kvs

I am currently using http://www.zotero.org/ on Firefox with Dropbox.

If you're good at interacting with people and a computer science major you do have an edge over most computer science majors:-)

Perhaps you can minor in business or other area and see whether that helps with your job prospects better. I think you're "selling yourself" wrong...

One still has to get the decoder through detection systems; just like encrypted code but it's impressive work & results nevertheless.

Nice! It sure does sound like a nice app. Congrats. BTW: Does the iPhone app syncs using mobileme (or at least automagically) or do we have to manually open both apps and sync (Things-like syncing)?

This might help (if you haven't read it already): http://www.mattcutts.com/blog/subdomains-and-subdirectories/

I use subdirectories.

Same here. Good one to monitor user interaction and flow in real-time. I also like the replay.

@Werner: Darn, someone figured out that Dynamo is a flawed architecture. Luckily its only use is storing hundreds of millions of shopping carts :-)

Yes. And the evidence is I find my self going to page 2 or 3 to find what I wanted and sometimes Bing!

No you're not. It is not the same quota.

"But that openness stands in sharp contrast to arcane U.S. immigration policies that discourage young scholars from settling in the U.S." I am an immigrant. Whether I am a "scientist" is debatable but the above statement is disingenuous. I think, sometimes people mix-up the highly talented pool of immigrant with IT works... I am yet to see Ph.Ds refused visas and what not (not to say one need a Ph.D. to do something useful).

"execution"? I doubt that but life in prison is more likely.

Paths: http://en.wikipedia.org/wiki/Path_%28computing%29

* Exercise

* One day without computer (mediafast) (thinking time)

* iPhone/Simplenote/Things

* Snow Leopard/Quick Silver/Things/Dropbox

* Friends & Family

* Absence of scheduled TV (no cable or disc) Hulu instead

I was confused too. Looks like they already have a technical co-founder there; why another? Is he leaving the company?

Nothing wrong with comparing against a single benchmark, regardless of how good/bad the benchmark itself is. On the other hand, how would you have gone about measuring the security of browsers in a meaningful way?

Good to see gov agencies mimicing NetFlix type contests to speed up innovation and adoption.

One thing I'd prefer is dropbox using _my_ S3 buckets so irony rely on them only for syncing. What are the economics of such an approach?

I wonder if the C/C++ code compiled with LLVM/Clang would make a dent in the run time?

Malware just need a C2 for a day or two and then they mve along to another. The days of clinging on to one source are long gone. Please see http://www.pandasecurity.com/usa/homeusers/media/press-relea...

"invented" something small but useful in CS. Won some reserach competition, got invited to the Turing Award ceremoney, got an award on the same stage right after Vincent Cerf and Rob Khan got thier Turing Award. Goose bumps :-)

I voted up and would do so multiple times if I could. I don't like those slideshow sites that want you to login either or those embeddable slides! They all break page on the web, IMHO.

FWIW: http://www.blyon.com/blog/index.php/2009/08/06/twitters-host...

> DNS is so trivial to distribute that it's not normally a worthwhile attack vector. F5 knows that.

Really? So what do you say for the evidence they provide in the article, from Arbor and from Twitter's DNS provider?

Funny how the company's employees need to fill in a pile of paperwork and wait for a month before they can buy/sell anything but the same rule doesn't apply to the company itself.

Yes. That didn't work to our satisfaction.

My recent allocator design is an attempt to address memory fragmentation in Windows XP & Vista. The unique problem I faced was that there was significant amount of code already in place which used C++ new/delete (about 300KLOC) and lot of coding is ongoing at breakneck speed. The allocator is written in C++.

My story:

- Profiled the code to figure out allocation sizes and number of allocations.

   * Found most objects are 32, 64, 128 bytes and we allocate millions of them.

   * Allocation rate is linear in the sense, rate of allocation is pretty steady within a time window.

   * Life time of these objects are long; Majority of them (99.9%) lasts as long as the application itself.

   * Found couple of large allocations in the range of 64MB to 128MB

Can you tell me yours? With such detail, then perhaps we can figure out what's best for you? :-)

This redirection is not happening at application layer, perhaps due to scale?

My question is, how much of the speed-up is due to RDB being append-only and how much of it is due to specialization to SSD.

I wonder how the IT/security staff react to such information? In the sense, what do you do? how to be "ready" for this?

I read the article, looked around their site, and bought the app. All in 5-10 minutes.