should we create an OS development magazine?

[01000101]

ok, i finished the upload page.
pm me if you're an editor/writer that wants to upload an article/tutorial to the site.

I'll start working on locking-it-down so that I can just allow uploads based off a password.

[jerryleecooper]

I have a problem with my article on paging, it's not doing well. I wrote 17 lines in Word and now I find it boring to continue it. I want to create things I would like to know and to read, not things I already know. But because I know well paging, I tought that it would be easy writing an article about it, it's not because it's boring And I don't what I wrote is of great quality

I have some other ideas. For example, we could make a bank of trivias to put into each issues. For example:

Code: Select all

WHy does DOS use the backslash to separate the folders?

Because when it was time to implement the hierarchical filesystem, Microsoft already used the forwardslash, as in UNIX, for the switches, because of tradition as in DECnet, so at the request of IBM, they used the backslash.

And there would also be "chipset" articles, about the chipset of something, architecture articles, etc.
Each kind of articles would have its own icon.

With this message, I want to say that I prefer to write a lot of little pieces of articles, or help others with finding information for their articles, than writing big articles . A big article for me is between 150 lines and forever . But it is not amusing because I promised to write the article about paging and well, I will continue it, I am inspired now that I wrote this message.

[jerryleecooper]

Here is what I wrote yet about paging, what do you think of it?


Paging, you surely know that your hard disk partitions get fragmented over time. What it means is that files in your partition are not stored linearly; part of them could be stored anywhere on the linear place defined by the partition entry. To permit this, file systems have a special data area where they store the position of each chunks of each files on the partition. With this system, there’s no need to move files to make space when you want to put a new file on the disk. You just need to know where there is space to put each file fragments to, and your file is on the disk. But that mean your files are not always stored linearly, and a performance penalty ensue. So you need sometimes to run the disk defragmenter.

Paging is a bit like hard disk fragmentation, but for RAM. With pagination, you get to decide how to link together each and every little chunk of 0x1000 bytes of size of RAM data in order to make from non sequential chunks, called pages, a unified bigger chunk of memory. The file system is to the hard disk what the memory manager is to RAM, hence pagination is useful to the memory manager. How?
Suppose, knowing what pagination can do for you, that you decide to use it in your scheme where you have a bitmap in memory where ON mean that the page is used, and OFF that the page is free. If you want to allocate 0x1234 bytes size of data, you search for two free pages, put their address into the page table *sequentially* in an area where there is nothing yet, and give to the function calling the allocator the address from the page table. This is said simply. It will be simpler with the code.

(edit, it looked bigger in WOrd)

[os.hacker64]

good, but obviously not complete.

[RUMyMummy]

You can get that much from any old textbook or internet article. Shouldn't you be more practical in a migazine of this nature? You know, start giving examples and source code after about a paragraph or two of introduction.

[t0xic]

Do you want me to post what I have on quantum computing so far? Not finished yet...

[os.hacker64]

do it

[01000101]

The paging article is good, but as stated earlier it is not complete.

examples are good, code-blocks are great, but try to drift away from the bland text-book lecture-on-paper style.

Please post anything you have guys. that goes for the quantum computing article as well. we need substance to the magazine, its a crucial part (obviously).

[t0xic]

Here's what I have so far...

Code: Select all

	Most of us have at one point or another heard the term "quantum computer". What is a quantum computer, you ask, and how do they compare to "normal" computers? A digital computer works in the 2 dimensional realm of bits - which can be read as either a '1' or a '0'. Digital computers can perform only one operation at a time - due to physical limitations and physics in general. A quantum computer on the other hand, doesn't work in the realm of normal physics that we are used to. It works with quantum physics (mechanics). Quantum mechanics deals with matter on the atomic level, which means that it represents data as quantum bits (qubits) - which can physically be either a '1', a '0', or a superposition of both. A superposition means that the physical representation of a qubit (an atom, photon, or other atomic/subatomic particle) is both '1', '0', and all/any values in between - at the same time. This possibility to represent multiple states (the "quantum superposition") is what allows a quantum computer to run much faster than a normal computer - and executing multiple operations at the same time. One limitation to this is that we when measure (or "read") the value of a qubit, it must choose a 2 dimensional state (a '1' or a '0').

	Why should we want quantum computers? Quantum computers are best suited for large scale operations, i.e. cryptanalysis and integer factorization. A quantum computer has the potential to solve these large scale problems very quickly because of the nature of a qubit. Each qubit has the potential to represent an entire computer core, and each can be used to solve an individual problem. 

	Why don't we have quantum computers today? Because of physical limitations. For example, reading ("measuring") a qubit actually causes it to choose a state of either '1' or '0'. Without measuring a qubit though, you must assume that it is in the superposition state. One possible solution to this problem is called quantum entanglement, where you have two qubits represented as electrons, with one electrons spinning (the superposition) and another electron with a value of '0'. You force the two electrons to collide, and the second electron will be equal to the opposite of the first (b = ~a). You can then just measure one qubit, and you will be able to figure out the value of the other one as well, without disrupting it's spin (remember, measuring or observing the electron will change it's value). 

	Why should I care? Quantum computers 

	Overall, quantum computers will probably not have any practical implementations within the next decade, so don't start writing your parallel computing operating system just yet.

Edit away

edit: code blocks

[01000101]

interesting material you have to work with. great read overall. there were some gramatical errors and the fact that you mention its 'physical constraints' a little too many times throughout the article.

one error thats a simple fix is to reverse two words in the sentence.

that we when measure

should be 'that when we measure'.

I like this article as it touches on the future of computing (possibly).

refine and expand.

[t0xic]

Sure thing!

[piranha]

I like the statistic of a future quantum computer breaking the U.S. security codes in one second.

-JL

[Dex]

What about hacking articles, will these be allowed in the mag ?

[01000101]

I'd say yes for now, unless someone presents a complaint or other reasoning I see no reason for it not to be in the magazine.

[Solar]

Bad idea. Such a complaint will usually be attached to a lawsuit, which usually means a four-digit amount of fines payable by the person signing responsible for the mag. Been there.

I like the statistic of a future quantum computer breaking the U.S. security codes in one second.

Haven't read the article, but that smells fishy - Slashdot type journalism, that which a hacker likes to hear and heard elsewhere on the 'net.

As an example of peer editoring, my take on the subject: For years to come, such a quantum computer won't be a PCIe slot-in card or a different brand of motherboard, but something that resides in the basements of organizations-who-can-pay. And it won't be just "the US security codes", but the GPG, SSL, and SSH you use for your privacy. Of course there will be quantum encryption, too, which will be (reasonably) secure from such quantum computing attacks - and again, available for organizations-who-can-pay. Then there's the history of the US banning strong cryptography from export - which effectively failed with stuff like PGP because it's software. Quantum Cryptography is hardware...

Now I may be paranoid. But whoever is the editor of that Quantum Computing article should have some pointers on how he could extend his article by some research, to make it into something that has not been widely circulated on the 'net already.

This is not for the sake of the article on Quantum Computing, but a suggestion for Modus Operandi. Find two (or more!) people with widely varying views on a subject. Talk to both of them. Follow their pointers. Do research. Write an article that reflects both POV's, without actually judging either, so the reader can widen his horizons instead of being spoon-fed whatever the editor feels about a topic.

Disclaimer: I still don't want a part in this magazine thing, I'm just being vaguely supportive.