I find it sad that the first post rated "Informative" is almost all political...
"This is one place where the advantage of FDR appears."...
So let's discuss FDR in a more sane manner (if that is possible on the flamefest that Slashdot has bcome).
FDR = Franklyn Delano Roosevelt = Liberal = one of the most politcally charged sets of initials in American politics.
SDR = software defined radio = not really a question of politics
I think you're wrong to be narrowly quantitative about this - I was impressed by the presentation and content that they packed into that slide show. That's the hard thing to do: say something of interest _and_ be short.
Here was what I found interesting:
should a CEO be a "products" person or a "manager"? I recently sat in at a CEO forum where exactly this was discussed, and I expressed strongly the view that the best company leaders have to understand products and take ownership of them. But - you may be surprised - this was definitely a minority view.
what usually happens in tech companies is something like this: a product visionary founds a company; if it's successful, it grows fast and the product visionary, being no manager, hands over the reins (maybe if forced to by worried investors) to a pro manager who can manage the process of change. if the company is lucky, this happens without the culture and the vision dying.
What is impressive about Jobs is that he grew himself from product guy to manager and back to product guy. that is quite a rare ability, let alone doing it _so_ successfully. There is an interesting hint in the slide show about his ruthless business style. This is not something that is very commonly found together with the creativity and visionary qualities you need to get something off the ground. Maybe the combination of those 2 traits is important for the super-hero tech entrepreneur.
This, love him or not, is something he shares with Gates.
I share your sentiment on paedophiles, but it has to be quite likely the database would also be trawled for traces of islamophilia, support for all sorts of fringe groups etc.
paedophilia is bad, but the fight against it shouldn't be the back door to an authoritarian state. If impacts on subtle rights like privacy are ignored in policy, you'll soon end up in a very nasty society where those in power wield much too much of it.
It's been a bad few days for googlers: first the thought that our search results might get sent to the DoJ; then the news that google is bowing to the Chinese government's censorship rules. As a user who cares about privacy, my rights and human rights, I'd like to vote with my click and search elsewhere.
What are the real low/no-evil search-site alternatives? Shouldn't search be opensource & p2p? If Google were to lose the confidence of its users, they will lose their route to profit.
Practically speaking, what does the/. crowd think can be done to keep off the evil google (geegil?) search engine?
When you're rounding in silicon, the same considerations can apply as pen & paper: carrying too much precision costs you gates, power consumption and even clock speed.
One of the methods not mentioned in the article and widely used in hardware - for example in GPUs - is rounding partial product arrays in multipliers. You can be sure that your favourite high performance gaming box would be more expensive and hotter without this odd rounding technique.
The idea is that you're doing binary long multiplication; you form the partial products (the rows of numbers you'll need to add), but you exclude the least significant N columns. This is clearly an underestimate of the real multiplication result. One trick for getting pretty close to the real result is to "promote" the most significant excluded column to the least significant included column. at first surprising, this actually produces a nicely bounded approximation.
On a big multiplier, you can often get good enough results at 1/2 the silicon area and power.
=======
Plug, admission and more information: I'm at http://www.arithmatica.com/ which has design software which allows designers to quickly, efficiently and verifiably build this kind of hardware.
I have pleaded with X1 for a Linux version of their desktop search tool. It is one of the 2 apps I miss from the bad-old-days of running a WinXP desktop.
What I really liked was this: emacs-search like, it produces answers to your queries as you type, providing immediate feedback as to whether you've got your search terms about right: when you start typing, the results list is long; as you keep typing, it shortens. You can stop typing when you have 5 or so results back, and check that they actually contain the record/file you're after.
I have a reasonably good memory, but I'm very disorganised; so the real-time feedback allowed me to home in on information really fast.
I've tried htdig to index my linux desktop, and maybe with lots of work it would become usable - but i don't see myself morphing it into anything as nice as X1. I look forward to trying the other systems metioned in this thread - maybe they'll get closer.
Oh - and the 2nd app I miss? Microsoft Streets & Maps.
Slashdot Mirror
you might be interested in this chart:
http://www.opendemocracy.net/blog/economics/admin/2008/11/18/mollom-beats-spam
that shows what Mollom did to our forum spam after just 2 months. The interesting thing is that the spammers actually seem to have stopped trying to attack this url.
Tony
FDR = Franklyn Delano Roosevelt = Liberal = one of the most politcally charged sets of initials in American politics.
SDR = software defined radio = not really a question of politics
a revealing lapsus??
Here was what I found interesting:
should a CEO be a "products" person or a "manager"? I recently sat in at a CEO forum where exactly this was discussed, and I expressed strongly the view that the best company leaders have to understand products and take ownership of them. But - you may be surprised - this was definitely a minority view.
what usually happens in tech companies is something like this: a product visionary founds a company; if it's successful, it grows fast and the product visionary, being no manager, hands over the reins (maybe if forced to by worried investors) to a pro manager who can manage the process of change. if the company is lucky, this happens without the culture and the vision dying.
What is impressive about Jobs is that he grew himself from product guy to manager and back to product guy. that is quite a rare ability, let alone doing it _so_ successfully. There is an interesting hint in the slide show about his ruthless business style. This is not something that is very commonly found together with the creativity and visionary qualities you need to get something off the ground. Maybe the combination of those 2 traits is important for the super-hero tech entrepreneur.
This, love him or not, is something he shares with Gates.
paedophilia is bad, but the fight against it shouldn't be the back door to an authoritarian state. If impacts on subtle rights like privacy are ignored in policy, you'll soon end up in a very nasty society where those in power wield much too much of it.
What are the real low/no-evil search-site alternatives? Shouldn't search be opensource & p2p? If Google were to lose the confidence of its users, they will lose their route to profit.
Practically speaking, what does the /. crowd think can be done to keep off the evil google (geegil?) search engine?
When you're rounding in silicon, the same considerations can apply as pen & paper: carrying too much precision costs you gates, power consumption and even clock speed.
One of the methods not mentioned in the article and widely used in hardware - for example in GPUs - is rounding partial product arrays in multipliers. You can be sure that your favourite high performance gaming box would be more expensive and hotter without this odd rounding technique.
The idea is that you're doing binary long multiplication; you form the partial products (the rows of numbers you'll need to add), but you exclude the least significant N columns. This is clearly an underestimate of the real multiplication result. One trick for getting pretty close to the real result is to "promote" the most significant excluded column to the least significant included column. at first surprising, this actually produces a nicely bounded approximation.
On a big multiplier, you can often get good enough results at 1/2 the silicon area and power.
=======
Plug, admission and more information: I'm at http://www.arithmatica.com/ which has design software which allows designers to quickly, efficiently and verifiably build this kind of hardware.
I have pleaded with X1 for a Linux version of their desktop search tool. It is one of the 2 apps I miss from the bad-old-days of running a WinXP desktop. What I really liked was this: emacs-search like, it produces answers to your queries as you type, providing immediate feedback as to whether you've got your search terms about right: when you start typing, the results list is long; as you keep typing, it shortens. You can stop typing when you have 5 or so results back, and check that they actually contain the record/file you're after. I have a reasonably good memory, but I'm very disorganised; so the real-time feedback allowed me to home in on information really fast. I've tried htdig to index my linux desktop, and maybe with lots of work it would become usable - but i don't see myself morphing it into anything as nice as X1. I look forward to trying the other systems metioned in this thread - maybe they'll get closer. Oh - and the 2nd app I miss? Microsoft Streets & Maps.