[alexander]

No one can know exactly what NSA is doing and talk about it publicly without worrying seriously about being sent to Gitmo in a private jet.

NSA's been putting lot of interest in SELinux though, some SE patches have even been written or more asked for by the agency, here's an article off of nsa.gov (http://www.nsa.gov/selinux/)
The thing is that SE has the potential to be the most secure, the permissions that you can set on there are crazily crazy, there was a box available online that you can telnet and ssh into as root, but you could not do squat, but you were root the permission set was insane, i think it never got cracked even after it got slashdotted and everything, it was kinda cool, my friends spent a few hours everyday for a few weeks trying to find a loopwhole, there were people who were even more obsessed, but natta, nothing to my knowledge noone was able to do anything signifficant...
And it does not mean that the NSA runs SE boxes everywhere, but it is safe to assume that they at least have a few, however it can be argued. (P.S. they probably run OpenBSD on some routers, it would make sense, OpenBSD is the safest OS on the market, its security measures cause unnecessary overhead, but on a decent box that is just supposed to be a router, its not that big of a deal. I admire OpenBSD, they have been progressing to provide a functionality of any Cisco device or software on your own system, you can already do things like load ballancing between nics, routing, firewall(full blown filtering by anything), IDS and some other nifty features that will allow you to turn a box into an awesome, smart and fast router for free.)

True to an extent, though some well designed GUI features can be nifty. In past years I was quite at home with VMS command lines. I also enjoy hacking in C/C++ in which you can do things that Coffee would never allow you to do, but I'm not enamoured with Posix notation and I think an OS command language should be clear and easy to remember.

well, let me put it this way, a programmer explanation, you can write a program to do anything you want in a text editor, and although it might take you a few more key presses to create windows and generate menus, it is all doable, so C++ compiler and a text editor is the most powerful tool you can have (yes you can argue that you can write in binary or assembly, but for the time matter, i'm just relating). Well, that would be your command line, bash is insanely powerful and useful, but you can also do what they refer to as Visual programming. You can drag things into places, right click on it and assign them actions and names, and yes if you are short on time and dont care about performance, sure there are some nifty features of Visual programming, but is it anywhere as powerful as a programming in a text editor?
Aah, Posix, why are you using posix again? there are better unix interfaces than straight up posix, try Bash, tab completion is your friend, plus bash scripting is many times more intuitive and if you are a scripter, you'll find many similarities between bash scripting and perl, actually more of perl/python mixeroo. I personally have nothing against looking stuff up in man pages, you need to know basic less commands and it becomes a whole lot easier as now you can search for whatever it is you needed, if i'm working on something i constantly have one to 2 (many times while programming in python i'll have 3-4, but those would be referencing pydoc) teminals with opened up man pages that way i can reference to them on the fly, and you have to remember, the man pages are 100% more informative, make much more sense, are written by the developers and turn out to be exact 999 put of 1000 times, and the 1 time left is when the program is still a beta and had recent impovements, but man pages have not been updated, then Windows help, and tutorials and wikis for linux are many times more easier to follow and figure out what's what then microsofts FAQ, oh and there is no "it sometimes works" that my frieds found on the M$ site while seaching for information about active direcory i think, I'll have to ask again, but it was hilarious

[Qfwfq]

We seem to have a few misunderstandings

Originally Posted by alexander

well, let me put it this way, a programmer explanation, you can write a program to do anything you want in a text editor, and although it might take you a few more key presses to create windows and generate menus, it is all doable, so C++ compiler and a text editor is the most powerful tool you can have

I won't argue that you can write in binary or assembly, although I used to do it around '80. I wasn't talking only about programming and programmers, what I said about developping included complex web development. Try doing maintenace on an Enterprise webapp like the one I'm currently on.

The author of one Wrox book writes that, when the PC-journalists ask him which IDE he uses, he smiles and answers "Visual Notepad" but I think he was doing only JavaScript, I can't remember for sure but he wasn't doing EJB, servlet and also the whole chain down to the HTTP response.

Originally Posted by alexander

but you can also do what they refer to as Visual programming. You can drag things into places, right click on it and assign them actions and names, and yes if you are short on time and dont care about performance, sure there are some nifty features of Visual programming, but is it anywhere as powerful as a programming in a text editor?

I've tried out direct use of Win32 for windowing, just to gain better knowledge and understanding of what's under the bonnet. Surely you wouldn't reccomend developping an actual management project for a customer without using some app framework or another?

Performance is hardly an issue for the bare GUI itslef. If the execution on a user's click can take noticeable times you might need or want to optimize it, but this isn't really a matter of the actual GUI. Unless part of the trouble is interacting with controls, but even the overhead of this can't be improved much by use of Visual Notepad or Visual VI.

If execution upon a user's click takes no more than a 10th of a second, how much sense does it make to reduce that time? Even if you bring it down to a few microseconds or less, no customer will be glad to pay you more for such a feat. One customer was hopping with content because, when he ask me to improve a functionality that was monthly used by accountants to produce bunches of Excel worksheets according to data on DB2, a worksheet for each product, I also reduced the execution time which was measureable in hours and minutes. I brought it down by perhaps half or a third. That's a different matter.

Originally Posted by alexander

Aah, Posix, why are you using posix again? there are better unix interfaces than straight up posix, try Bash

I said Posix meaning the whole Posix compatible category. I have tried bash and I like it better than the other shells, though it is limited to its purpose and it certainly isn't C! Bash would be better, even just if you could define a function to return more than a byte, and use the return value in a more direct and comfy way. When I read that functions can be called recursively I thought, OK, let's try it, classic exercise: factorial(n). What a nightmare! It takes getting used to but, even when I got it straight, the limitations are obvious. I guess I'm just a C/C++ guy and I'm discontent enough with Coffee too. I've heard Perl is great, I'd probably like it if I took the time.

Originally Posted by alexander

if i'm working on something i constantly have one to 2 (many times while programming in python i'll have 3-4, but those would be referencing pydoc) teminals with opened up man pages that way i can reference to them on the fly

I take it you're not working on VTXYZ terminals but console emulator windows. If so, you are nevertheless exploiting a GUI feature!

Originally Posted by alexander

there is no "it sometimes works" that my frieds found on the M$ site while seaching for information about active direcory i think, I'll have to ask again, but it was hilarious

Yeah, I've seen and heard hilarious things. Have you ever heard the rumour about the back door in some versions of IIS?

[alexander]

I take it you're not working on VTXYZ terminals but console emulator windows. If so, you are nevertheless exploiting a GUI feature!

i do run a gui, Fluxbox, why, because this is a laptop, and performance is not that big of issue and its not a server, a small gui, with semitransparent everything does a good job at impressing friends one of which already switched to linux, so its for the good. But many times if i'm coding on my friends box over ssh, screen does a splendid job at creating as many virtual terminals as i need, so, no, not necessarily exploiting gui features... (I'm actually working on a good configuration of fvwm, there are a few problems still, but eventually i want to be able to press a key combo, and see all my windows move and hide at the nearest edge, oh and i love transparency )

Have you ever heard the rumour about the back door in some versions of IIS?

Which one lol, there were quite a few

I said Posix meaning the whole Posix compatible category. I have tried bash and I like it better than the other shells, though it is limited to its purpose and it certainly isn't C! Bash would be better, even just if you could define a function to return more than a byte, and use the return value in a more direct and comfy way. When I read that functions can be called recursively I thought, OK, let's try it, classic exercise: factorial(n). What a nightmare! It takes getting used to but, even when I got it straight, the limitations are obvious. I guess I'm just a C/C++ guy and I'm discontent enough with Coffee too. I've heard Perl is great, I'd probably like it if I took the time.

there is are packages that allows you to script C in which case you win no matter what
Perl is great for text edditing anything, although as programmer i think that Python is a better language as a whole, I'm waiting for the new perl to come out, then i'll learn it, it will be object oriented and have a whole bunch of cool new features worth my time. but perl does have the best libraries for dealing with strings in general, their regexp library so good that it is used in PHPs engine, they decided not to screw with the best regular expression engine ther is
I dont know whether you'd like perl, the syntax can be very discombobulating like the attatchments:
code1 - insane perl program, just open it up, if you see a picture of a guy, you're golden
code2 - gives a new meaning to hello world (should see globe)
code3 - a little some something for science (sould see e)
code4 - a little something for hypo (H)
(no none of this code is mine, it is however from my collection of code so i dont know the authors, but most of it is from perlmonks.org... (actually think that all of it appeared on perlmonks))

[Qfwfq]

Thanks for the files, I'll have a look shortly.

Originally Posted by alexander

I dont know whether you'd like perl, the syntax can be very discombobulating like the attatchments:
code1 - insane perl program, just open it up, if you see a picture of a guy, you're golden
code2 - gives a new meaning to hello world (should see globe)
code3 - a little some something for science (sould see e)
code4 - a little something for hypo (H)
(no none of this code is mine, it is however from my collection of code so i dont know the authors, but most of it is from perlmonks.org... (actually think that all of it appeared on perlmonks))

C/C++ can also be discombobulating, if properly used!

I got TeleMad, your old firend, fuming. I posted something that can be a help with factoring, on that thread, hoping things had calmed down enough, but he came back in a rage. I'll be seeing what sort of a mood he's in, when I get round to it.

[Qfwfq]

I saw they guy, despite thinking I was opening n° 4 and having to scroll!

Do these actually execute and do something? I don't have perl installed here and I'm not sure I can try them.

[paultrr]

Code 1-use strict;BEGIN{$SIG{qq)__WARN__)}=sub{$..=$_[$===!$=]}}eval,$..=$@for
qq}bless("Wall","Larry"),Not("Wall","Russ");#},qq! ;print Magic,counts!,
qq#use strict;\$Perl=42#,qq>goto"WALL, LARRY ONTO (OR ON?) SOMETHING">,
q~warn+Randal::Nevermind()~;s~~pop~e;$~=/^[1-7]$/?~-$&:rand$=>>$^F+!$%;
@; =(
q~ -F+=->*/+/*+*W*. ->
*/ *.+=*+*.*/+/*+*3*/-B*+* ,+
1* +*1*\*\*\*,*2*D*2*:*c*W*,+1*\* \*
\* ]~,q~ -Z*- *c
*/ *:*2 *c* W*
,+ 1*2+ 1*/ *7
*< */*. + =*2 */
+1 *++` * /*c +`
*/ +1*+ *G , I+= *.
*2 *:*/ * A *+* c*
2* /*W *, -> */* 1*
7* 7,G *2 + 1*/ *7
*D *2* :*/* c +`* /*
3+ =*7 *G*1 + 1*2 *:
*/ *+* .*]* 3 *7* c*
2* /,, *2*:*1*/. + */* A*
7+ 1*2 -> * /* +*
,+ 1*/ *3 * +* W+
1- >*/ , * -: +=
*+ *.* / *W*3* /*
.+ =*+ * .- >*
/* 3+/ * 2*+*:*W*,,I-Q,**]~,q~-F*0*D*W*7*G*3*1+ `*
/- Z*/ +/*+*3*/*2*W *.+=*2* : * /*7*,*. * 7*
/* 3*7 * c *2* .+= *W * , ,I ->* /*7 * :*
/* 7*, * / *3*7*c*2*.+= * W *,,I*\*]~,q~ - ;*
1* 1*/ * 1 * + * ,,
I* G* +,I * 2 * /+ 1 *2
*3 *W , I*,*2 * : *3 * /*
+* :* 2 * / * + * :*
:* 7, I *+* ,*.*\*/+X*7*2*3 * /+/*W*.+=*/* .+
=* 2* / * . * 2*
:* : * W * . *7
*: + ` * \*\ * \ */
.R * 9*T *] ~ ,q
~- Z * .*-*3*/*.+=*2*/-V*+,I * W*
A* / * .+=*+*.*/*A*7*G*,*.*3*\* ] ~,
q~ - B *+*,+1*+*1*/*A*+*,*/+/*:*W * .*
2* / * 7*,*2 *9*1*W*,*2*:*3 */*+ , I*
+* W * ,*\* / - W*D * 2*
:+ `* 7*, *2*/*W*3*/+=*+,= ,=+ `-
>* / * +*
,+ 1 * /,
=* 2 * +*
A* 2 * / *3
,= * : * 2*
++ 1 * 3 *]
*7 * D * 2*
:* / * . + =*
2* / + / + =*
7* 1 *2* /- W *+
*: * .+=*\*]~,q~ - Z*
/, G*
,*7+/*/*W*.*-*3*/+/*2*W*:+1->*/*0*G*.*/*W*.*/+1*7*2*3*/*c*+,G*2*/*W*.*/
*2*+*3*W*2*:

Is what is come out to in text.

[alexander]

C/C++ can also be discombobulating, if properly used!

aah, job security
here's a good one: http://www1.us.ioccc.org/2004/newbern.c

I got TeleMad, your old firend, fuming. I posted something that can be a help with factoring, on that thread, hoping things had calmed down enough, but he came back in a rage. I'll be seeing what sort of a mood he's in, when I get round to it.

thats so easy to do though. I am trying to learn to take it easy with him, if he posts something that enrages you, ignore it, just post your thoughts in peace. Oh, and by the way, i"ve rewritten that source, it ended up being in a class with 2 overloaded functions, one allows you to do the recusrion method, while the other one would allow you to do it all the fast way. The thing is i really didnt feel like posting that code, i dont know what the fuss is, its not that i dont want to write bad code, its that i really dont care for a program like that for a datastructures class, although i do understand the point of factoring. I'm currently working on my final project for that class, but i've also been writing a large number class that would easily allow me to do math with large inegers, so far adding 2 10000 digit positive numbers with their generation takes less than 1/2 a second, but my multiplication process is crapping out, some logic flaw that's bothered me for a while, so i've been avoiding it hoping to get a revealing idea about how to do it all better...

[Qfwfq]

That's a good one, the IOCCC. I don't do it for job security though, and I don't go to the extent of using #define solely for illegibility.

The only time I delighted in making code as tricky as could be done in Coffee in a job for a customer, was when I had been put under a young project leader that was treating me like an idiot and there wasn't much chance of otherwise changing her attitude. Boss thought the world of her but imv she was little more than a student and certainly not enough of a genius to justify being so haughty. It was a non-trivial thing anyway but I was packing as much as I could into single expressions and so on. One evening she tried looking at it because of new specs, when I had left for home, she had hoped to make a little change or two herself and she had a fit. It caused a great row but eventually she realized she had no grounds to treat me that way.

Originally Posted by alexander

thats so easy to do though. I am trying to learn to take it easy with him, if he posts something that enrages you, ignore it, just post your thoughts in peace.

Yeah, I'm tring to ignore his outright insolence while showing that my execution times are good.

Originally Posted by alexander

but i've also been writing a large number class that would easily allow me to do math with large inegers, so far adding 2 10000 digit positive numbers with their generation takes less than 1/2 a second, but my multiplication process is crapping out, some logic flaw that's bothered me for a while, so i've been avoiding it hoping to get a revealing idea about how to do it all better...

I developed a class Rational, which uses a class Looong, but there is still room for improvement. Looong can add and subtact integers of over a thousand bytes, it would be straightforward to change the max length, it builds them up by multiplying unsigned longs each with an exponent and it is used by Rational to factorize the results. Instead of giving Looong full arithmetic, Rational has it but this implies the sometimes heavy factoring after algebric sums with the limit on prime factors.

For your query about the count of a number's factors, prime and not prime, I thought it would be a good idea for Rational to also give users a way of iterating over the prime factors, without making the class fragile. This could be done easy enough, when I get round to it.

[alexander]

Looong can add and subtact integers of over a thousand bytes, it would be straightforward to change the max length, it builds them up by multiplying unsigned longs each with an exponent and it is used by Rational to factorize the results. Instead of giving Looong full arithmetic, Rational has it but this implies the sometimes heavy factoring after algebric sums with the limit on prime factors.

what i'm writing can add over 3 billion digit long numbers, but again, so far its only capable of adding 2 positive integers with no floatpoints, cant do much else, but when i get time to start working on it again i hope to get it to multiply (not by adding the values over and over and over again), subtract and divide as well as finding the modulus (pain, if you have some kind of an algorithm, please i'll be looking for it eventually) and do exponentiation (not via multiplication (also need a mathemagical solution, maybe some base shifting math exponentiation craziness))... eventually hoping to come up with something that will do RSA encryption the way that it was intended to be done (with really really large numbers), really stinks that bc does not accept input from the command line, bc is so awesome...)

[alexander]

my execution times are good.

build in the timing function, store the system time before start in one var... you know the drill, then you'd have exact time of execution of your function, then run TM's program, compare the results on the same system, then you'd have the necessary data to back up your point
In any case his algorithm is not that great either take some 50-60 digit long number and it will take him years to get the results, if you are talking about factoring here is a few thoughts:
here's wikis on most commonly used factoring algorithms (I personally prefer gnfs although some curve factoring is needed in large numbers):
http://en.wikipedia.org/wiki/Categor...ion_algorithms
RSA's favorite algorithms:
http://www.rsasecurity.com/rsalabs/node.asp?id=2190
and little something in the quantum world and Shor's algorithms:
http://ece.gmu.edu/crypto/student_projects/physics.htm