As they say, if buying isn't owning, piracy isn't stealing.
The story is an exclusive interview with a significant player in the OSS community, and it addresses an important and pressing matter. That is entirely aligned with Slashdot's mission, no matter who's publishing it.
I don't really like El Reg either; its general style can only really be described as a self-parody of whiny bitterness, and I've always found it rather dated, like something that should've been left for dead between the stickiest pages of a White Dwarf magazine in the late eighties. However, this story seems to be free of any of that, which is perhaps a good sign.
I don't disagree with you, but as far as I can tell even the colon isn't necessary for Crockford's syntax to be unambiguous, which annoys me.
It didn't take long to read through the entire Misty spec. Basically it's the software equivalent of a conlang: there's only one short program written in the language, no actual implementation, and it includes a bunch of slightly insane features that would be annoying as hell to actually use. Here are some highlights:
- The syntax for variable assignment is set foo: bar — I initially let this one slide since Doing Weird Things With Colons is a big part of the Make Parsers Dumb Again revolution of the past 10 years or so, but in retrospect it kinda looks like the creator's imagination is being subsumed by the legacy of JSON...
- Many operators are not ASCII. Throw out / * = and adopt their Unicode equivalents! Have fun learning how to make your keyboard do any of that.
- Question marks in variable names for marking predicates and booleans! Schemers, start your grudging!
- Some very conspicuous Not Invented Here terminology: arrays are either "stone" or "antestone" (immutable or not—just call them 'final'!)
- Lots of Crockford's other pet projects get screentime: numbers are always stored in his DEC64 format (a 64-bit float with a 56-bit mantissa and an 8-bit decimal exponent), there are native functions for handling strings encoded in a variant UTF-8 format that's slightly more dense, and it seems JSON has been updated to Nota (mainly by making commas and quotes around field names implicit where possible).
- Probably the biggest pet project, patterns are embedded directly into the spec and don't seem to have any other documentation. The gist is that they're regexes with indenting (not an innovation) and slightly less hairy names for common character classes. The goal of fixing regexes is admirable, but I think more research should have been done on what the state-of-the-art in regex legibility actually looks like, rather than starting from vintage 90s PCREs. At any rate, I don't think the rework is radical enough.
- An attempt at handling parallelism through a fairly unremarkable message-passing system called actors, which is more like a multi-process RPC API than a threading model; there is no way to have multiple threads sharing access to a variable. Should the sharp knives be hidden from the children? You decide!
Must you refer to Sundar Pichai this way? It's very offensive to the shareholders.
For asking such an inconvenient question, you are hereby sentenced to three HP InstantInk subscriptions.
...And, to clarify my own remark, the issue is literally an artificial 5-second delay in a javascript file that was added out of pure spite:
setTimeout(function(){c();a.resolve(1)},5E3);
(5E3 = 5000 ms)
That's not even the half of it—the summary is missing the heart of the story. Changing Firefox's User-Agent string removes the delay.
UNIX is hot. It's more than hot. It's steaming. It's quicksilver lightning with a laserbeam kicker. -- Michael Jay Tucker