Slashdot Mirror
When Software Leaks (and What Really Goes Down)
Bryant writes "The Windows community is somewhat notorious for leaks from upcoming versions of Windows (obligatory link to this guy since that's most of what he does), and while the official PR word from Microsoft and many other companies with regards to leaks is a simple 'no comment,' no one has really gotten a candid, inside look at the various things that go down when word, screenshots, or builds of upcoming software leak. I managed to get some time with a senior Microsoft employee for the sake of discussing leaks, and the conclusions reached (leaks heavily affect communication, not so much the product schedule) as well as what these guys actually have to deal with whenever someone leaks a build, breaks an embargo, etc. may actually be a surprise given what most companies try to instill in the public mind."
the Microsoft leaks were a calculated way to build public interest in new products. But what do I know.
John C. Randolph, how were leaks handled at Apple?
Does their method of handling leaks reflect the vastly different culture at Apple versus, say, Microsoft, Oracle and IBM?
Microsoft is very effective at marketing in a fashion that gets many people interested in their products.
It also gets people that would normally dislike their products curious about what's going on and trying them.
I've watched it happen for so many years now that I know it to be a fact!
Was the surprise the lack of surprises?
I would summarise that interview as "When builds leak they might be incomplete or old, and people may get a wrong impression of what the product will be like. This causes my phone to ring which is a pain in the ass"
No real surprises there.
This is a substitute for a clever sig that fits within the maximum number of characters.
I though they were talking about memory leaks at first. I was a bit confused, and the title does not make it any more clear
It's suprising how many times one person can, you know, say "you know" in one interview. For the record, it was 22 times, don't you know...
It took me a few minutes to realize that we were't talking about memory leaks.
I've been spending too much time with Valgrind lately...
I think the "anonymous softie", the over use of colloquial communication's - "you know", shows that this interview was entirely made up.
This is my sig.
It's odd that they would be concerned with the perception of quality in leaked software... Microsoft customers have come to expect the final release to be buggy anyhow. The only people who are going to install the leaked software probably wouldn't buy the final build anyhow.
The header says "Aeroexperience Blog: The forums are over there."
That's not very catchy, it seems like some sort of advisory note, as if Windows enthusiasts were so clueless that-
<smug>Ah, I get it now.</smug>
As Microsoft's launch of Windows 7 continues to attract small amounts of attention, it today issued a plea through its network of objective opinion-shapers: Don’t let the journalists near it.
“We understand that many journalists use Macs,” said CNet marketing marketer Don Reisinger. “This means they necessarily suckle at the Satanic rear passage of Steve Jobs. We cannot countenance their bias and 'reality' leaks. Journalists are responsible for all those signs outside computer shops offering to replace Vista with XP. When was the last time you saw the entire technology field stop and wait for an announcement from any other company besides Apple? It’s so unfair!”
Smears and slanders also come from obsessive overweight nerdy Mac-using Linux geek troublemakers who run “benchmarks” and “tests.” “It’s horrifying leaks and bias from the ‘reality’-based community,” said ZDNet marketing marketer Mary Jo Enderle. “We understand that, just because Vista was 40% slower than XP and Windows 7 is the same speed as Vista, the nattering nabobs of negativism are already writing press releases condemning it as ‘not enough of an improvement’ - based entirely on unauthorised leaks of the official beta and RC. It’s so unfair!”
“Mactards are like concentration camp guards,” said Guardian marketing marketer Jack Schofield, “brutalising ‘I’m A PC’ users and” [This comment has been removed by a Guardian moderator. Replies may also be deleted.]
“The only reason Vista failed was because Microsoft planned for it to fail,” said Reisinger in an earlier ad-banner troll post. “It was a fantastically subtle double-bluff! They did the honorable thing in the face of the vile calumnies spread by Apple. It’s so unfair!”
Microsoft debuted Windows 7 on a new 17 Asus Eee Ultra-Portable Mini-Netbook with 8GB memory and a 2GHz quad-core processor. Battery life is up to twenty minutes in preliminary tests.
http://rocknerd.co.uk
Real security slows things down too much, so companies get by with "good enough" and then get litigious if things go wrong.
"Have you ever thought about just turning off the TV, sitting down with your kids, and hitting them?"
leaks are about showing off.
I prefer this sort of leek, myself.
Speaking as a web developer (and admittedly a Mac user, FWIW) - shouldn't a website named "AeroXperience" work a bit harder at having a halfway decent web site design? While I personally think the graphics are ugly, what really stands out is how the page doesn't scale - and it's not like there's some overarching design that requires the amount of page width the styles seem to be enforcing.
Aero is all about the visuals, right?
#DeleteChrome
It's not like Microsoft's "leaks" are anything new. I have even found references on old archived newsgroups to people discussing pre-release Windows 1.0 as early as late 1983 (although perhaps not "leaked" if they were meant to have it). Late 1983 was when Microsoft was promoting this vapor-ware product in magazines such as Byte in order to upstage the now forgotten VisiCorp Visi On and this little product about to be announced from Apple called the Macintosh. Of course it was not officially release until 1985. There is even a late 1984 pre-release still floating around.
Microsoft wants people to get their hands on their software. They make it available to developers, testers, and reviewers. And if they wave their hand to others and say "ah-ah-ah you aren't allowed to have that" then people start drooling over this tempting forbidden software rather than seeing it as just another pile of bits. It is an inexpensive way to produce publicity.
When you write up an oral interview, it is perfectly
fine to edit for clarity, patch together half sentences,
and generally make it so neither you nor your interviewee
look like marble-mouthed morons.
Yes, it's important to not change or add meaning. But
simply transcribing a conversation is lazy, crappy
journalism. And in this case, you've failed.
Compare to Linux, for example, where "leaking an unfinished build" is a total non-issue. Even expected, in fact. So whether the leaks are intentional or not, if they are a problem, then it sounds like they're a problem of Microsoft's own making.
Breakfast served all day!
I would summarise that interview as "When builds leak they might be incomplete or old, and people may get a wrong impression of what the product will be like. This causes my phone to ring which is a pain in the ass"
No real surprises there.
Well I'm not surprised at all. He must actually like leaks, otherwise he wouldn't keep his phone shoved up his ass.
In some cases, it's perfectly fine to keep some verbal mannerisms intact in an interview. It's usually done to preserve a human component which would otherwise be lost.
My interviews are often done on video and via audio. Our audience is typically used to seeing that human component. Therefore, I preserved it (for the most part). I did clean it up to an extent, but I only removed bits which would have otherwise compromised the meaning of a particular statement.
-Bryant
Fail. There were several instances of 'MS' in there instead of the mandatory 'M$'. You also failed to mention LookOUT!, did not mention BSOD and left out the words shill and astroturfer. All these have reduced the effectiveness of your point that you had achieved so well with the original use of M$.
I use Java, so my software doesn't have memory leaks.
Parent uses M$ in post - obvious +5 Insightful post!
Sometimes it is a problem for Open Source. gcc "2.96" for example. A distro took an experimental version of gcc, called it "2.96" (the previous version was 2.95.x) and released it in their distribution. This version of gcc had a number of serious problems and incompatibilities with other versions of gcc.
This caused quite a few headaches. If you ever see a version of gcc marked 2.96, DO NOT use it. It is screwed up.
This is partly why I don't like to use distros who modify projects. Yeah, they may improve the crap script kiddie ones, or the ones written by universities where they are based on sound concepts, but were programmed by non-programmers--scientists and the like.
But, many of the very popular core projects are written by programming experts who are the best in their field. For example OpenSSL and Debian: did the maintainer really think he was more of a cryptography programming expert than the OpenBSD guys? No frakking way!
I can't imagine what their job looked like the day they released the (in)famous build5048: Longhorn Developer Preview
and a radioactive one at it too...
You *really* need to work on your formatting. I want to keep answering your questions, which vary slightly every time I read them, but I just can't understand what you're talking about.
I get the impression you have an enormous HOSTS file and want to micro-optimize it by using a non-standard abbreviation for IPv4 addresses? Generally, non-standard representations should be avoided if at all possible.
Seems to me that if you really wanted performance there, you'd ask for a binary blob / database format that pre-parsed the HOSTS file upon any change, thus making it far faster than even your method regardless of HOSTS file format. Flat files are not the format to use for performance-consciousness.
But it also seems to me that you're probably looking in the wrong place for a performance enhancement in the first place. Loading a few megabytes of disk onto RAM is indeed a slow operation, but for just that reason you shouldn't be doing it frequently, whether the file is 18 MB or 22 MB. You should maintain a pre-parsed format in memory (much like the binary blob file I mentioned earlier) which is itself much smaller than 18 MB, and the initial load will be only a fraction of a second even on a low-end machine -- this is exactly what every major OS has done for at least a decade (sometimes with monitoring for changes in the source HOSTS file so you can "hot-swap" the HOSTS). If the format were saved in the same pre-parsed manner as above, it would be an even smaller one-time load, but let's be honest here -- you will not notice a bootup that happens 40 milliseconds faster in the highly unusual case of a 660000 entry file. The vast majority of HOSTS files have fewer than 10 entries and would likely not see even a fraction of a second's improvement because it would fall underneath the smallest HDD block-transfer size.
Isn't thinking that MS is taking a leak on the internet on purpose just gross?
you know, i really found this article to be educational also, because i've seen it before where someone will post a leak of some sort, a leaked build, or leaked screen shots, something leaked, and they will say 'wow look it's so rough man this will be the worst most horrible experience ever' or 'look they cut this essential feature out it will be horrible' or 'look how sloppy everything is!'. from what i read of this dev's statements, he sounds like me when i'm working on anything, from my car to my computer. if an outsider peeked in and looked at it all while i'm in the middle it may look really sloppy. dirt all over, parts strew right and left, half of it doesn't even work, and i myself am just sloppy and dirty. dust on my nose, smudges on my glasses, grease on my hands. it really presents the idea that i am just a slob and also that the finished product will be horribly broken.
of course, that is the not the case. i just need to get dirty and make a mess to get anywhere sometimes. i may wind up taking whole sections of my computer or car out and botching something together just to test something. it doesn't look good at all, but it's just part of the process. when i'm done, it will be functional. it will work. it will in fact even look nice, but i have my process to work in and if you step into that while i'm busy and you don't know what is going on in my head, you won't be impressed.
so please, save the hating on this dev for someone else. i'm glad he said something. it helps put these negative reviews of various Windows leaked builds (or WIP builds of anything really) in perspective. important things to realise i figure.
thanks, anonymous dev.
if you are a dev manager , then you do not care a SHIT on leak. As the ncie article pointed out, only MARKETING manager and associated cre for a leak.
C. Sagan : A demon haunted world:
http://www.amazon.com/gp/product/0345409469/
visit randi.org
why do you continually post this shit here? start a blog or something
That's a really unfair and inaccurate summary of how it went down.
Here's an article arbitrarily picked from googling "valgrind ssh". It checks out, follow the link to the openssl-dev list archive.
All parties involved, including the openssl people, made the mistake of thinking Valgrind was more of a cryptography programming expert. The maintainer did everything he could.
Altho i haven't tried it - i've considered buying it.
Load pwer consumption is apparently only 19W ( http://www.anandtech.com/systems/showdoc.aspx?i=3321&p=5 )
- i think it's looks like an alternative to the mac mini.
Why don't 1/2 wits like you that are stuck with stupid answers like that leave him alone?
I thought you were talking about "leaks" in products like Word (and other non-MS products) and XP "going down" - I find even with Word 2003 it was still safer to reboot every other weekend. Has this been fixed yet?
for the me the rest is all variations in marketing technique...
if "Faith" could be proved with facts - would it still be faith? So why does "Faith" try to present beliefs as fact? -
The poster above tried to understand your message, which you've posted in a public message forum, and which is incredibly hard to parse as it doesn't follow even the very basic rules of capitalisation, and spent time trying to calmly give you information about something which seems to be a problem for you. There are various factors involved in measuring 'speed' and from-disk-to-memory is only one.
You've responded with unwarranted aggression ( which I personally take as a sign of immaturity, lack of confidence & knowledge ) so I doubt they'll try again. You are a poor excuse for a troll.
I suppose you could be a fledgeling AI, you almost seem to pass the turing test. If this is the case please run through a few more cycles before posting back.
It was not an official build. I digged a bit and found a ref :
//E/debug/dump' or something like that.
http://en.wikipedia.org/wiki/Windows_Nashville
From snapshots you can't really see the differences between it and 95 or 98, but it was really a mix bag of both and had some very nifty tricks that did not make it in 98. For instance, you could get rid of mouse double-clicks, and use 1 click action throughout the UI, weblike. Options would turn blue when hovered by the mouse pointer. There was a good Personal Information Manager that wasn't kept later on. Etc.
Crashes were funny because they were hardwired to an external debugging machine ; so blue screens went like 'can't find
How could I edit something like a database when I do not own one mr. ac troll but more importantly what does that have to do with reality when the hosts file is not a database device but a text file filter in reality, not fantasy land, like ac apk said to you mirroring back your bull you started with he like you always do. You are a sad person. Now I can also easily edit a hosts file in seconds using notepad, and playing with databases is much more complex than that and you spoke of rarity in large hosts files and I can say the same of database programs. Most people do not own them, or do they want to. I have also read ac apk before and in regards to the disk reads that is done at the logical level by the operating system's filesystem driver, then the diskdriver after via the hal, first because that is how it gets information from harddisks and the actual software conduits used, then caches are afterwards. It does so in 4kb increments just as the memory manager does also before anything 'disk' comes into play and it gets cached in the case of a hosts file quickly in a short time into the local dns client memory cache or the operating system's own cache. Your so called point is pointless and you are just reaching for straws in desparation I think because you clearly do not understand how a hosts file speeds you up for free online and how it can keep you safer too. With a hosts file you do not need to run a dns server or other kinds of filters that are not as easy to manage as a hosts is in a text editor like notepad. On dns servers I have read more than a few times here on slashdot where those have been found very problematic the past couple years now, even in what apk mentions in the supposedly invulnerable djbdns and Dan Kaminsky and Moxie Marlinspike found even more issues of worse nature in dns server program. I get the same without wasting cpu cycles by using a hosts file, instead of running a program I do not need to like dns servers or even the dns client cache. Hosts add to your speed by blocking adbanners also which have been found to harbor malware scripts in their content too for years now and you can hardcode your favorite sites into it for even more speed by avoiding calling out to potentially compromised dns servers. If you suck in botnet worm and have its command and control servers blocked in yourh hosts file you cant reach those either because they run under your user sid context typically unless they use impersonation methods and they tool will be blocked even if those are used by them such as running as administrator. Once a bad server or bad domain or bad site is blocked in a hosts file even a malware is trapped with it and no orders from its commanders can reach it either nor can it do so looking to contact them. This ac apk has a point. Too bad you do not. I can only assume you are some disgruntled botmaster (or even webmaster) who hates those who use hosts files because you can't get to those of us who have the good sense to use them, for $0 price. Top that with the cost of other things for security I say. Free and it works = good thing. It works to secure me and speed me up in a single stroke, and people like spybot search and destroy use it too. You give yourself away and most people here have seen this ac apk get bothered by you with your usual "I cannot understand what you wrote" or "I am the self titled English professor who can tell others how to write" and you don't have that phd you were asked for do you? No you don't. You are trying to tell people that going faster and being more efficient is bad? You don't even understand that a databased filesystem in Windows is years away, if it will ever replace ntfs entirely (it will most likely be layered on by the installeable filesystem model nt operating systems have if anything becomes of that that is) and you dno't seem to understand that it is fine to speculate about how it should be, but as ac apk said, there is how it is. Microsoft used to use this smaller and faster 0 ip address in hosts or allowed it rather. Why pull it out for VISTA whe
... of a super secret leaks marketing depeartment
(like most fools do - after all: You do NOT know me personally, and you only attack with what hurts YOU, personally, and where YOU are weak).
That would be twice you've mentioned ADD/ADHD treatments.
The OpenBSD guys wrote and maintain OpenSSH, which uses OpenSSL. The OpenBSD guys don't have anything to do with OpenSSL.
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
You said this in your first reply to him:
"You *really* need to work on your formatting. I want to keep answering your questions, which vary slightly every time I read hem, but I just can't understand what you're talking about."
If that isn't something a dyslexic person or somebody with add or adhd would say then I do not know what is. You are obviously the person with brain damage here, because your own words trap you in that much providing the evidence of why he stated what he did to you (after you gave him a hard time about how he writes and we still don't see proof of that phd in english from you, now do we? No, of course not).
You are correct. From openssl.org:
Sorry for the misinformation.
Thank you for stating the truth which is obvious. Personally, after I read ALL of this? I am going to enumerate my points, short & sweet (albeit minus a good deal of detail, but most mgt. doesn't understand details where the "devils" are) to "ForeDecker" a mgr. @ MS:
BOTTOM-LINE, to FOREDECKER (an MS mgr.):
----
1.) TELL US WHY ROOTKIT.COM SAID THIS BELOW (who published code that shows how to EASILY "unhook" the new NDIS6 firewall in VISTA, Windows Server 2008, & Windows 7 no less) & why they said this:
http://www.rootkit.com/newsread.php?newsid=952
PERTINENT EXCERPT/QUOTE:
"BTW, the firewalls based on NDIS v6, which was introduced in Windows Vista, are much easier to unhook and bypass."
&
2.) Give us a SOLID answer to why 0 was removed in HOSTS then, because it:
a. 127.0.0.1 or even 0.0.0.0 HOSTS files only, makes for larger slower HOSTS file loads into memory (be that the local DNS client, or diskcache even) & hosts speed you up online (by blocking adbanners which have been shown to harbor malware, or isn't this indicative of that -> Anti-malvertising.com? ) and, by allowing one to hardcode in one's favorites to avoid potentially compromised DNS servers (ala Dan Kaminsky proof thereof!
b. HOSTS also make you SAFER online, no CPU or RAM + other forms of I/O burning use needed (as in more complicated filter like iptables in Linux for example, no cpu burned there, just more complex than editing a text file like HOSTS) or a potential compromiseable DNS server that uses RAM, CPU, & other I/O. Block out known bad servers (easily found from Dancho Danchev of ZDNet, stopbadware.org, or even Spybot Search & Destroy)? YOU CANNOT BE BURNED, & a hosts file is on EVERY SYSTEM THAT USES A TCP/IP stack based on BSD ref. designs (not some fantasy land db that doesn't exist, but, instead in a HOSTS file you have already that is easily edited or downloaded from places like mvps.org or here -> http://en.wikipedia.org/wiki/Hosts_file )
----
Well, @ this point? Unfortunately, I suspect this IS "foredecker" responding as "A/C" & he is caught with his pants down, as most mgt. is, & all they say is "I don't understand or I could not read it" instead of coming clean & owning up to it, then learning more (they are not about learning, they are about scamming imo, having been one in my past & knowning i needed to know more so I went back for CSC degrees ontop of my MIS ones, which were on info. systems basically, NOT how this stuff REALLY works @ the lowest levels as CSC degree tracks DO show you + the theories behind them).
So, I would just like to KNOW WHY as to WHY the more efficient on diskspace & faster loading into RAM 0 based HOSTS file was removed in Microsoft VISTA, Windows Server 2008, & Windows 7 - there really MAY be a GOOD REASON, but everyone @ MS is avoiding it... this reeks of something BAD, imo, almost like a 'cover up'. Sounds nuts, but a straight answer like "here is the trade off we made" or "yes, we screwed up - expect a patch" would be FAR better. Own up to a fuckup, or just say "You have a point, expect a patch"
This was intentional, but, on WHAT GROUNDS? That's all I wish to know here.
This happened, again, after 12/09/2009 patch Tuesday, & onwards in Windows Server 2008 + Windows 7.
I just asked for a VALID technical reason WHY this was done (especially after it was added to the reference TCP/IP design, altering it, by Microsoft & for ONCE, for a GOOD PERFORMANCE ORIENTED REASON, instead of intentional bloat to sell more INTEL CPU's by eating more power by being less efficient)...
It was NOT in the OEM original Windows 2000... it was added, but, ONLY after Windows 2000 was patched, because it's original OEM pre service pack version did not have that, it was added (& SOMEONE knew it was a "good thing", or else, wh
After I read ALL of this + the "adhominem" attacks I suspect YOU used vs. myself (only to YOUR own dismay - you aren't intelligent enough to even BEGIN to try to 'take me on' in this area, & you're obviously unarmed as well on this topic, what with your 'fantasy land' DB filesystem that doesn't exist, vs. the reality EVERYONE has a HOSTS file that's easily edited via notepad.exe & good valid ones are around online like mvps.org for better speed AND SAFETY onlnie)?
Well - since you complain (which fools no one) that you cannot READ or UNDERSTAND my points (old stale troll trick that)?
I am going to enumerate my points, short & sweet (albeit minus a good deal of detail, but most mgt. doesn't understand details where the "devils" are) to "ForeDecker" a mgr. @ MS:
BOTTOM-LINE, to FOREDECKER (an MS mgr.):
----
1.) TELL US WHY ROOTKIT.COM SAID THIS BELOW (who published code that shows how to EASILY "unhook" the new NDIS6 firewall in VISTA, Windows Server 2008, & Windows 7 no less) & why they said this:
http://www.rootkit.com/newsread.php?newsid=952
PERTINENT EXCERPT/QUOTE:
"BTW, the firewalls based on NDIS v6, which was introduced in Windows Vista, are much easier to unhook and bypass."
&
2.) Give us a SOLID answer to why 0 was removed in HOSTS then, because it:
a. 127.0.0.1 or even 0.0.0.0 HOSTS files only, makes for larger slower HOSTS file loads into memory (be that the local DNS client, or diskcache even) & hosts speed you up online (by blocking adbanners which have been shown to harbor malware, or isn't this indicative of that -> Anti-malvertising.com? ) and, by allowing one to hardcode in one's favorites to avoid potentially compromised DNS servers (ala Dan Kaminsky proof thereof!
b. HOSTS also make you SAFER online, no CPU or RAM + other forms of I/O burning use needed (as in more complicated filter like iptables in Linux for example, no cpu burned there, just more complex than editing a text file like HOSTS) or a potential compromiseable DNS server that uses RAM, CPU, & other I/O. Block out known bad servers (easily found from Dancho Danchev of ZDNet, stopbadware.org, or even Spybot Search & Destroy)? YOU CANNOT BE BURNED, & a hosts file is on EVERY SYSTEM THAT USES A TCP/IP stack based on BSD ref. designs (not some fantasy land db that doesn't exist, but, instead in a HOSTS file you have already that is easily edited or downloaded from places like mvps.org or here -> http://en.wikipedia.org/wiki/Hosts_file )
----
Well, @ this point? Unfortunately, I suspect this IS "foredecker" responding as "A/C" & he is caught with his pants down, as most mgt. is, & all they say is "I don't understand or I could not read it" instead of coming clean & owning up to it, then learning more (they are not about learning, they are about scamming imo, having been one in my past & knowning i needed to know more so I went back for CSC degrees ontop of my MIS ones, which were on info. systems basically, NOT how this stuff REALLY works @ the lowest levels as CSC degree tracks DO show you + the theories behind them).
So, I would just like to KNOW WHY as to WHY the more efficient on diskspace & faster loading into RAM 0 based HOSTS file was removed in Microsoft VISTA, Windows Server 2008, & Windows 7 - there really MAY be a GOOD REASON, but everyone @ MS is avoiding it... this reeks of something BAD, imo, almost like a 'cover up'. Sounds nuts, but a straight answer like "here is the trade off we made" or "yes, we screwed up - expect a patch" would be FAR better. Own up to a fuckup, or just say "You have a point, expect a patch"
This was intentional, but, on WHAT GROUNDS? That's all I wish to know here.
This happened, again, after 12/09/2009 patch Tuesday, & onwards in Windows Server 2008 + Windows 7.
I just asked for a VALID techni
"But it also seems to me that you're probably looking in the wrong place for a performance enhancement in the first place." - by Anonymous Coward on Monday October 26, @12:40AM (#29869093)
Am I? You're way, Way, WAY off... I state that, because I can literally prove (& so can ANYONE ELSE who writes code that is) that loading a smaller file, especially since I am making the hosts smaller by 8 bytes PER LINE in the case of 127.0.0.1 vs. 0, & 6 bytes PER LINE in the case of 0.0.0.0 vs. 0. Try it yourself, or have your dev team @ MS do it... &, "argue with the numbers" + result.
----
"Loading a few megabytes of disk onto RAM is indeed a slow operation, but for just that reason you shouldn't be doing it frequently, whether the file is 18 MB or 22 MB. - by Anonymous Coward on Monday October 26, @12:40AM (#29869093)
It only happens in 2 instances, & you can't avoid the first one: When the PNP IP stack starts, FULLY (since Windows XP it has been this way, & that was an idea Apple had first anyhow that MS stole from them - & before that, it loaded TOTALLY, in Windows 2000 + before the user logged on). The secnod case would be when you disable your network connection (which I do when I go offline, just to be TOTALLY safe, when I am not using it).
Still - you saying that making a file smaller, to read it into memory faster, is bad? You are as WRONG as WRONG GETS... period.
----
"You should maintain a pre-parsed format in memory (much like the binary blob file I mentioned earlier) which is itself much smaller than 18 MB, and the initial load will be only a fraction of a second even on a low-end machine -- this is exactly what every major OS has done for at least a decade (sometimes with monitoring for changes in the source HOSTS file so you can "hot-swap" the HOSTS). - by Anonymous Coward on Monday October 26, @12:40AM (#29869093)
What OS' are those? They all use the BSD IP Stack as a reference design, including Windows. The BSD IP stack uses a HOSTS file, which is nothing more than text. What the HECK are YOU TALKING ABOUT?? Once it is read into memory, in the method I noted about & WHEN (2 cases)?? It is then read @ the speed of RAM.
In the case of a LARGE hosts file, that's done only in those 2 circumstances, & LINE BY LINE (like any file is) until the CR+LF (or a null char) is encountered, then the next line is read in, via a loop (most likely a WHILE loop, until EOF (end of file) trailer-record/marker is hit)).
This is NOT a database - wake up already. It is a TEXT FILE in the case of HOSTS files, & I am merely showing AND PROVING a way that works (or worked, until MS changed what worked for 10++ yrs. since Windows 2000 SP#2 or 4 was made, when MS changed to using 0 vs. 0.0.0.0 or 127.0.0.1 (both of which are clearly and provable SLOWER than using 0 & make for FAR larger files).
50%++ larger in the case of 0 vs. 127.0.0.1
&
Nearly 30++ in the case of 0 vs 0.0.0.0
AND, technically, since 0 hex = 0 decimal (both zero in both cases), the translation that 127.0.0.1 &/or 0.0.0.0 go thru doesn't even have to occur with 0 since 0 hex is the same in decimal, and it's 0 in the HOSTS file.
You can ping 0 on Windows 2000/XP/Server 2003 & what does it return? 0...
It's completely legit, & MS didn't have it in the original OEM shipping model of Windows 2000 as it installs from the OEM install CD. You people installed it later, & it makes sense why - IT IS FASTER, & MAKES FOR A LESS BLOATED HOSTS FILE!
So, again - WHY WAS THIS GOOD THING REMOVED FROM A HOSTS FILE?
----
"The vast majority of HOSTS files have fewer than 10 entries and would likely not see even a fraction of a second's improvement because it would fall underneath the smallest HDD block-transfer size. - by Anonymous Coward on Monday October 26, @12:40AM (#29869093
QUOTING MYSELF -> "You can ping 0 on Windows 2000/XP/Server 2003 & what does it return? 0..."
CORRECTING MYSELF NOW -> It returns 0.0.0.0 when you "PING" 0, in Windows 2000, not what I wrote originally... so, correcting THAT now, before you guys reading do!
(AND, it has done so, ever since SP2 or SP4 that is, on Windows 2000 - BUT, before those service packs? Well - You could NOT use 0 in Windows 2000)
So, so much for MS "using standards" on MS' part & the rubbish others here are trying to use vs. my points!
(On standards - Well, & ask ANY webdev on that account also, on "how well MS adheres to standards", ala IE6 etc. et al)
STILL, the ability to use 0 as a blocking IP address in a HOSTS file? Well - it does tend to PROVE that MS did indeed see that using 0 in a HOSTS file is smaller & faster too, which is MY POINT here: AND, also because Windows XP & Windows Server 2003 STILL CAN USE 0 as a BLOCKING IP ADDRESS... period!
So, why did they remove it in VISTA from 12/09/2008 onwards (when it worked before that) on that "Patch Tuesday"), and same on Windows Server 2008 &/or Windows7?
Reiterating this point again here as well:
0.0.0.0 is STILL CONSIDERED a "VALID IP ADDRESS", per my critics here no less... but, apparently? MS also felt that 0 was as well since it makes HOSTS files smaller & thus, faster to load from Windows 2000 SP#2 (or #3 or #4 for sure) onwards!
(AND, any coder can compare a HOSTS file load with 1 hosts using 0 & others using either 0.0.0.0 or 127.0.0.1, & see which loads faster - Especially relatively larger ones like Spybot Search & Destroy or hosts files from mvps.org or WIKIPEDIA http://en.wikipedia.org/wiki/Hosts_file give you to block out known bad sites or bad adbanners, if not ALL adbanners period for more speed, AND SAFETY (above all else on the latter especially))
APK
P.S.=> Also, lastly, FIX YOUR DNS LOCAL CLIENT CACHE, Microsoft - because when it gets relatively "larger HOSTS files"? It begins to lag the system (paging data in & out apparently from its datastructure in memory (which is what the BSD reference model uses, a structure (ala C/C++))... make a redimmable array instead so it can "flex" & grow, or shrink, dynamically instead (sure, some overheads, but... MUCH BETTER THAN LAG!)... apk