Slashdot Mirror
Windows RT Jailbroken To Run Third-Party Desktop Apps
An anonymous reader writes "We all knew it was just a matter of time, now it looks like Windows RT has been Jailbroken. From the article: 'The hack, performed by Clokr, exploits a vulnerability in the Windows kernel that has existed for a long time — since before Microsoft ported Windows from x86 to ARM, in fact. Basically, the Windows kernel on your computer is configured to only execute files that meet a certain level of authentication. There are four levels: Unsigned (0), Authenticode (4), Microsoft (8), and Windows (12). On your x86 Windows system, the default setting is Unsigned — you can run anything you like. With Windows RT, the default, hard-coded setting is Microsoft (8); i.e. only apps signed by Microsoft, or parts of Windows itself, can be executed.'"
Microsoft locked Windows RT down because it wanted to slowly get rid of the Win32 cruft dating back to the 80s and 90s. That cruft does exist now and is used to run things like Office and Notepad etc. but Microsoft can easily rewrite them in the future. What will happen to Putty, VNC and the like then? They will break,and then again we will blame Microsoft for it. That's the reason to lock it down.
This space for rent.
All 3 of them.
Never heard of it.
This may border on being pedantic, but I'd call this a crack instead of a jailbreak. It sounds like they're just patching a kernel value ... not breaking out of a jailshell.
I expect MS will probably just find a way to patch it up in the near future.
Or Android? If so it might be possible to render these gadgets useful, even if it does require going through a song and dance every time you reboot.
Warning: this article may contain humor, sarcasm, parody, and perhaps even irony. Read at your own risk.
well then the appstore will NEED NO censorship other apps that crash the system. Also need to have a 3rd part app store like amazon app store for android
This trend of making it hard/impossible to run what you want on your computing devices is just despicable. I predict that not many years from now there won't be a commonly-used platform where you can download whatever you want and run it. We may be way past the year 1984, but we sure seem headed for 1984.
Imagine!
"Need" implies there are people using it, which is a conclusion we might want to take off our mat for the time being.
I foresee an update to Windows RT tomorrow (or soon thereafter) to plug this serious threat to user security (have to secure users from getting apps somewhere else that Redmond doesn't make money from)
Am I missing something here? How can anyone develop new applications for Windows RT and test run them?
"Windows RT Gains Solution to Allow Customers to Run Any Software They Choose"
And we wonder why people don't "get" Software Freedom. Somebody please remember to name the next software-freedom work-around "murder" just to keep the bad PR going.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
Still have to be complied to ARM right?
Interesting post. I'm no security buff, but whitelisting doesn't sound like it's inheriently a bad thing, and I don't think anyone would argue so, but if that's the route you go, the default should have to be that the user themselves is gatekeeper, with the option of enabling it such that they can use another party to manage their walled garden for themselves.
:P
Really, building your own walled garden of executables from places you trust actually sounds like a pretty clever idea. It also sounds like Linux repositories with a filemask of 110. Or maybe using a host file instead of DNS.
Support the EFF and Creative Commons. The war is coming, and they're supporting you...
Except the problem with your whole premise is that you forget the user.
Basically Apple "whitelists" what Apps can run under iOS (and are clearly moving that way for OSX too), yet people rail against it and even go so far as to remove the "whitelist" (e.g. jailbreak).
The problem comes down to who does the vetting and testing of an application to add it to a whitelist? If it is the user, they've proven they can't be trusted because they'll "vet" any new screensaver/antivirus/cursor application that comes along. If it is a central organization (Microsoft/Apple/Google/etc..) you then run into conflicts of interest in what they think you should do with the platform and what you actually need/want to do (e.g. what happens when you have a problem that can't be solved by any existing approved application?).
There is no simple single solution to the problem of security. A real solution by nature needs to be multilayered which means there is some complexity and ultimately users have to take responsibility for their actions. The idea that a single company/program can keep you safe just keeps perpetuating this idea that you don't have to pay attention to what your are downloading/executing and it's that mentality that allows malware to continue to be so successful.
They can set this selectively per environment as well. Microsoft sets it to "Unsigned" for the desktop but to "Microsoft" for "formerly-Metro" applications.
When someone says, "Any fool can see
Now that there is a win8 version in circulation which allows it, it is too late.
Interesting post. I'm no security buff, but whitelisting doesn't sound like it's inheriently a bad thing, and I don't think anyone would argue so, but if that's the route you go, the default should have to be that the user themselves is gatekeeper, with the option of enabling it such that they can use another party to manage their walled garden for themselves.
Except that the root problem with security is the user. If users could be reliably trained to not follow links in phishing emails, to turn off cookies, or not execute random files the find on the metaphorical floor, there would be no need for any kind of whitelist.
Since that hasn't proven to be feasible, any system that lets the user override the system (by for example adding a third party repository to their trusted list) is doomed to fail.
"* E.G.-> It'd be a LOT simpler for say, a home user all the way up to a network administrator on a HUGE corporate WAN to setup a list of PROVEN & fully vetted/code reviewed allowed apps to run (& all the rest would be disallowed...)" - by Anonymous Coward on Monday January 07, @12:17PM (#42506533)
See subject-line, & that quote of myself, vs your misinterpretation of what I wrote (or perhaps you just missed it)...
---
"Except the problem with your whole premise is that you forget the user." - by iamgnat (1015755) on Monday January 07, @12:36PM (#42506753)
Once more - See above...
---
"The problem comes down to who does the vetting and testing of an application to add it to a whitelist? If it is the user, they've proven they can't be trusted because they'll "vet" any new screensaver/antivirus/cursor application that comes along." - by iamgnat (1015755) on Monday January 07, @12:36PM (#42506753)
See above again - in corporate environs, where THE MACHINE IS NOT THE USERS but the companies? That'd be the network admins doing the testing (hopefully).
---
"There is no simple single solution to the problem of security. A real solution by nature needs to be multilayered which means there is some complexity and ultimately users have to take responsibility for their actions." - by iamgnat (1015755) on Monday January 07, @12:36PM (#42506753)
You're "preaching to the choir" here man... seriously, take a look below
(I.E.-> I've been doing security guides based on "layered-security"/"defense-in-depth", especially geared to 'end users' @ home with single systems, since 1997 online & doing pretty well @ it):
To "immunize" a Windows system, I effectively use the principles in "layered security" possibles!
http://www.bing.com/search?q=%22HOW+TO+SECURE+Windows+2000%2FXP%22&go=&form=QBRE
I.E./E.G.-> I have done so since 1997-1998 with the most viewed, highly rated guide online for Windows security there really is which came from the fact I also created the 1st guide for securing Windows, highly rated @ NEOWIN (as far back as 1998-2001) here:
http://www.neowin.net/news/apk-a-to-z-internet-speedup--security-text
& from as far back as 1997 -> http://web.archive.org/web/20020205091023/www.ntcompatible.com/article1.shtml which Neowin above picked up on & rated very highly.
That has evolved more currently, into the MOST viewed & highly rated one there is for years now since 2008 online in the 1st URL link above...
Which has well over 500,000++ views online (actually MORE, but 1 site with 75,000 views of it went offline/out-of-business) & it's been made either:
---
1.) An Essential Guide
2.) 5-5 star rated
3.) A "sticky-pinned" thread
4.) Most viewed in the category it's in (usually security)
5.) Got me PAID by winning a contest @ PCPitStop (quite unexpectedly - I was only posting it for the good of all, & yes, "the Lord works in mysterious ways", it even got me PAID -> http://techtalk.pcpitstop.com/2007/09/04/pc-pitstop-winners/ (see January 2008))
---
Across 15-20 or so sites I posted it on back in 2008... & here is the IMPORTANT part, in some sample testimonials to the "layered security" methodology efficacy:
---
SOME QUOTED TESTIMONIALS TO THE EFFECTIVENESS OF SAID LAYERED SECURITY GUIDE I AUTHORED:
Microsoft decides to suck it up and puts in a simple UAC-like system where the user has to confirm that they want to run a potentially dangerous application that is not signed by Microsoft itself.
Once you can attach a remote debugger to a process you can pretty much run whatever code you want, it's just not user-friendly. The big thing here is that a system process is bypassing sanity checks on API calls (for speed, I assume) and so it's exploited to run arbitrary code in kernel mode, and then you have the whole system (in this case, it just flips the switch to allow any app to run, for the current session only I assume, it won't persist to the next boot).
MS may restrict the processes to which the debugger can attach to fix this, so you can't attach to any system process which uses the faster API calls lacking sanity checks. Assuming there's no way to get other programs to use those versions of DLLs, this would close the exploit, unless the user removes the hotfix (can you do that in RT?) or reinstalls Windows (if that's easy to do).
Either way a tool to package up the remote debugger side of things into something usable would be fairly trivial to make, just gotta capture the network activity of the exploit and then automate it so normal users just push a button and then trigger the proper breakpoint by adjusting the system volume.
See subject-line, & that quote of myself, vs your misinterpretation of what I wrote (or perhaps you just missed it)...
You mean the subject line that is simply showing as "Whitelisting of a sort (& the future of securi"?
As far as possibly misinterpreting you, I will admit your writing style is not as clear as it could be but you clearly go on about whitelisting being most/all of the security solution (to the extent you talk about it possibly replacing AV software). If that was not the point you were trying to get across, then I apologize but that is how it came across to me.
See above again - in corporate environs, where THE MACHINE IS NOT THE USERS but the companies? That'd be the network admins doing the testing (hopefully).
In this case the network admins/company are still the end user even if the result is that they represent more than one physical person. Using the iOS example without buying into Apple's development system there is no "authorized" method for the company to build an internal application and deploy it to their employees iOS devices. So in that case the corporate environment operators still have limited control to do their own vetting. Even still, I've been on the receiving end of "by god you will install this" in the corporate environment so you still have the "users can't be trusted" element there as well.
Whitelisting is the security holy grail, but as with all hardline security measures it forgets that there needs to be a balance between letting the user perform the work that needs to be done while still protecting them from themselves.
I spent some time in a secure environment that tightly controlled what ran on desktops and needed an application that was allowed, but not for the role I was filling. We spent 6 months going back and forth before finally getting approval and getting it installed. Because what I was doing couldn't wait those 6 months we had to work around the restrictions in the meantime. While I am sure of my personal computers, that I had to use them and email the data back and forth opened a vector for a potential problem (as well as violating the corporate rules so you can be sure I had it in writing from a couple levels of management that they approved of what I was doing).
The flip side of course is that I've also worked in environments where everyone had admin rights and could install anything the wanted (though the written policy said they "can only use approved software"). That environment was a constant headache for security and the help desk due to the near constant malware issues (which almost always manifested as performance problems).
Those are the reasons that a rigid whitelisting policy can't work in the real world. Exceptions have to be able to be made in a responsive manner, but that control still needs to be somewhat centralized. In a corporate environment this is relatively easy to do (in principle anyway), but when you start talking about home users that becomes near impossible as there is no way a large company (Apple/Google/MS/...) know what all their users need to do (and really have no business knowing that level of detail in my opinion) and the individual can't really be trusted either. Really the only way to possibly do it would be a community based system, but even there you need some kind of control to keep the likes of 4chan from polluting it by tagging malware as "safe" and Photoshop as "unsafe".
If people are so stupid as to buy a locked device, why make it better for them? Otherwise the might buy and unlocked device next time=better.
I think this article linked through TFA reviewing the WOA appstore sums it up nicely "But for now, x86 compatibility isn't just a check box: It's a doorway back to a land of sanity.". Kinda sad they are actually charging more than iPad for Surface when its quite obvious just from reading the reviews their appstore is completely broken and worthless.
BTW it may be a little petty of me, but since i called it months ago that the WOA and Win 8 appstore would be a trainwreck, since they couldn't make GFWL functional after years and a competitor that would be easy enough to copy they sure as hell wouldn't be able to pull off an appstore for a different arch so I'd like to say "I told you so" to those that doubted me and do the dance of smug superiority.
ACs don't waste your time replying, your posts are never seen by me.
Since when attaching the debugger constitutes a jailbreak?
Not only did you falsely accuse me of NOT covering end users, WHICH I PROVED I DID QUOTING MYSELF IN MY OTHER REPLY TO YOU IN REGARD TO THAT -> http://news.slashdot.org/comments.pl?sid=3364039&cid=42506997
---
NOT ONLY THAT!
You also MISSED that I noted "layered-security"/"defense-in-depth" (as I've been 'into that' for decades now via the proofs of that I put up in extremely successful security guides for end users, the ones that need them most)!
That quote? It's per the termination of my initial post here -> http://news.slashdot.org/comments.pl?sid=3364039&cid=42506533
PROOF again, via this salient pertinent quote of myself from my initial post here:
---
"Whitelisting COULD help stop that too, per what I stated above, along with other "layered-security"/"defense-in-depth" measures commonly used today already." - by Anonymous Coward on Monday January 07, @12:17PM (#42506533)
---
Especially on THAT note?
You are PREACHING TO THE CHOIR on layered security/defense in depth especially to myself, and you also falsely accused me of NOT COVERING END USERS too... in regards to whitelisting!
---
AND, yet you got a +5 upward mod & my posts are downmodded to -1...
APK
P.S.=> Again, as I asked you in another post: EXPLAIN THAT, please...
... apk
Please take your meds. Thanks.
Why was my post downmodded here http://news.slashdot.org/comments.pl?sid=3364039&cid=42507429
?
* Especially since I covered END USERS being 'gatekeeper' as well as network admins in corporate environs in the link above!
(This I gotta see - Why? Well... Just to see if others notice what I do around this website, noted below...)
APK
P.S.=> So, why the downmod of my posts (nearly every one of them since my initial post you replied to)?
Again - I'd love to see YOUR answer to that!
(Not saying YOU did it or anyone I replied to - it's MORE to make a point how BOGUS the moderation system is here @ times, when trolls misuse it)...
... apk
Quit projecting your own "issues" onto others.
AND, yet I get a downmoderation here -> http://news.slashdot.org/comments.pl?sid=3364039&cid=42507627
Purely in response to HOW I use both custom hosts files AND DNS servers (external to my home filtering DNSBL based ones no less, vs. online threats no less, both in my routers AND network connection, layered security again, vs. using a home DNS system (which would be a waste of electricity, cpu cycles, RAM, & other forms of I/O as well as introducing the possibility of failure + DNS redirect poisonings).
"?"
* I'll be waiting...
APK
P.S.=> THIS I JUST GOTTA HEAR - not saying YOU did it, I could care less - it's MORE to point out how BOGUS the moderation system here really is!
E.G.-> So far in this exchange beneath my initial post?
---
1.) I've been falsely accused of 'forgetting end users' in this exchange - which I PROVED, literally quoting myself to do so, that I didn't (& 2 times to 2 different repliers who obviously "skimmed" my reply), both here http://news.slashdot.org/comments.pl?sid=3364039&cid=42506997 and here http://news.slashdot.org/comments.pl?sid=3364039&cid=42507429
AND
2.) I also listed using layered security methods in addition to my noting whitelisting (of which I am an expert for decades now on the former, & proved that too with examples of security guides I've done that did GREAT online since 1997 no less that employ it), then, the poster I replied to 'went off' on it, until I showed him I've been doing THAT for coming up on 2 decades now, implying I didn't note it -> http://news.slashdot.org/comments.pl?sid=3364039&cid=42508567 when I clearly did (again quoting myself to prove I did )
---
Come on - "something's up" here... lol!
... apk
You do realize that sideloading "Modern" (a.k.a. "Metro") applications is fully possible and officially supported, right?
Microsoft states that it "can detect fraudulent use of a developer license on a registered machine." What information is sent back to Microsoft when a developer license is used to allow Microsoft to "detect fraudulent use"?
You can sideload "Metro" applications just fine (after running one command to unlock this capability).
How hard is it to script the periodic commands to renew this capability?
I use a "white list" DNS server written w/ "python" running on my computer. python => smaller than an /etc/hosts file, less memory and i/o than an /etc/hosts file, & NO BUGS (pythons eat bugs lol => true fact!)
BQL
Ms only hardware??? then they will need to make alot more choices then what apple has and lunix will get all the high end systems / good video cards.
Ms may even try to lock in video cards as well.
1st of all - I program in Python... it's LIMITED, compared to my favs. in C++ &/or Delphi. How MUCH so? See next:
Secondly: How the HELL do you figure running Python's RUNTIME ALONE, never mind the code for a DNS server is smaller than hosts?
You STILL have to give it a DNSBL for blocking AND it consumes memory to run period!
---
1.) DNS server, written in python? Yea, ok... sure - see subject-line above & this -> http://stackoverflow.com/questions/981200/can-windows-drivers-be-written-in-python
That means you're NOT RUNNING IN RING 0/RPL 0/kernelmode (fastest there is) because you CANNOT CREATE A KERNELMODE DRIVER WITH PYTHON, first of all!
Which MEANS YOU ARE RUNNING FAR SLOWER than tcpip.sys/IP stack, & less efficiently! Hosts are a TIGHTLY INTEGRATED PART OF the IP stack mind you...
---
2.) How could it also consume less RAM, cpu cycles, & other forms of I/O than a TIGHTLY INTEGRATED PART OF THE IP STACK in hosts files running in kernelmode/rpl 0/ring 0 & RUN FASTER + MORE EFFICIENTLY than the fastest part of the OS with over 40-50 yrs. of optimization put into it?
---
THIS?
* LMAO - This I just gotta hear... lol!
APK
P.S.=> Hosts files, don't crash (DNS servers, do). Hosts files aren't subject to recursive DNS setup attacks either... & more - but I want to hear your "explanation" above, lol! This ought to be good...
... apk
This would be good if they keep their independence from Microsoft and allow these phones to do some good.
Twitter supports and protects racists - by smearing their critics with the "Hate Speech" label.
& upmodded my initial post from -1 to 0 so far -> http://news.slashdot.org/comments.pl?sid=3364039&cid=42506533 & also where I cleared my name vs. false accusations here -> http://news.slashdot.org/comments.pl?sid=3364039&cid=42506997 taking it from -1 to 0 again.
APK
P.S.=> Which proves /. isn't COMPLETELY controlled by trolls (though it is infested all to hell with them, multiple accounts and all - anyone want proof of it? Ask, I've actually GOT it!)...
... apk
Whitelisting will need NO Centership other then apps the crash the system.
and more then 1 app store as well a 100% free zone for both dev's and users.
Per this post of thanks of mine earlier for it -> http://news.slashdot.org/comments.pl?sid=3364039&cid=42510145
* Probably was trolls "playing games" again, what with their multiple registered 'luser' accounts they use to upmod their own posts (Jeremiah Cornelius, webmistressrachel, & the 'dual identity' troll tomhudson/Barbara, not Barbie - as just a FEW I know do it, & how, why, when too (if ANYONE wants proof of that too? Ask, & "ye shall receive"))...
APK
P.S.=> See subject-line above, & thank-you to whoever set those 2 posts of mine right again @ least (many more of mine here are STILL downmodded, but those are a start)...
... apk
http://betanews.com/2012/01/25/the-top-10-web-security-threats-you-should-avoid/
Pertinent quote/excerpt:
---
"The compromised website is still the most effective attack vector for hackers to install malware on your computer with 47.6 percent of all malware installs occurring in that manner, says security firm AVG.
Another 10.6 percent are tricked into downloading exploit code -- many times, without their knowledge -- by clicking on links on pages to sites hosting malware...
It also found that faked pharmacy sites are a popular attack method, seen in about 10.4 percent of all attacks.
Fake antivirus scanners remain a popular malware injection method at 8.4 percent"
---
* Fact is, what I noted, in compromised sites, comprises 77% of malware installations - not what users download & install themselves (ala shareware/freeware sites like download.com etc./et al)...
APK
P.S.=> That's the TOUGH part - the 'driveby download attack' in maliciously scripted sites, bad links to sites, etc.-et al, & why I built this app:
---
APK Hosts File Engine 5.0++ 32/64-bit:
http://start64.com/index.php?option=com_content&view=article&id=5851:apk-hosts-file-engine-64bit-version&catid=26:64bit-security-software&Itemid=74
---
It speeds you up, makes you surf more reliably, but mostly more SAFELY (you can't get burned by what you can't touch)...
It works, to STOP USERS from 'hitting' or rather, being HIT, on known maliciously scripted hosts-domains/servers/sites... & it's updated every 15 minutes here in a 'special build' I use, but users can do it every 12 hours on THEIR end using it.
The rest of what I do?
"Std. layered security/defense-in-depth" ala these guides I've done online since 1997-2008 for Windows users -> http://www.bing.com/search?q=%22HOW+TO+SECURE+Windows+2000%2FXP%22&go=&form=QBRE
HOWEVER: What I personally find VERY EFFECTIVE vs. the 'driveby attack' on sites with malicious code or bogus bad links? In COMBINATION with blocking I do in custom hosts files noted above??
Opera & it's "By Site Preferences"... how so?? Ok:
By ONLY allowing scripting, cookies, plugins, frames/iframes, javascript, java, etc. on SOME sites only that REQUIRE THEM FOR FULL FUNCTION!
The rest are in global policy, disallowing their usage (lessening the chance of attack since those items are OFTEN exploited, massively)...
... apk
No the idea is to get you used to the fact that you can't have anything that's not approved
or otherwise certified. The day will come when you buy a new refrigerator but before you can
turn it on you have to visit a website to register it and accept an energy plan for it. And no...
you won't be using that electricity to grow pot in your garage anymore.
I assume people and companies (governments too) are honest (mostly)... that can be a "problem", for me (or essentially, I figure, anyone who operates honestly period).
Why? Well... simple:
I.E.-> If you don't THINK like a crook? You'll GET "CROOKED"...
APK
P.S.=> It's tough for me to "channel that 'inner-criminal'" here so I don't think about points like yours - HOWEVER/again: That is 1 HELL OF A GOOD POINT you made
(That is, IF things get so 'nuts' that corporations get THAT bad, that is)...
... apk
"I'd like to hear why u think a DNS server is a kernel driver? lol" - by Anonymous Coward on Monday January 07, @05:33PM (#42511265)
Did I state DNS servers are? No. I said the IP stack is (tcpip.sys, & it loads @ bootup + turns on fully once you are in Windows logged on & services start making request on it with client apps too)...
Fast as it gets, written in C & Assembly, and with over 40-50 yrs. of optimization poured into it as well!
---
"Python is written in C so it's as fast as C." - by Anonymous Coward on Monday January 07, @05:33PM (#42511265)
It's INTERPRETED 'code', not a TRUE 'stand-alone' self-contained executable... & there is a LARGE DIFFERENCE in speed right there, for starters!
APK
P.S.=>
"/etc/hosts is a text file so it must be parsed -- very slow!" - by Anonymous Coward on Monday January 07, @05:33PM (#42511265)
Clue - ANY lists you load into DNS? Same thing... loadtime from disk (they are NOT 'automagically' in RAM)...
Secondly:
The app I wrote? Makes hosts files SMALLER & FASTER by using smaller blocking addresses, AND, removing redundant entries (as well as problematic ones with say, CDNs)... & more.
THIRDLY:
Parsing a hosts file in RAM is fast once it's loaded!
Loaded - either by the FAULTY with larger custom hosts files DNS clientside cache in Windows (Linux has no such fault, it is 1 thing I will give it hands-down in fact over Windows), OR, by the kernelmode diskcaching subsystem, as I do it here...
... apk
linux installs and runs with ease on arm-based winrt tablets (e.g. surface).
Accessibility by definition grants extra ability to those who have less or no skill.
There's nothing wrong with accessibility in general. But it incurs a cost. By granting the unskilled extra power that they have no skill in weilding, accessibility makes the landscape that much more dangerous. There are those who are willing and able to become skilled, and through an increase in accessibility, there will be an increase in these people. But this happens at the cost of having to deal with everybody else, who are not interested in such things and willingly accept the collateral damage.
It's like guns (which I mention because it is a popular topic being debated), or vehicles (which I mention because this is Slashdot, and a car analogy is necessary). Accessibility to guns or cars makes the landscape that much more dangerous for the very same reason. For cars, it's getting into accidents. For guns, it's hitting an innocent bystander (or oneself). Accidents are fairly frequent. Bystanders getting shot happens less so, but occasionally when police start firing into a crowded area (individuals are not bystanders when they are the target).
One can argue that in the hands of a skilled operator, a gun would be that much more dangerous for everyone. There certainly is merit to this argument. But that particular argument comes with a presumption of a self-destructive actor, which in all cases would be equally devastating (within context). The same self-destructive actor in a car analogy would drive a car just to ram other people's cars off a cliff road, or a semi just to T-bone a full schoolbus. There are such actors, these events are impossible to stop irrespective of overall population skill (though one can argue that a skilled operator has a small chance of preventing such actions). A better, more effective method of prevention would be to attack the self-destructive impulse itself, and operating skill does not factor into this at all.
I don't support strictly regulating computer usage the way that cars are (albeit loosely) regulated. But I do understand where companies are having trouble finding that middle ground, the spot where they can offer a powerful product to everybody such that anyone with skill can make full use of it, but where they can also limit the unskilled to very specific abilities. Companies are at both the mercy of the consumer who can choose to use a competing product, and of the 3rd party vendor who can choose to develop for a competing product.
Android, I think, comes very close to this, but there are still numerous destructive (within context) things unskilled individuals can inadvertantly do with an Android device. For starters, surrendering habits and other personal information to Google or other companies, and allowing and encouraging this behavior, constitutes dangerous behavior, though possessing a cell phone effectively amounts to having a 24/7 tracking device so the point is moot for many pieces of information.
Windows, stemming from the computing paradigm of yore, did not control the users actions at all, which resulted in the mess that is Windows. Developers and users alike made a mess of the ecosystem. Meanwhile iOS and Windows RT is a bit too controlling, resulting in developers' annoyance for the former and full abandonment for the latter. There are other problems with Windows RT (and Windows 8) from the user interface side. but that I leave for a separate discussion.
(As for gun control, there are certain measures which I support, in particular, focusing the control on bullets and taking the mental health (history) of a gun buyer into account, but there are other measures I oppose on the grounds of being ineffective, draconian, or both. But that is neither here nor there.)
"If a nation expects to be ignorant and free in a state of civilization, it expects what never was and never will be."
Still interpreted code slow, no getting around that - you use a Python created DNS server>
That's just the way it goes for you... period!
* You FAIL #1 of 2 (see next below on your 'point')...
APK
P.S.=> On hash tables - ok: Even binary search patterns & indexing CANNOT beat my placing my top 20 favorites into the host @ THE TOP OF THE FILE!
That's what your hash table does, & I beat it with SMART PLACEMENT, lol...
(Try it, do the math over 2 million entries, since that's what my hosts file has - the way I do it noted above? Well - I actually come out FASTER for my favorites top 20 @ the TOP OF MY HOSTS FILE once loaded into the kernelmode diskcaching subsystem!)
The rest of what's in my hosts file? Blocked - I could care LESS about 'getting to them' since they are known malicious & BLOCKED - (you "FAIL" yet again!)...
... apk
I agree that it is illegal to distribute Android without agreeing to the GPL. The difference between Android and GNU/Linux lies in how much code is affected by the GPL. Unlike GNU/Linux, Android limits the extent of GPL covered code to the kernel, so any changes to user space need not be distributed to the public as source code.
Ok - What's @ the top of my hosts file? My 20 favs.!!!
(Which beat up to 2++ million INDEXED by hash table entries)...
Please - Do the math!
You'll see EXACTLY what I mean, mathematically (even in a binary seek, which is generally easier to use than hash tables, is sorted, & faster usually too).
Hash Tables: Which are a BITCH to get right for diff. kinds of datasets too mind you... they're touchy shit!
(Binary trees & seeks are better & take less RAM to implement, & I still beat those too, up to 2++ million entries by putting my favs I spend a GOOD 95% of my time online at, which my hosts NEARLY has... very close now!)).
* You FAIL again...
Why?
I rarely DO DNS lookups!
I figured out WHERE I spend my time online is why/how, & put them @ the top of my hosts file (loaded into RAM via the kernelmode diskcaching subsystem).
HOWEVER - When I do DNS lookups?
They're from filtered vs. malicious threats FAST online DNS servers I noted earlier...
Double-win!
Lastly/Again: You use a DNS server WRITTEN IN PYTHON for Pete's sake - SLOW INTERPRETED CODE!!!
vs.
Tcpip.sys which uses hosts, in ring 0/rpl 0/kernelmode + the diskcaching kernelmode subsystem in RAM, housing my hosts once loaded!
APK
P.S.=> You STILL don't "get it" do you? Usermode stuff is way, Way, WAY slower than kernelmode, & always WILL be - I use smart placements to beat even hashtable indexing OR b-tree seeks, because I rarely DO dns lookups (since I know where I spend my time online, mostly here in fact)...
... apk
Android caters to manufacturers that don't like GPL compliance by minimizing the amount of what they have to do that they don't like. Say a manufacturer has two relevant options: either A. do what it doesn't like to a small extent, or B. do what it doesn't like to a large extent. The manufacturer is more likely to tolerate doing A as what it considers the lesser of two evils. In this case, Android is A and a GNU stack is B. With Android, they have to distribute source code only for changes to the kernel. With GNU/Linux, they would have had to distribute source code for all modifications, including the launcher modifications that manufacturers use to distinguish their products from those of other manufacturers.
"I use a "white list" DNS server written w/ "python" running on my computer. python => smaller than an /etc/hosts file, less memory and i/o than an /etc/hosts file, & NO BUGS (pythons eat bugs lol => true fact!)" - BQL by Anonymous Coward on Monday January 07, @03:34PM (#42509307)
1.) You use Python interpreted SLOW code (many orders of magnitude slower than a single self-contained TRUE 'stand-alone' executable, let alone DRIVERS (tcpip.sys & the local kernelmode diskcaching subsystem, which power mine - far, Far, FAR FASTER... tons so!)
2.) You use USERMODE/ring 3/rpl 3 SLOW code (vs. tcpip.sys + the local kernelmode diskcaching subsystem) since PYTHON CAN'T DO KERNELMODE DRIVERS (many orders of magnitude faster than usermode, & certainly even moreso than your INTERPRETED SLOW python code).
3.) My 20 favorites @ THE VERY TOP OF MY HOSTS file outperforms hashtable indexes (up to 2++ million entries). You don't deny this - the math proves me right!
---
A.) You also CLAIM a 'whitelist' DNS list - what about blacklist for protection vs. KNOWN online threats? You WILL perform DNS lookups eventually, you know, and in recursive mode (more than potentially hazardous)!
B.) Your DNS locally IS in recursive mode, opening up the door to DNS-poisoning redirection too!
C.) Even I perform remote DNS lookups, albeit only perhaps 1 in every 100 lookups since I know where I go online & spend 95%++ of my time, & perform OCCASIONAL dns lookup queries, albeit, to FAST remote DNS servers NOT WRITTEN in SLOW INTERPRETED CODE, & that also filter vs. online malicious threats - yours doesn't apparently: YOU'RE LEAVING THE BARN DOOR OPEN TO "LET THE TRASH COME BLOWING IN"...
---
Here's what custom hosts files can do, for SPEED alone (from your own /. peers, & a security-expert from a division of SYMANTEC as well):
(Especially on the note of loads & parsing (which you claim is 'slow'))
---
SLASHDOT USERS EXPERIENCING MORE SPEED USING HOSTS FILES QUOTED VERBATIM:
---
"I want my surfing speed back so I block EVERY fucking ad. i.e. http://someonewhocares.org/hosts/ and http://winhelp2002.mvps.org/hosts.htm FTW" - by UnknownSoldier (67820) on Tuesday December 13, @12:04PM (#38356782)
"this is not a troll, which hosts file source you recommend nowadays? it's a really handy method for speeding up web and it works." - by gl4ss (559668) on Thursday March 22, @08:07PM (#39446525)
"I actually went and downloaded a 16k line hosts file and started using that after seeing that post, you know just for trying it out. some sites load up faster." - by gl4ss (559668) on Thursday November 17, @11:20AM (#38086752)
"I'm currently only using my hosts file to block pheedo ads from showing up in my RSS feeds and causing them to take forever to load. Regardless of its original intent, it's still a valid tool, when used judiciously." - by Bill Dog (726542) on Monday April 25, @02:16AM (#35927050)
"I have several notorious slow adservers in my /etc/hosts" - by jandrese (485) on Friday August 17 2007, @01:00PM (#20263547)
"They're visually annoying and distracting. They're a waste of bandwidth. Sometimes they're even noisy. I block them with a hosts file" - by Kris_J (10111) on Monday October 10 2005, @11:12PM (#13761572)
"Am I the only one that uses a hosts file? Takes care of more than just ads. It's to the point now that when I see ads, I'm shocked. I've had them blocked for years. They may be able to stop adblock, but good luck trying to outlaw a hosts file." - by mrbcs (737902) on Friday November 23, @06:59PM (#42077997)
"127.0.0.1's in my hosts file. Some shady ads do cause trouble, and similar methods can be used
"I use a "white list" DNS server written w/ "python" running on my computer. python => smaller than an /etc/hosts file, less memory and i/o than an /etc/hosts file, & NO BUGS (pythons eat bugs lol => true fact!)" - BQL by Anonymous Coward on Monday January 07, @03:34PM (#42509307)
What about BLACKLISTING PROTECTION vs. AdBanners (+ known malicious sites-servers/hosts-domains & botnet C&C servers + bogus DNS servers as well as trackers/spammers/phishers) filtering for security on your 'whitelist' DNS slow Python usermode interpreted code (ontop of speed savings this all yields)?
Adbanners (& other items noted above), that:
1.) SLOW DOWN YOUR WEBSURFING
2.) INTRODUCE MALICIOUS CODE (many evidences thereof next below)
3.) ROB YOUR BANDWIDTH YOU PAY FOR?
4.) TRACK YOU
& more...?
You don't GAIN SPEED, PRIVACY, or SECURITY THAT WAY EITHER! I do, by using a custom filtering hosts that is BOTH a 'blacklist' & 'whitelist'... lol!
Even Spybot "Search & Destroy" IMMUNIZE feature (fortifies HOSTS files with KNOWN bad servers blocked) uses this feature of hosts (it is a HIGHLY RESPECTED antispyware software).
Additionally - SINCE MALWARE GENERALLY HAS TO OPERATE ON WHAT YOU YOURSELF CAN DO (running as limited class/least privlege user, hopefully, OR even as ADMIN/ROOT/SUPERUSER)? HOSTS "LOCK IN" malware too, vs. communicating "back to mama" for orders (provided they have name servers + C&C botnet servers listed in them, blocked off in your HOSTS that is) - you might think they use a hardcoded IP, which IS possible, but generally they do not & RECYCLE domain/host names they own (such as has been seen with the RBN (Russian Business Network) lately though it was considered "dead", other malwares are using its domains/hostnames now, & this? This stops that cold, too - Bonus!)...
---
Adbanners slow you down & consume your bandwidth YOU pay for:
ADBANNERS SLOW DOWN THE WEB: -> http://tech.slashdot.org/article.pl?sid=09/11/30/166218
---
And people do NOT LIKE ads on the web:
PEOPLE DISLIKE ADBANNERS: http://yro.slashdot.org/yro/08/04/02/0058247.shtml
---
As well as this:
Users Know Advertisers Watch Them, and Hate It:
http://yro.slashdot.org/yro/08/04/02/0058247.shtml
---
Even WORSE still, is this:
Advertising Network Caught History Stealing:
http://yro.slashdot.org/story/11/07/22/156225/Advertising-Network-Caught-History-Stealing
---
PLUS:
---
THE NEXT AD YOU CLICK MAY BE A VIRUS:
http://it.slashdot.org/story/09/06/15/2056219/The-Next-Ad-You-Click-May-Be-a-Virus
---
Yahoo, Microsoft's Bing display toxic ads:
http://www.theregister.co.uk/2011/09/16/bing_yahoo_malware_ads/
---
Malware torrent delivered over Google, Yahoo! ad services:
http://www.theregister.co.uk/2009/09/24/malware_ads_google_yahoo/
---
Rogue ads infiltrate Expedia and Rhapsody:
http://www.theregister.co.uk/2008/01/30/excite_and_rhapsody_rogue_ads/
---
Google sponsored links caught punting malware:
http://www.theregister.co.uk/20
On both SPEED & SECURITY: Speed -> http://news.slashdot.org/comments.pl?sid=3364039&cid=42516867 and Security -> http://news.slashdot.org/comments.pl?sid=3364039&cid=42516923
* Reply to each point in each post...
(With your "rebuttals" & good luck - you'll NEED it!)
(You're OUTNUMBERED, OUT-THOUGHT, & just plain 'outgunned' by many, Many, MANY orders-of-magnitude on BOTH speed & security - From myself, your /. peers' findings, security experts, + just plain FACTS!)
APK
P.S.=> I look forward to it actually - as there is NOTHING QUITE LIKE "busting up a troll" with facts - especially one doing ONLY a WHITELIST using usermode SLOW code (vs. drivers in kernelmode) that is INTERPRETED SLOW in Python on top of it (as far as performance) - for security? A whitelist doesn't cut it alone, not until the topic of my initial post "kicks in" & I don't think it ever will on PC desktops (users LIKE control + personalization, not lack of it)...
... apk
You also understand HOW EASY/SIMPLE it is to have multiple accounts here to mod others down, and yet not remove moderations from "other alternate registered 'luser'" guises too (as well as modding up multiple accounts)...
THAT'S ALL I WANTED TO SEE - THAT OTHERS HERE KNOW HOW IT WORKS THAT WAY... why?
---
I've literally CAUGHT, red-handed, an entire PACK of trolls around here that hangs out, literally, @ a domain they own called "TrollTalk.com" (not kidding on that either). They are:
1.) gmhowell
2.) Jeremiah Cornelius
3.) webmistressrachel
4.) countertrolling
5.) PLUS lastly - the WORST OF THE LOT & STUPIDEST -> tomhudson alias Barbara, not Barbie (whom I ran out of here since I caught he/she doing that very nefarious thing).
* I've had others attempt to "impersonate me" here as well... but I am 'long-winded', like the post of mine you replied to - but, I did in fact, CONSIDER USERS (though you thought I didn't).
---
On HOW I POST (long-winded etc./et al):
Sometimes, there is NO 'short & sweet' way of putting a point out, so I have to go into detail, provide examples, & go on 'tangents' as you say ( but I usually ALWAYS "circle back" to where I was & my main point I was attempting to make IF you read my entire posts...).
It's just "my style"...
---
Oh, lastly - if you DOUBT any of the above (on the 'trolltalk.com' crew I noted above, multiple account using trolls? Don't - I can LITERALLY provide proofs of them doing so AND 'gaming' (cheating) the moderation system here in fact...).
APK
P.S.=> Anyhow/anyways - Above ALL else: Thank-You for replying & thanks for your time + understanding of "how it really works" around here, & the BOGUS broken so-called "moderation system"...
... apk
It worked - I was getting REGULAR frequent upmods (like I used to before trolls around here began 'targetting' me)... you're correct, but "been there/done that" too!
E.G./I.E.-> Posting as "pure AC" (minus the 'APK' @ the termination of my posts as a signature of sorts, & the ellipses ... apk on my subject-lines also).
This life, imo @ least? IS a matter of style... it's ALL about style in fact. If folks want to be 'cookie-cutter' clones of 1 another? Fine... it's just NOT for me is all!
* It seriously "boggles my mind" that people act like that online, because I sincerely & seriously DOUBT they would in 'the real world'... why?
It'd get you a PUNCH IN THE JAW most likely!
APK
P.S.=> Again thanks for your time & understanding... apk
I hate to be that guy, this being slashdot and everything... ... but does this mean that it might be able to run Linux?
If I knew that I could get a tiling wm, a term, and a browser running on Linux on one of these, with a minimal amount of fucking around with it, I'd go out and buy one this instant.