381 posts • joined Saturday 14th January 2012 10:26 GMT
Re: Sums up the Reg audience
I may have some distinctive dislike for Microsoft (date back of DOS days) but anything that doesn't endorse ads is good in my book.
Re: Dear MS...
And FFS stop calling programs bloody "apps". I wanna puke each time I hear that apps and the marketing is abusing it too much, too.
Re: “advance the bold vision”
That article is by a troll. Not wanting to dilute your argument with any facts or anything, but in most functional benchmarks Windows outperforms Linux on the same hardware (e.g file server, database server, directory server, NFS server, Hypervisor, graphics, local large file copy, etc.)
Citation needed, I mean badly!! In my humble experience Windows is nowhere near Linux's epoll .
Not wanting to dilute your argument with any facts or anything...
Show some references, please. Otherwise it's just trolling.
Re: ability to discern pictures more quickly than words
I distinguish text faster and more accurately than icons, always have done and I don't see much chance that will change.
I just can't express how true this statement is. Google have similar approach actually, so Microsoft should not be singled out.a Unless you are the designer remembering/associating pictures to actions always seem extra trouble till the point I have to hover the icon to realize what it is.
Re: Biometric will never work
Actually no good copy of a finger is needed, just a digital copy the data of a comparably sensitive sensor and pass it - sockets transfer data in digital format. If the raw (non hashed/salted) data is acquired, basically the entire system is entirely compromised.
Re: I can change my password ...
I know quite little about finger print recognition but I always assume it'd be some hashing of the actual fingerprint. Adding 256bit salt to the fingerprint would solve the case.
My main gripe are injuries, cutting (or esp burning) your fingers leaves out handing dry.
Beer or... bust
I'm a Diet Coke drinker, not pepsi. fwiw. diet coke/diet shmoke, see above.
It's Friday, at least be serious about something as you fail at propaganda.
A nifty tip: if you can cause more eye discomfort: pressing 'windows' key, ask the guys to code few extra full screen stroboscope effects and randomly pick one each time. Later you may explain how that innovative feature helps better out cognition.
Re: That sounds useful if I only want to read the document
That should save the time taken to load full fat Word.
Dunno about Word, open office opens in around 1.4sec from SSD (i.e. no disk cache) on a 3years old laptop (core duo). So the entire copy-paste human time will take a lot more than just double-click the file.
For the quoted amount of time, one get can significantly better looking girls (or boys) [yeah plural]. That's way more overpriced compared the entire line of products of Apple.
Re: In it for the kill
Besides, hedge funds are major stake holders in many, many companies and most don't specialize in dismantling them.
This needs a citation badly.
Re: Using Japanese logic.
Actually, it's a common practice to extract forced confession in Japan. The conviction rate is extremely high according to wikipedia as well mostly due to the mentioned confession rate and lack of jury.
So the logic is right on the mark.
Seriously, it'd be quite easy to connect the voice to any other available voice even if the userId has been "lost".
That even doesn't touch like write mail to XXX.
Yeah, what a dishonest statement.
The companies are "too big to fail" and cannot be tried under criminal law. As we all know money laundering for drug cartels is a fine only verdict. Something among the lines.
Few simple firewall rules: blocking anything that doesn't access the said stock exchange.
Use proxy for any regular web (and block the social crap site), setup a local mail server and then block the rest. The trader won't need more network access.
Banks are usually flooded by super paranoid "security guys", if they take care VNC alike stuff won't be an issue.
Re: I don't think it's a good idea...
There are tons of easily available tools to allow capturing the screen(s) content. DRM is waste of time (incl. CPU), coding and servers no purpose, add unnecessary bloat and obscurity.
I can affect only legitimate customers and I bet "how to save files from netflix " in a search engine will yield a very simple step-by-step. To put it simply in order DRM to work it has to violate the laws of physics.
Use Ctrl+U to view source or proxy.
The movie greatly reminds about "Moon" w/ significantly larger budget... and chicks, erm female characters. And tons of bright scenes, easy on the eyes. It's not a bad movie, especially if you are a sci-fi fan, though. There are some clues about how the story is to unfold and for picking on them I was labeled 'geek' by my girlfriend.
Perhaps the greatest downside would be pretty straightforward storyline.
To get a blood sample you need some scratches, use a modified bathroom door handle.
Alternatives include sedatives in the proposed wine. I am sure you can figure a way once you lay hands on the portable tester.
Re: Is it snowing in hell? - the IT angle
if it snows in hell it is probably raining.
The IT angle: w/ enough rain and brimstone fire you get a nice cloud, albeit sulfur one... but cloud.
Re: Perfect phone
James, you could have skipped malware slur infestation in that marketing rant.
If blink is bad for accessibility popping out ladies, walking half way and offering support in the middle of screen is a truly magnificent idea.
Apple... can't patent writing.
but selection over that writing is there to take
I'd not be surprised if Apple has managed to patent even fart w/ overlays.
Back in the days text selection was simple xor but functionally was exactly the same thing. How on earth can anyone hold patent for such a trivial functionality.
I mentioned that was the way I read it first but it required the voodoo bag so I discounted the option.
Re: More marbles
@Arbee, or they are color blind and mistook red for green. :)
The riddle is rather trivial, so I wonder how they managed to screw the answer... or lost the count of their marbles in this particular case.
The 1st time I read it it appeared to me blue and green marbles also had 3/5 odds to be picked and that would have made the bag sort of magic or voodoo magic. The latter would explain adwords deals nicely.
The most shocking...
...is that bounce back mambo-jumbo could actually be considered patentable. The game is so rigid it's just stupid. Laws from the lawyers for the lawyers, etc. Software patents help no one but the trolls.
Re: humm x86 anyone?
Google/Facebook/Twitter other big boy do not care about consistency.
Who cares if you have receive a different search result ?
Re: humm x86 anyone?
The main problem w/ x86 is the CAP theorem: out of Consistency, Availability and Partitioning you can pick only 2.
So you either sacrifice the Consistency or Availability as you go for the Partitioning . To mitigate the problem you might need to roll into 40Gb infinity band.
Re:30% chance of pregnancy after 50 'perfect uses'
I didn't discuss getting pregnant in general as that involves ovulation, the swimmers actually being able to make it to the final destination and so on.
It was only regarding Math and probabilities. But yeah, probably unclear.
I am suggesting that citing current condoms as efficient birth-control is overstatement.
For instance: Fertility "Awareness based methods e.g. standard days method, symptothermal method." are reported as 75-99% (compared to the 85% of the condom). So no exactly true "more effective than not using them".
I cannot support more condoms for STD/I prevention, yet I do not see them as effective birth-control method.
Re: @stanimir - bollocks
Ohh, I seriously thought I had provided the source (I do not quote w/o sources). Firstly I looked up wikipedia which has adopted it from http://www.fpq.com.au/pdf/Contraception_Efficacy.pdf
Looks pretty fine source and the 98.x % I have seen written on the leaflet w/ the condoms.
The main problem w/ condoms is the friction and then then not-proper placement or size. Wrong sizes can result into a burst/tear and men are often unlikely to pull out to check if the condom is still whole.
Church propaganda would be the last thing I'd do as the Church is significantly unrepresented where I live and has close to zero influence.
Having yourself as anecdotal evidence doesn't work either. The reported rates are from females using the method.
Also 2% fail rate based on attempts would be 0.98^50 ~ .36 chance to get pregnant after 50 attempts w/ 2% chance - I mean your math is incorrect.
also make an effective family planning tool
Unless someone considers one night stands family planning, the condoms are very far from effective w/ their fail rate:
With proper knowledge and application technique—and use at every act of intercourse—women whose partners use male condoms experience a 2% per-year pregnancy rate with perfect use and a 15% per-year pregnancy rate with typical use
I understand Bill Gates may not take security seriously but that's beyond laughable - 2% w/ perfect usage and 15% w/ typical :). Here, pals we have 15% fail rate as "effective"
Moral of the story...
you can't sell democracy to People's Republic of China.
Well played by Baidu.
This is not how it works, to use more than one core the application has to be designed to do so. Writing concurrent stuff aint trivial and it's quite error prone.
At some point you need to synchronize everything just b/c everyone rushes doesn't mean it's done concurrently.
Some tasks can be easily parallelized like path finding, sorting large arrays, etc however when operating on small structures the lack of locality and price to fork/join outweights ability to run in multi-core.
For games usually single threaded performance is the most important factor when picking the CPU (unlike servers [usually])
Re: Google does no evil (tm)
Where did you read "google" in the article?
Simple rule - if a resource can be served by that algorithm it'd be done once only - also the algorithm will not be so "good" on smaller resources.
Compared to the video sizes - 8k (optimistic ratio of your quote 100k) once is less than a drop in the sea.
Re: -bzip2/xzip is the way to go
Neither of them will decompress with zlib
I quite realize that (having 1st hand experience porting zlib to java, java lacks Z_SYNC_FLUSH support - lame as it gets). However, the idea is that even if one browser starts providing support for a better algorithm and there is a mod for apache/tomcat/nginx/whatever it will be automatically available.
The use of compression is quite flexible w/ http and the user agent (browser) can announce any supported one, it's up to the server to decide. It could be a good selling point and it doesn't require a lot of work actually.
Also, LZMA SDK is placed in the public domain. So there is no issue w/ the licensing terms.
Btw I posted results about lzma (xzip) and the compression level is noticeably improved compared to bzip2 but the cpu cycles are increased an order of magnitude too - still a lot better than the "new compression".
@Eddie Edwards -- Re: It is even slower than you say
The idea is that the compression works w/ deflate/inflate that's readily available on every browser - so there is no need to change anything on the client side
On a wimpy Opteron 2372 (@2.1GHz) -- benchmark on sizes/speed vs
corpus - 100000000 enwik8
gzip -9 36’445’248 enwik8.gz (~10sec) - matches the PDF results
bzip2 29’008’758 enwik8.bz2 (~17sec)
xz 26’375’764 enwik8.xz (123 sec) [lzma -- the algorithm is just better]
Zopfli 34’995’756 (didn't wish to way 15+ minutes to complete, data from the linked PDF)
Re: Its a standard (zip format)
The CEN is at the end of the file to allow easier adding of new files w/o necessarily decompress/compress of the entire archive
...but zip file format has nothing to do w/ the article.
Re: It is even slower than you say
The client enjoys less CPU cycles and less bandwidth - it's the server/compressor that has to do way too much work for little benefit.
As for storage -bzip2/xzip is the way to go if you wish to save space (bzip2 is 20% better than zopfli on enwik8 test... and orders of magnitude faster)
and an 8% saving in traffic sent externally for the cost of a set of Zopfli boxes on the outgoing lines
The quoted 8% would be applicable to large enough static text files that are actually cacheable.
Facebook and BCC would have dynamic content mostly, so they need online compression. Zopfli compression algorithm is way way way too slow for online compression. Deflate is fast enough, though.
Side note: sites shall use 'deflate' not 'gzip' - gzip comes w/ crc32 and extra header.
For reference compared to LZMA algorithm Zopfli still ~25% worse in size and still slower processing.
To save bandwidth some browsers shall start supporting LZMA and announce that in Accept-Encoding header - the server would pick the best from gzip/deflate/lzma and server the content w/o wasting tons of cycles and bandwidth on a dated 16bit compression.
Re: The question I have is,,,
From the linked PDF -> the decompression is the same (inflate) as any other algorithm compared to. Probably even slightly better due better cache hits (smaller size).
By default deflate (the zlib compression) does not use maximum compression nor max memory for dictionaries. -9 would be the max compression but not max memory
Also gzip uses crc32 that's practically unnecessary over tcp, so they should have tested via simple deflate.
80times as cited in the pdf is just too slow for few percentages for files in sizes of megabytes. The deflate algorithm is essentially 16bit as well. When transferring such files they would be compressed w/ better algorithms no necessary compatible w/ inflate on the client side. Since "web content" like images and video that are large enough are already compressed internally and their is nothing much to gain.
bzip is not "inflate" compatible on the decompression end.
Not the 1st SSL accident by Microsoft, it happened (iirc 2000) once w/ hotmail / password. Does anyone have fresher memory?