Distributed darknet crypto-geeks have a pretty poor track record of creating unbreakable systems, too.

SSL/TLS: Launched 1994, sold as impervious, significantly compromised roughly once a year ever since

PKI: Same story as SSL, except you also get fun design decisions that allow foreign governments and corporate IT to impersonate any host they like

Tor: Launched 2002, all onion layers pierced by 2012, requiring only sufficient funding

Bitcoin: Decentralized, anonymous, encrypted to the hilt, billed as economically sound, it goes through market crashes that make Wall Street look sane, is infested with more scum an villainy than a Tattooine nightclub, and in the end isn't really all that anonymous anyway, for much the same reason that a Tor veil can be broken.

Silk Road: Launched 2011, collapsed 2013.

Silk Road 2.0: Re-launched 2013, shut down 2014.

DNS was created in 1984 to replace the old flat HOSTS.TXT system, at which time the file contained only "several thousand" entries, according to one source I found. Maintenance and distribution of the file was already becoming a problem by that point.

The oldest actual HOSTS.TXT file I found for download was from 1990, and contained about 9,200 lines. (No link; don't want to spam someone's Internet history server just to prove a point. Do your own Googling if you don't believe me.)

There are single data centers with more than a few thousand public-facing IPs in use.

As for this vague handwavy idea of a shadow domain name system, what's going to make that immune from the same sorts of attacks? There's this vague notion that if it's distributed and encrypted, it will be impossible to kill, but guess what? DNS is distributed and encrypted already.

That's exactly my point. Whatever goal the CS community sets for itself on the road toward AGI, as soon as we achieve it, we redefine "intelligence" to not include it.

At some point, we're going to have a machine competent enough to demand its voting rights, then we get to fight the 1860s-1960s civil rights battles all over again. "It can't vote, it's just a computer!"

I'd say we keep raising the bar.

"If a computer can play chess better than a human, it's intelligent."
"No, that's just a chess program."

"If a computer can fly a plane better than a human, it's intelligent."
"No, that's just an application of control theory."

"If a computer can solve a useful subset of the knapsack problem, it's intelligent."
"No, that's just a shipping center expert system."

"If a computer can understand the spoken word, it's intelligent."
"No, that's just a big pattern matching program."

"If a computer can beat top players at Jeopardy, it's intelligent."
"No, it's just a big fast database."

That last paragraph isn't too clear. Allow me to clarify it.

If if all rights are individual rights, then any law that denies specific individuals the freedom to exercise that right is unconstitutional. You can't say, "This right belongs to these people over here, but not to you because you are not in this special class of people." There are no special classes, as far as the US Constitution is concerned.

Some people interpret the 2nd amendment to the US Constitution as protecting a right of the states or of "the miltia". This legal dodge is intended to let them say that the right does not belong to the citizenry individually, giving justification for gun bans. That legal theory was shot down six years ago in DC vs Heller, but that doesn't stop some from persisting in misinterpreting the amendment.

That is to say, if all of the rights protected by the US Constitution are individual rights, laws that deny free exercise of those rights are unconstitutional.

You're arguing my position. If your needs are relatively simple, you can use Gimp, Elements, or Pixelmator roughly interchangeably. If you need to do anything tricky, you'll probably run into a wall with all three.

Gimp is only comparable to Photoshop if you don't know the extent of Photoshop's capabilities or wouldn't push it to its limits if you did. Gimp is closer in capability to Photoshop Elements or Pixelmator.

At least they don't commit the inverse problem, where a single word is blindly assumed to be a search term just because it doesn't contain a dot or slash.

Chrome, Safari and IE all commit that UX sin, which is really annoying when you're trying to go to an internal LAN web server by name.

Chrome and IE let you hack around this by appending a slash (e.g. "myserver/") but Safari doesn't. You end up creating bookmarks purely to avoid having to type the FQDN or explicitly prepend "http://"

FORTRAN wasn't the language in 1964 that you think of as FORTRAN today.

Most people's concept of FORTRAN is FORTRAN 77 or its descendants, which was 13 years in the future from BASIC's introduction.

At the time of BASIC's introduction, FORTRAN IV was the current version.

FORTRAN wouldn't be ANSI-fied for another two years as FORTRAN 66, so every version had machine-specific features. Also, because FORTRAN's development was largely driven by IBM until FORTRAN 66, all the non-IBM versions were "nonstandard." Imagine if, today, every computer came with a C compiler and there were no ANSI or ISO standard to constrain its behavior. The last common reference would have been K&R '78.

Another fun feature of early FORTRAN was fixed column layout, common among languages invented in the punched card era. That is, you had to do things like start all statements in column 7 or later, because the first 6 columns had other meaning.

Early FORTRANs also had very primitive program structuring concepts, hardly raised from the level of assembly language.

Read through the Wikipedia article. You'll probably be shocked at how primitive FORTRAN was in the early 1960s.

Are you aware of even one substantial program that builds without changes under both Ocaml and F#? I'm not.

F# is a "port" of Ocaml in about the same sort of way as C# is a port of Java.

Familial resemblance, yes. Compatibility, definitely not. And as with C# and Java, the gap widens year by year.

IMHO, F# is a vast improvement on Ocaml.

F# and Ocaml are pretty different, actually. They share about as much familial resemblance as C# and Java, or C++ and Objective C:

- The default "light" syntax in F# does away with most of the semicolons and other noise required in Ocaml

- The OO facilities are entirely incompatible

- The standard libraries are almost entirely different; there's only a tiny bit of overlap

- The mere fact of F# being a .NET language has many practical effects on the language, down to strange implementation details like what it means to rotate an 8-bit integer right by 8. F# and Ocaml do not do the same thing! F# differs because it behaves like C#, whereas Ocaml behaves like C in this regard.

- Ocaml has separate operators for integer vs floating point arithmetic; F# overloads the traditional set, though like Ocaml, F# won't let you implicitly mix integer and FP arithmetic

I could go on, but suffice to say, knowing one only gives you a bit of a leg up on learning the other.

It typically takes about 60% of the number of code lines to accomplish something in F# as with C#.

Don't take it from me. Go to the MSDN library page for any given .NET API, scroll down to the Examples, and compare the F# and C# versions. (Not all pages have F# equivalents for the C# examples, but an impressive number do.)

It's no mystery why that is so. F# is highly expressive and doesn't require much boilerplate.

You can mix an F# library into a C# "solution" (urp), so the next time you need a new class or module in a .NET project, I challenge you to write it in F#. You might just get hooked. If not, learning an FP language changes how you think about software, in a good way, even if you never use a pure FP language again.

I also find that correct F# code is more obviously correct than equally correct C# code. It just seems to snap into focus at some point. This might be due to the mathematical rigor of pure FP, or it might just be another benefit of its expressive syntax.

Not much chance of that. F# just hit #12 on the Tiobe Index, up from #69 this time last year:

http://www.tiobe.com/index.php...