It is the approach that matters, not the particular library. There is no great trick to Boost signals/slots, I coded a servicable one myself using libsigslot to get the basic idea. The thing is, once you have done so the whole misbegotten MOC mess becomes irrelevant and you can integrate your signal/slot code with, for example, your project templates.
I keep seeing these elaborate explanations full of bafflegab about introspection and scripting support to justify QT's big design mistake with MOC, but the arguments just don't hold water. You can do all the introspection you need without any preprocessor pass.
This brings up a question I have been wanting to ask, and since i'm sure this article is full of hardware guys maybe someone here can answer...how low CAN you go before electron leakage makes it no longer worth it?
I'm not a hardware guy, but my understanding is, 22nm is already on the other side of that point where standard design techniques fail and the transistors have to be redesigned to make it work. Because of leakage the gates had to go 3D, that is, wrap around the channel as much as possible. Intel called its 3D design trigate, which is just branding, the normally accepted term is FinFET.
Of course someone like Zuckerberg prefers kids that don't have a life, will put up with any crap their fed by the boss, and won't contradict management.
The same goes for your other bean counters.
True. Unless you are pre-IPO you better regard orgs like Facebook and Google for that matter as just a pit stop to pick up the resume item. Optimal in-out time is roughly two years. Wait for your full vest and you'll look like a lamer while your pals are rolling in trajectory goodness.
Bloomberg sees everything through their beancounter goggles and can't imagine such a thing as "engineering passion" to save their lives. Sure, if you are dead from the neck up you better plan your exit by age 35. But if you have passion to stay on the cutting geek edge you only get more valuable as your engineering discipline matures.
the whole analogy is retarded anyway, because it's invariably qualified with a sentence explaining that it's a die shrink or a new architecture. So why bother with the whole tick-tock business?
Because Intel wants it to sound like a clock ticking away the life of its competitors?
I know this is how intel defines it, but that's always seemed odd to me. Tick comes before tock. A new architecture comes before the refinement of that architecture. Seems like the tick should be the new architecture. and the tock should be the refinement.
It should come as no surprise given that Intel also got the order of bytes in a word backwards.
You twisted my words up entirely. Let me put it more simply: I am underwhelmed by the "tock" this time. As is every commentator with a clue. This process node appears to be a fail for Intel.
Power consumption is important to some people.. Gamers won't care about either.../quote.
Wrong. Power consumption determines cooling requirements, which determines fan noise. Power consumption also determines whether your long suffering power supply needs yet another upgrade. We are already in circuit breaking blowing zone on a lot of gaming rigs.
For people familiar with Intel's Tick-Tock cadence - this should not come as much surprise....
It comes as a great surprise. Normally, the process shrink delivers either or both a clock boost or power efficiency improvement. Normally, also a speedup due to additional superscalar hardware, but Intel explained that one away as "improved graphics". Well, where did the clock boost go then? Power efficiency? OK, all missing in action. So, the big unwritten subtext here is: Intel's 22nm node has got problems. Big problems. Trigate not working out so well?
...even Android hasn't yet found any footing there without the carrier infrastructure that helped it to compete with the iPhone in the smartphone industry...
The security risk isn't always acceptable. Google banned Windows laptops after they got owned by China, now you need executive permission to connect one to the internal network.
Oh right, what these corporations did was "somewhat evil" but not "completely, totally evil". Right, we understand that. Now lets move on to the "illegal" aspect of it.
We're dealing with an evil/delusional Googler here. They do exist, I met some myself. Just notch that one up and move on. Not all Googlers are evil, in fact not even most. But more than should be.
Slashdot Mirror
It's going to be rather hard to get Boost libraries out of the official C++ standard. You may have to wait 10 or 20 years to depreciate them.
Boost libraries are not in C++03 or C++11 or any other official C++ standard that I know of.
It is the approach that matters, not the particular library. There is no great trick to Boost signals/slots, I coded a servicable one myself using libsigslot to get the basic idea. The thing is, once you have done so the whole misbegotten MOC mess becomes irrelevant and you can integrate your signal/slot code with, for example, your project templates.
I keep seeing these elaborate explanations full of bafflegab about introspection and scripting support to justify QT's big design mistake with MOC, but the arguments just don't hold water. You can do all the introspection you need without any preprocessor pass.
C++ is how old now? C? Java? Hell even .NET is no spring chicken. Frankly only the "fad language of the week" is gonna actually be very new and if all
As exhibit "A" your honor, I give you C++11
This brings up a question I have been wanting to ask, and since i'm sure this article is full of hardware guys maybe someone here can answer...how low CAN you go before electron leakage makes it no longer worth it?
I'm not a hardware guy, but my understanding is, 22nm is already on the other side of that point where standard design techniques fail and the transistors have to be redesigned to make it work. Because of leakage the gates had to go 3D, that is, wrap around the channel as much as possible. Intel called its 3D design trigate, which is just branding, the normally accepted term is FinFET.
Of course someone like Zuckerberg prefers kids that don't have a life, will put up with any crap their fed by the boss, and won't contradict management.
The same goes for your other bean counters.
True. Unless you are pre-IPO you better regard orgs like Facebook and Google for that matter as just a pit stop to pick up the resume item. Optimal in-out time is roughly two years. Wait for your full vest and you'll look like a lamer while your pals are rolling in trajectory goodness.
Bloomberg sees everything through their beancounter goggles and can't imagine such a thing as "engineering passion" to save their lives. Sure, if you are dead from the neck up you better plan your exit by age 35. But if you have passion to stay on the cutting geek edge you only get more valuable as your engineering discipline matures.
...it's naive to think that Microsoft will not expand on their background tasks...
Too late.
It's not a hurdle that they cannot overcome...
Perhaps, but it is a marketing fail that they many not recover from.
A 50% GPU improvement over Sandy Bridge is VERY significant.
Compared to other Intel. But compared to AMD and NVidia it still sucks major donkey poo.
And, oh yes, I am underwhelmed by the "tick". On the face of it, Intel would have accomplished more with another go around at 28nm.
Now for the Intel fanboys in the thread, let's shed some authoritative light on the subject.
the whole analogy is retarded anyway, because it's invariably qualified with a sentence explaining that it's a die shrink or a new architecture. So why bother with the whole tick-tock business?
Because Intel wants it to sound like a clock ticking away the life of its competitors?
I know this is how intel defines it, but that's always seemed odd to me. Tick comes before tock. A new architecture comes before the refinement of that architecture. Seems like the tick should be the new architecture. and the tock should be the refinement.
It should come as no surprise given that Intel also got the order of bytes in a word backwards.
You twisted my words up entirely. Let me put it more simply: I am underwhelmed by the "tock" this time. As is every commentator with a clue. This process node appears to be a fail for Intel.
I think bitching about it this extensively advertises an irrational need to defend the brand on your part.
It's all just part of Microsoft's death spiral.
Power consumption is important to some people.. Gamers won't care about either.../quote.
Wrong. Power consumption determines cooling requirements, which determines fan noise. Power consumption also determines whether your long suffering power supply needs yet another upgrade. We are already in circuit breaking blowing zone on a lot of gaming rigs.
For people familiar with Intel's Tick-Tock cadence - this should not come as much surprise....
It comes as a great surprise. Normally, the process shrink delivers either or both a clock boost or power efficiency improvement. Normally, also a speedup due to additional superscalar hardware, but Intel explained that one away as "improved graphics". Well, where did the clock boost go then? Power efficiency? OK, all missing in action. So, the big unwritten subtext here is: Intel's 22nm node has got problems. Big problems. Trigate not working out so well?
x86 and ARM are going to performance and battery life parity by the time Windows8 Launches
Maybe, but not cost parity. Intel would die if it had to sell high performing chips at the price ARM vendors charge.
It says everything about Microsoft.
...even Android hasn't yet found any footing there without the carrier infrastructure that helped it to compete with the iPhone in the smartphone industry...
Since when is 35% market share not "footing"?
I'd say that's a bad bet - trying to "out Apple" Apple.
True. Apple has had more success at trying to out-evil Microsoft.
The security risk isn't always acceptable. Google banned Windows laptops after they got owned by China, now you need executive permission to connect one to the internal network.
No, this was evil, plain and simple. Working to keep salaries down for people is evil.
Particularly if it is a conspiracy.
Oh right, what these corporations did was "somewhat evil" but not "completely, totally evil". Right, we understand that. Now lets move on to the "illegal" aspect of it.
this is going to fly way under the radar, as it affects their employees, not customers
Nice theory, but I have one word for you: conspiracy.
We're dealing with an evil/delusional Googler here. They do exist, I met some myself. Just notch that one up and move on. Not all Googlers are evil, in fact not even most. But more than should be.