Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror

Comment Re:ChromeFrame & Chrome (Score 1) 104

WSUS only equals delayed updates if it is managed wrong.
If all you want to do is minimize bandwidth, you can set WSUS to auto-approve updates.With the WSUS server checking for updates regularly, when the 3am install domain policy ticks over, all updates would be ready to roll anyway. It may mean that some updates get delayed by 1 day, but if your patches are that critical then
a) you should not be auto-accepting updates direct from microsoft via windows update (the chance of things breaking automatically is higher than the chance of something happening from a day or 2 delayed patch), and
b) you've probably thought this through anyway.

Comment Re:Here's a thought (Score 1) 295

correct.
However you or basically anyone with a SSD drive can easily do better than 125MB/s -- you can even do that with a good mechanical drive. (you don't even need raid).

The real question is: what domestic use scenario regularly requires that much bandwidth, other than copying an entire cd's worth of information in 2 seconds rather than about 10. (the overhead of finding what you want to copy, and the time it takes to issue the paste command at the destination will possibly be larger than the time it takes to copy in either scenario; for most things the computer is "sufficiently" fast.

Comment Re:Not as real a threat as on Microsoft Windows (Score 1) 252

There are common code segments across large parts of that ecosystem. eg: what fraction are running any kernel between version 2.6.37 and 3.8.8? (http://arstechnica.com/security/2013/05/critical-linux-vulnerability-imperils-users-even-after-silent-fix/ ; top google link for 'may kernel exploit')

There are many different combinations of the same software options, with a few extra patches; I don't believe there are nearly 'thousands' of unique code bases, and even then there has to be very large exposed code segments common to many of them. (eg: What if a remote code exploitation flaw was discovered in Apache? )

Having said that, the variability in file paths, memory locations, patches, versions available, or even what windowing system libraries are would make any potential issue much more limited in scope compared to over 1/3 of all computers on the net affected by either a xp, or windows 7 flaw.

Comment Re:Damned if they do... (Score 3, Insightful) 275

the other thing here is this only makes it clear that the link is "accessed" -- it's quite possible that the link is not persisted in any way. In that case this would just be an automated part of the message passing process, and not a record of the conversation.

It depends on if skype is sending all chats, or just the links. It depends on if microsoft is archiving what it receives or just checking them for malware. As usual, more information is required to make an informed judgement on this issue.

Comment Re:Sounds good except that I ..... (Score 1) 133

probably number 2 -- all you need to do is to have you wear levelling software swap infrequently written cells onto frequently written ones, once some write disparity has arisen. something like:

onWrite(data,location){
    if(location.writeCount>threshold*drive.writeMinimum){
        write(drive.writeMinimum.data,location)
        write(data,drive.writeMinimum)
    } else
        write(data,location)
}

(i'm sure this is a sub-optimal implementation, but then i'm not officially a hardware expert, or even a programmer, just a science grad student)

Comment Re:These peanuts are the BOMB! (Score 1) 337

overwrite with 0's / random data multiple times to ensure that not a single bit of usable data remains.

(not that it has been shown that you can recover data which has been overwritten even once with any degree of certanty (i believe the best that was done was recovering data overwritten once with zeros at between 80 and 90% success per bit, and that was an old drive using techniques which are not realistic today. the cost would be quite a few seconds electron microscope time per bit, so unless you potentially have data worth several thousand billion $, return on investment of extracting data is negative on any modern hard drive)

Comment Re:My business plan includes world domination (Score 1) 259

the grand parent wasn't promising 2x the computational speed, and neither was I.
As someone who does a fair amount of image processing i'm well aware of single processing/multi processing limits, and how 2x3gh ~= 6ghz, it can be quicker or slower, depending on where the bottleneck is.

eg: the duel processor, quad core mac downstairs is the quickest box i have access to, unless you need lots (>12G) of memory and cpu's (>8), in which case the older altix box (11nodes, 2 cpu's,4G/node)is faster, or if it is a low mem (fits in 2G), sequential job, then my core2 laptop is the quickest box.

Pure Ghz wise, the altix box wins hands down, and my laptop comes in last, but for most serious jobs, the 8cpu's are better, and for minor jobs, 1 cpu wins easly.

Comment Re:Wolf 360 hack (Score 2, Funny) 683

my favorite hack is doing object click detection by rendering objects in a function pointer (4 byte pointers, rgba colour buffer .... ) -- when you click, a render pass is done with no lighting, effects, ... , and with each object coloured according <rgbaColour>object->onClick. To act on a click, read of the colour of the pixel under the mouse, and call it :)

(oh, and solaris lies about giving you rgba -- it only gives you rgb -- my code would work on one of the uni servers, but not on another)

Slashdot Top Deals

"It says he made us all to be just like him. So if we're dumb, then god is dumb, and maybe even a little ugly on the side." -- Frank Zappa

Working...