICS and IMS just turned 40).
PS. There are CICS and IMS paladin around. There wouldn't be any for your "solution". Also, IMS can use direct storage links, and, because data in IMS is hierarchical, even IMS can outperform "B-Tree".
Britt
Feedback -- sure, why not. But, it has to be in the form of questions.
1 - does your powerpivot appliance work with IBM ISAS appliances?
2 - does your powerpivot appliance work with IBM BCU?
3 - how does your appliance interoperate with Ab Initio? McSource? COGNOS? (8.4 and 10). If a customer has, say 6000 or more ETL jobs running with Ab Initio, and wants to start adding powerpivot functionality, is your appliance represented by a new GUI box to simply drag-and-drop in that environment?
4 - your appliance is configured to provide "backup" -- now, I don't have time (right away) to dive into your offering (week after next is the earliest, I have to present a Data Warehouse/ETL BI datacenter design this week). How do you intend on competing with TSM?
5 - please provide performance figures I can use (simply 24 core/96 GB doesn't tell me much). SPECint or nPerfs to allow me to compare against IBM P5 or P7; I don't really care about comparing against Oracle in this space, but T3 may be useful as well.
Thanks in advance. I presume that you have the answers -- Give a reply here, or simply email me at fred (dot) weigel (at) treklogic (dot) com
But sharing the information here may increase your exposure. Honestly, excel is used to deliver information views for BI, but I have never considered Microsoft as being a "core element" in a DW/ETL environment. And, given licensing fees for Ab Initio and the IBM appliances (a BCU is just south of a million), there certainly is room for competition.
This solution bundles the graphics adaptor with the sending of screens.
Yes, that's wrong. But its wrong because it sends too much data. The WORST amount of data is the screen itself. However, the graphics card itself takes information and generates that screen. For example, texture mapping. Why send all that texture data? Instead, send it once!
The idea would be to send higher level graphics commands via wireless, and then generate the actual screens at the device. An external, wireless graphics card. Now, this model fits the older "obsolete" OpenGL better than the newer OpenGL, which may be its downside.
But, the model should be similar to the "X" protocol (on which OpenGL rides). Commands at this (transport) level to be sent to the device. The software vendor optimizes to this level, and the card vendor implements, either using local transport, network transport or wireless transport.
This fits well into the Unix GUI approach -- maybe someone wants to carry this further into a Windows(tm) DX10/11 analysis?
I thought I had covered your points in my submission. I am going to quote (ellided) myself:
"In the case of my neighbor, that isn't a problem. But my best friend has his own network; but is located across the country. Data exchange with him would be impaired without my ISP (and, my use of /. would be curtailed)...In previous years...I shared this data with my distant friends via transport of CD, or hard disk drive...So, the mail and airplane system was part of my "Internet". It can be again, if those physical pipes are regulated to the point where the data will be inspected."
Note that I "mesh network" with my neighbor; does that make us ISPs?
But why is the reference a POINTER?
How about an implementation that uses (say) 4 bytes for an object reference. But directly encodes small integers and characters up to +/- 2 ** 30, if the upper bit is zeroed (31 bit integer). If the upper bit is one, the lower bits can represent an index into a pointer table, which can, in turn be a 64 bit pointer.
This would allow efficient representation of up to 2 ** 64 bytes of object space, and a unified efficient representation of self-representing integers and characters, and allow up to 2 ** 30 objects.
In the memory space used by the OOP programming system (running 64 bit), map the entire first 4GB to "not accessible".
And, the object space can be completely separate, and not directly reachable or corruptible AND an "object reference" would not be usable as a pointer to subvert the type system.
Just saying... I would prefer that sort of design. As a nice bonus, memory could be completely reorganized without invalidating any object references.
And, you still have the property that only 4 bytes needs to be copied instead of an entire object. Note that Windows uses this sort of design -- the object reference is called a HANDLE.
Copyright
Do you imagine that you have the right to copy a song from a CD to an iPod? In some jurisdictions, you do; in others, you don't.
Indeed, in MAI v. Peak, the Ninth Circuit Court found that copying Software into RAM infringed Copyright. Which was not the intent of the law (US); so changes were made allowing for this, and archival copies. If it were not for those changes (remember they are quite recent, MAI v. Peak was decided in 1993), it would be a Copyright violation to transfer a CD to an iPod in the US as well. But I don't know the law in France (and my French is not good enough to do French legal research).
I think that, perhaps, you are confused by what the "Internet" is.
The Internet is the network of networks. When I connect my network, which, for the record, is under my sole control, and not under any other regulation, to another network, we have a piece of the "Internet".
The Internet is facilitated, but not defined, by the "backbone" -- fiber optic connections (currently). The Internet does not mandate the form of these backbone interconnects. They exist because, simply, there is a need or desire to send that much data between the interconnected systems and networks.
So, what are you proposing should be regulated?
Nothing in the description of the "Internet" mentioned content, or physical interconnection requirements, or physical systems.
Read the article -- the point is brought home when the author mentions "Bob" and "minitel" as being REQUIRED to connect... not to the "Internet" but to regulated information providers.
As an example - my neighbor has a network of around 5 systems, and I have a comparable network. We connect these networks using two different ISPs (a cable provider and a DSL provider). We also occasionally directly connect with 802.11g wireless. The connections are made to exchange data (photos, videos, Linux distributions, it really is none of anyones business). What regulations would you want to put into place, and where?
Because 802.11g, ethernet, etc. are just convenient standards that permit us interoperability. We could come up with our own local protocol if we had to. The ISP provides a higher degree of interoperability (an IPv4 address that allows my network to be reached from other ISPs that co-operate, packet services that allow data to be exchanged, and access to DNS services to provide a convenient naming service). There really isn't any service my ISP offers that I couldn't easily replace, except for access to the high-speed backbone.
In the case of my neighbor, that isn't a problem. But my best friend has his own network; but is located across the country. Data exchange with him would be impaired without my ISP (and, my use of /. would be curtailed).
I thought the article was actually very well thought out. The point is that the "Internet", being a simple connection between systems and networks, only exists because it is beneficial to the connecting parties, and the method of connection is irrelevant.
Now, as soon as enough systems and networks participate, there is a "network effect", and some of the "Internet" may degenerate into a pure producer/consumer relationship. I imagine that this is what you want to regulate. Be careful though, because the end game of regulation is simply that some parties may decide to pack up their dolls and leave. We are not TIED to the infrastructure. We have always had a bit of balkanization in the Internet. Certainly, not all of my content is publicly available. Some is simply not directly addressable, or findable via Google (darknet, an example of this would be my music collection, family photos and videos, which I share to friends but no one else). Some is SSL only, and password controlled (my TV watching schedule, my calendar. And some is only locally available (banking statements, financial records).
In previous years (when high-speed meant 1Mbps down/128kbps up), I shared this data with my distant friends via transport of CD, or hard disk drive. 5 Gbytes of music was a lot to jam down the limited communication pipe I had. So, the mail and airplane system was part of my "Internet". It can be again, if those physical pipes are regulated to the point where the data will be inspected.
In other words
The Internet is dead; Long live the Internet
Why would an object reference be a pointer? A more useful construct would be an index to a pointer; then the object can be moved and all pointers would still be valid. Or, possibly a hash.
Anyway, I would definitely shy away from making an object reference even APPEAR as a valid pointer. The benefit is that "cracking" a system, and supplying an object to an OS level routine just wouldn't work. (eg. Put all object references in one 32 bit space, and map that as read-only). Characters and smallints can live in object reference space "as-is", everything else would need the double indirection. Yes, performance may suffer a bit, but wouldn't the system be more robust?
And, as a side-benefit, any low-level code introduced would be forced to use the correct access mechanisms, making it more likely to be portable.
It isn't necessarily obvious.
Basically, the idea is that bits of the key leak. And how is this accomplished?
For example - if a key bit is 0, you take one code path, if 1, another. Make the two paths different lengths. It may be possible to affect packet timing. Or... A function may end with "x - y" and then return "z". No leak? Not so clear, the carry/borrow may be leaking information to the caller (on x86 style hardware).
Anyway, it probably isn't a "back door", just some means of determining enough key bits to make brute force practical is enough. And this sort of thing can be subtle. It can even be based on the machine code generated for certain sequences by a particular compiler (the "x-y; return z" sequence above, for example).
In 1977/78, I worked on a system called PEAC. It was developed for CTW (Childrens Television Workshop), and used for Sesame Street production (among other programs).
A number of clicker boxes were plugged into a charging/syncing station. All of the clickers could be synchronized by a single button press. They were then disseminated into an audience. Both demographic questions (are you male/female, what is your age, etc.) were collected, and then polling data for the program was collected (like/dislike on a scale of 1 to 5, for example).
Afterwards, the clickers were put back into the syncing station, and the data was transferred to an Apple ][. If the material was sourced from a video tape, a secondary audio channel was written to the tape, with time marks.
The time marks were fed into the Apple ][ audio controller, and the floppy based demographic and polling data was then set up to produce interactive cross-tabulations, color swash (called "mondrian" analysis), and other investigative approaches. The bulk of the analysis software was written in Apple ][ Integer BASIC, along with machine code routines for actually performing the cross tabulation.
The Apple ][ was chosen because it supported "color" graphics, audio input, and floppy discs.
The system was delivered in 1977/78.
I don't know when it was decommissioned. But, I imagine it wasn't unique (well, the time synchronization made the "user interface" simple -- just fast forward, rewind and play on the video tape deck! The audio channel was composed of many short independent "files", each recording a time-stamp; and, I have never seen the "mondrian" style display for analysis again, but maybe it was a bust). But the patent is deliberatly vague about that end of things, anyway.
Now, honestly, we would have liked to use RF, infrared, or "mobile device" clickers. Instead, we had to develop our own stand-alone clicker device (and, honestly, I don't know if you can run 30 IR clickers in a classroom at the same time!). RF would have been an option -- but we put storage into each clicker, and synchronized the lot in the charging station. But, the need for clicker synchronization was a weak point, and we recognized that (back then). We could have used low frequency FM (47Mhz required a license, and cell and "wifi" didn't exist). So, low power 88Mhz was an option. Still, not as reliable as the hard-sync approach.
Another problem (back then) was that available AFFORDABLE computers would have been hard-pressed to reliably support 30 concurrent clickers. We would also have needed to add a processing layer in to support that. With a sync box, each clicker could be read out and converted to floppy in turn.
We did think of it. It was documented. Question is, does the paper from back then even exist? We never considered it patentable material.
Insanity: doing the same thing over and over again and expecting different results.
Albert Einstein
I am (finally) forced to use Microsoft Office(tm) -- the one without the menu bar (2007). With Microsoft Windows(tm) (7? not sure). I have to have it plugged into a network in order to use "domain login".
It runs anti-virus software, some kind of remote access software, and Office (because heavily formatted documents from others don't work well with OpenOffice.org). I don't have "administrative access".
Since it's plugged in, I decided to portscan it. 10 open ports (AFAIR). Compare and contrast to 1 open port when I open my (considered non-secure) laptop (22, for ssh).
Yes, SSH has had two exploitable vulnerabilities in the past 15 years. (well... 1, unless you use debian/ubuntu in which case 2). But, client computers really don't need any other incoming ports. I guess it's a good thing my Windows laptop doesn't have port 22 open!
As to JScript? I have been using NoScript. Blanket denial of JScript unless I specifically enable it. We know that pattern-matching heuristics don't work (even though I am forced to put up with it on my MS Office laptop).
Still, I guess that the universal access to computers that has been enabled by Microsoft Windows and cheap Intel hardware has been a net positive to society.
The solution would be to produce two operating environments -- a "starter" and a "normal". We already have this, in the form of Windows vs. Everything Else. But, due to the commercial nature of Windows, there is limited compatibility across the tiers.
tibit
What I find fascinating is your distinction between "toy" and "real" functionality. Canvas, to be useful to me, must remain simple. I only use it every 6 months or so, and I don't (typically) have time to read up on it.
I "hack" up a quick & dirty thing, run it a few times and then throw it away. Really, the original question was about "personal" programming.
Typically, I try to generate the graphic (if static) with a spreadsheet first; or gnuplot. If interactive, I try to feed gnuplot instructions, or use OLE and DDE (if I remember how, normally, my attempts in that space don't work out quickly enough). If that doesn't do it, or I need mousey input, I turn to tcl/tk.
Qt is way overkill. The only way I could do that is by building a bunch of templates, and cargo-culting it from there.
If I want Qt, I'll throw the prototype over the fence, and get some folks in development to redo it.
Sorry for the reply to myself.
It occurs to me that an example is in order.
Imagine an artist creates a piece of performance art. A photocopier, with a pushbutton that is simply marked "for content, press me". The input to the photocopier is sealed by being screwed shut, and a single page is loaded, ready to be copied.
You press the button, and a copy is made. Out roles a sheet that says "This work is copyrighted, and All copying rights are reserved". Behind the curtain is a government official, hiding.
You press the button, and find yourself arrested, and have to pay a $750 fine.
(If I were building this, I would simply write that on the paper)
What is interesting is the concept of "automatic copyright" introduced in 1976. If copyright had to be registered, a person MAY be expected (it would be difficult, but may be possible) to search copyright status and allowed rights prior to taking an action.
With the "automatic copyright", the only legally correct action is NOT TO PUSH THE BUTTON. But, what if the button said "For content, push me. I give you permission to push me." I argue that a transitive copyright has now been granted, especially since there is no other way to determine copyright status.
In other words, downloading is legal.
Downloading is NOT copyright infringement. Everything is "copyrighted" on the web. Even the google home page!
So, if downloading copyrighted material WERE infringing, we could make it really simple and just close the internet.
As a corollary, there is nothing wrong with searching for material, and including terms like "torrent", or "bittorrent".
Not the slightest hint of illegal. NOT THE SLIGHTEST.
Downloading is... just downloading. Searching is... just searching. Copyright infringement is copying a work without permission of the copyright holder.
Let's bring this back to web pages -- it is perfectly legal to DOWNLOAD that web page, but may not be legal to copy the bits and store them on your hard drive. It is perfectly legal to DOWNLOAD that web page, but may not be legal to print out a copy.
Interestingly, if the page has "Copyright (c) year, some person / All Rights Reserved", or NOTHING on it, a strict interpretation of copyright would lead to the above. Since browsers make a COPY in the ordinary course of operation, and the web site content provider is (should be) aware of this behaviour when preparing content, it can be argued that, at least, that copyright was granted. In other words "All Rights Reserved" makes no sense in the context of the web. This applies further to other downloaded content.
It is perfectly legal to download, as long as further copying is not engaged in (actually, in Canada, you can download music, and then further distribute as well, under certain circumstances).