There does seem to be more violent left-wing thugs than there are right-wing thugs...
That's because if they're right-wing we don't call them "thugs", we call them "domestic terrorists" (or, even more likely, "lone crazy actor, nothing to see here"), because they tend to skip the whole "trying to shut down this talk" part and go straight to the murder.
More than that, there's probably some moral licensing going on.
The USA elected a black guy. This proves that Americans aren't racist assholes.This absolves America from the moral culpability of subsequently voting for a racist and sexist asshole.
My high school history textbook doesn't have to interoperate with yours, either, but for some reason most publishers are only producing ones that satisfy the needs of Texas politicians as opposed to actual history education.
If there's one thing I've learned in my 25 years in the software business, it's that common standards are better than mutually incompatible competing "standards".
It really doesn't matter who does it, as long as it happens.
There's no revenue in telling people that they can't buy stuff so they throw a fit in the hope to find enough judges that think they can pass laws from the bench.
I know, it's hard to RTFA, but let's be clear on what's happening here.
The DoE is legally required to have published the standards by now. It hasn't done so. This is not "pass[ing] laws from the bench". This enforcing laws already passed by the legislature.
If you don't like this, campaign to get the law changed. Be angry all you like, but be angry at the right target.
They're suing to make the Department of Energy do what it is legally required to do. If you don't like it, by all means contact your legislator about amending or repealing the relevant legislation which requires them to do this.
Oh, yeah, and good luck getting Congress to pass legislation at the moment.
Good point. I have not done C++ coding in the last decade. But it does make me wonder what a C++ project looks like when in brings in 100s of open source libraries into its build and 10s of libraries produced by others in the organization.
In most programming languages, namespaces are somewhat independent of filesystem directories.
C++ libraries are usually very good about using a distinguishing top-level namespace, with nested namespaces only if needed (e.g. if it's a big library with major subsystems, or following Boost, a "detail" nested namespace for implementation-specific detail which clients should not use). So, for example, suppose you want to use cpp_redis in your project. The top-level namespace is cpp_redis. In Java, that would have to be be org.cylix.redis or something.
Now to be completely fair to Java, there was a good reason for this design decision. Java was designed for grabbing possibly untrusted code and all its dependencies from across a network and running it locally on the same VM. Tying the namespace system to DNS makes names globally unique by default. In a language like C++, you can control all your dependencies and their namespaces at compile time. In Java, you can't.
Making white space significant is python's biggest flaw.
I see this claim a lot, but I don't buy it at all. There are lots of languages with significant whitespace (e.g. Haskell, makefiles) that nobody complains about (for that reason).
Anyone who claims that significant whitespace is Python's biggest flaw has obviously never tried to use Python on a big project. Lexical syntax is easily the least flawed thing about Python.
Not that sure --- R is such a clusterfuck that someone managed to write a succesful 150-page book about the different ways in which you can shoot yourself in the foot with it.
R is a terrible language with great libraries. I would heartily recommend R for any statistics job. Just don't try to write any programs in it.
In order to make the comparison fair, you should limit your examples to other popular programming languages that made a successful breaking change, since invalidating large, established codebases of user code is precisely what makes such a change so expensive.
Python has completely stepped on it's own dick by making v3 scripts incompatible with v2.
I remember when they said that of the libc5 vs libc6 changeover. Then it was C++9x vs C++1x (which thanks to certain LTS Linux distributions still isn't entirely behind us).
Python had to make a clean break because it was impossible to JIT-compile Python v2 into efficient code due to its broken variable semantics. The problem was that Python v2 was a victim of its own undeserved success, so the migration to v3 was painful and expensive.
Haskell has the right attitude: "Avoid success at all costs."
Deeper name-spacing nesting. Typical dynamic project might go two deep, a typical mature Java package may be 6-10 deep in nesting package name-spacing.
That's very much a Java thing. I've never seen that level of nesting in C++, or Haskell, or ML, or pretty much any other statically typed language designed for large-scale.
Coding in static languages is more laborious, has more up-front costs, and in general can be a large pain when compared to dynamic languages.
This is a crude way of saying that the compiler doesn't accept code with certain classes of bug still present. Those "up-front costs" are, in reality, costs shifted from later in the development cycle.
If it were a different country, I bet that many here in the US would opine that they should hold a new election and call shenanigans if the ones in power refused.
There is precedent. Generally speaking, this is a matter for the appropriate court of disputed returns, which I assume in this case is the US Supreme Court. And presumably they can only act if there is an actual controversy.
The attempted/failed wiretapping of the Democratic National Committee's office in the Watergate Hotel during the 1972 Presidential Campaign didn't change the outcome of that election either.
We could definitely do with another round of Church and Pike committee investigations.
When I hear that Russia bought votes, we can talk, but sending spam at voters?
No, not just voters, phishing attacks against voting hardware/software vendors and election officials. Basically, anyone who might have access into the equipment and procedures which manage the electoral process.
It's more a question of scaling and what politicians call "optics". Beyond a certain scale, you need professional help if you want your protest to present a coherent message.
Slashdot Mirror
The GSM network that most countries use doesn't count?
There does seem to be more violent left-wing thugs than there are right-wing thugs...
That's because if they're right-wing we don't call them "thugs", we call them "domestic terrorists" (or, even more likely, "lone crazy actor, nothing to see here"), because they tend to skip the whole "trying to shut down this talk" part and go straight to the murder.
More than that, there's probably some moral licensing going on.
The USA elected a black guy. This proves that Americans aren't racist assholes.This absolves America from the moral culpability of subsequently voting for a racist and sexist asshole.
My high school history textbook doesn't have to interoperate with yours, either, but for some reason most publishers are only producing ones that satisfy the needs of Texas politicians as opposed to actual history education.
Oh, did it pass in the Senate?
If there's one thing I've learned in my 25 years in the software business, it's that common standards are better than mutually incompatible competing "standards".
It really doesn't matter who does it, as long as it happens.
I know, it's hard to RTFA, but let's be clear on what's happening here.
The DoE is legally required to have published the standards by now. It hasn't done so. This is not "pass[ing] laws from the bench". This enforcing laws already passed by the legislature.
If you don't like this, campaign to get the law changed. Be angry all you like, but be angry at the right target.
They're suing to make the Department of Energy do what it is legally required to do. If you don't like it, by all means contact your legislator about amending or repealing the relevant legislation which requires them to do this.
Oh, yeah, and good luck getting Congress to pass legislation at the moment.
Why don't the manufactures set their own standards?
Yeah, because that's worked so well for the software industry.
Good point. I have not done C++ coding in the last decade. But it does make me wonder what a C++ project looks like when in brings in 100s of open source libraries into its build and 10s of libraries produced by others in the organization.
In most programming languages, namespaces are somewhat independent of filesystem directories.
C++ libraries are usually very good about using a distinguishing top-level namespace, with nested namespaces only if needed (e.g. if it's a big library with major subsystems, or following Boost, a "detail" nested namespace for implementation-specific detail which clients should not use). So, for example, suppose you want to use cpp_redis in your project. The top-level namespace is cpp_redis. In Java, that would have to be be org.cylix.redis or something.
Now to be completely fair to Java, there was a good reason for this design decision. Java was designed for grabbing possibly untrusted code and all its dependencies from across a network and running it locally on the same VM. Tying the namespace system to DNS makes names globally unique by default. In a language like C++, you can control all your dependencies and their namespaces at compile time. In Java, you can't.
Making white space significant is python's biggest flaw.
I see this claim a lot, but I don't buy it at all. There are lots of languages with significant whitespace (e.g. Haskell, makefiles) that nobody complains about (for that reason).
Anyone who claims that significant whitespace is Python's biggest flaw has obviously never tried to use Python on a big project. Lexical syntax is easily the least flawed thing about Python.
Not that sure --- R is such a clusterfuck that someone managed to write a succesful 150-page book about the different ways in which you can shoot yourself in the foot with it.
R is a terrible language with great libraries. I would heartily recommend R for any statistics job. Just don't try to write any programs in it.
In order to make the comparison fair, you should limit your examples to other popular programming languages that made a successful breaking change, since invalidating large, established codebases of user code is precisely what makes such a change so expensive.
C++11
Python has completely stepped on it's own dick by making v3 scripts incompatible with v2.
I remember when they said that of the libc5 vs libc6 changeover. Then it was C++9x vs C++1x (which thanks to certain LTS Linux distributions still isn't entirely behind us).
Python had to make a clean break because it was impossible to JIT-compile Python v2 into efficient code due to its broken variable semantics. The problem was that Python v2 was a victim of its own undeserved success, so the migration to v3 was painful and expensive.
Haskell has the right attitude: "Avoid success at all costs."
Deeper name-spacing nesting. Typical dynamic project might go two deep, a typical mature Java package may be 6-10 deep in nesting package name-spacing.
That's very much a Java thing. I've never seen that level of nesting in C++, or Haskell, or ML, or pretty much any other statically typed language designed for large-scale.
Coding in static languages is more laborious, has more up-front costs, and in general can be a large pain when compared to dynamic languages.
This is a crude way of saying that the compiler doesn't accept code with certain classes of bug still present. Those "up-front costs" are, in reality, costs shifted from later in the development cycle.
I don't know if there ever was a time in history when one programming language was dominant. If there was, those days are long gone.
Current practice is that almost all nontrivial software projects are multi-language. This trend will grow for the forseeable future.
Haha, comment is a user story.
C maps remarkably well to ASM [...]
It maps well to 1975-era ISAs. Modern ISAs... not so much.
#1 habit of highly ineffective people: Don't bother understanding it, just get someone else to produce an executive summary for you.
Posterity is a nosy bastard.
If it were a different country, I bet that many here in the US would opine that they should hold a new election and call shenanigans if the ones in power refused.
There is precedent. Generally speaking, this is a matter for the appropriate court of disputed returns, which I assume in this case is the US Supreme Court. And presumably they can only act if there is an actual controversy.
The attempted/failed wiretapping of the Democratic National Committee's office in the Watergate Hotel during the 1972 Presidential Campaign didn't change the outcome of that election either.
We could definitely do with another round of Church and Pike committee investigations.
When I hear that Russia bought votes, we can talk, but sending spam at voters?
No, not just voters, phishing attacks against voting hardware/software vendors and election officials. Basically, anyone who might have access into the equipment and procedures which manage the electoral process.
This is crazy, I know, but I don't mind the price of a printer being what the printer costs to make plus a reasonable markup.
It's more a question of scaling and what politicians call "optics". Beyond a certain scale, you need professional help if you want your protest to present a coherent message.