Seconded. I bought a pair of the HD-280's 6-7 years ago. They're durable and comfortable with and without music. I still use them 5 days a week at the office.
You must be thinking of some other processor. The first released Alpha silicon, Alpha 21064, had a pipelined FPU for adds/subtracts/multiplies and a non-pipelined floating-point divide unit.
And yet another plug for Linode. I have been with them for over four years. Their infrastructure staff knows what they're doing (tech support has responded in 3 minutes on a Sunday night!) and they're hands-off with respect to how you want to run your box. Disk space is a little expensive, but it's not oversubscribed. Even the smallest accounts are well worth the money.
Minor nit: fetch prediction logic in modern processors has to deal with unconditional procedure calls. Fetch pipelines aren't shallow anymore, so you need to predict the target address and speculatively fetch that cache line. Often this is before you even know you've just fetched a branch/call/jmpl.
The jump tables/switch statements in Ruby/Python/PHP make target prediction a necessity. The target of this jmpl is rarely the same twice in a row.
Someone tell that to Verizon. They seem to think it's best practice to send the same marketing email to both the original address with the + and the same address without. Better yet, their unsubscribe facility refuses to accept the +.
I wish more people understood the +. I've used it to make incoming mail self-sorting for well over a decade.
There's rounding in virtually every transaction you already encounter. Do you live in a location with sales tax?
In pay periods where my paycheck is mathematically supposed to be consistent, it also fluctuates by a cent sometimes. The value averages out but there's still rounding and it's quite obvious.
For the purposes of this argument, why does it matter at all who runs it?
Reed Elsevier does not get its content for free. Part of my thesis made it into one of their textbooks. The author/editor of the textbook does, in fact, get royalty payments from them. I neither know how much, nor if he also received an advance or other lump payments upon reaching various editing milestones.
At least in engineering and computer science, Lou Dobbs is completely wrong. Heck, as a grad student at a top-tier research university, you can get ~$24K (and a highly tax-advantaged $24K, at that). Depending on the location, that can mean a comfortable five years.
Sure, except Balmer thinks ahead: he ties a rope to the deck chairs. That way, he can pull up whatever muck gets dredged up, slap a dictionary word on it, and sell it as the next New Thing.
Finding these blips is the easy part. Any first year grad student can do it. They will even learn something from the process.
The interesting part is figuring out which blips are important and which don't matter, then explaining why. Pushing the identification part to an algorithm is a waste of time and I don't expect computers to be taking over research part any time in the foreseeable future.
Slashdot Mirror
Bro, do you even 5-digit UID?
Bok bok bok! Just barely!
Case law could be considered LD_PRELOAD.
Seconded. I bought a pair of the HD-280's 6-7 years ago. They're durable and comfortable with and without music. I still use them 5 days a week at the office.
You must be thinking of some other processor. The first released Alpha silicon, Alpha 21064, had a pipelined FPU for adds/subtracts/multiplies and a non-pipelined floating-point divide unit.
And yet another plug for Linode. I have been with them for over four years. Their infrastructure staff knows what they're doing (tech support has responded in 3 minutes on a Sunday night!) and they're hands-off with respect to how you want to run your box. Disk space is a little expensive, but it's not oversubscribed. Even the smallest accounts are well worth the money.
Companies already do for collecting sales tax in states where the companies have a physical presence.
Alright wise guy. Explain twitter.
Minor nit: fetch prediction logic in modern processors has to deal with unconditional procedure calls. Fetch pipelines aren't shallow anymore, so you need to predict the target address and speculatively fetch that cache line. Often this is before you even know you've just fetched a branch/call/jmpl.
The jump tables/switch statements in Ruby/Python/PHP make target prediction a necessity. The target of this jmpl is rarely the same twice in a row.
Or they have something to sell you. Marketers: Shooo! Go away! Leave me alone.
Let me guess, you inadvertently also found the candidates who eat candy constantly while at their desks.
There is no beauty in perl. Period.
Your example is spot-on, though. It can be completely obfuscated ASCII art or an unreadable single line.
Someone tell that to Verizon. They seem to think it's best practice to send the same marketing email to both the original address with the + and the same address without. Better yet, their unsubscribe facility refuses to accept the +.
I wish more people understood the +. I've used it to make incoming mail self-sorting for well over a decade.
File a bug report on it. .tsv files renamed to .csv work as you might expect.
I'm not sure how you have StarOffice configured, but it opens .csv files in calc for me with 'soffice foo.csv &' on Unix.
In Windows I had to change the default handler from Excel. That's a simple one-time switch by right-clicking and changing the "Open with..." setting.
There's rounding in virtually every transaction you already encounter. Do you live in a location with sales tax?
In pay periods where my paycheck is mathematically supposed to be consistent, it also fluctuates by a cent sometimes. The value averages out but there's still rounding and it's quite obvious.
The one more common typo that comes to mind--and the one alias that I do have to fix it up--is "dc" to "cd". The mind boggles.
For the purposes of this argument, why does it matter at all who runs it?
Reed Elsevier does not get its content for free. Part of my thesis made it into one of their textbooks. The author/editor of the textbook does, in fact, get royalty payments from them. I neither know how much, nor if he also received an advance or other lump payments upon reaching various editing milestones.
Okay, I'll bite: Reed Elsevier. Net profit margin: 23%.
This initialization is called BIST (built-in self test) and it has been a standard feature on processors since caches were introduced.
Clearly this is a problem that can be fixed with appropriate AdWords
At least in engineering and computer science, Lou Dobbs is completely wrong. Heck, as a grad student at a top-tier research university, you can get ~$24K (and a highly tax-advantaged $24K, at that). Depending on the location, that can mean a comfortable five years.
Sure, except Balmer thinks ahead: he ties a rope to the deck chairs. That way, he can pull up whatever muck gets dredged up, slap a dictionary word on it, and sell it as the next New Thing.
I beg your pardon. It's typically all dead weight.
Finding these blips is the easy part. Any first year grad student can do it. They will even learn something from the process.
The interesting part is figuring out which blips are important and which don't matter, then explaining why. Pushing the identification part to an algorithm is a waste of time and I don't expect computers to be taking over research part any time in the foreseeable future.
Good to see Rumsfeld invoked to explain databases. It really warms my heart.
If only he had remembered the unknown knowns--the data your DBMS just lost--the set would be complete.