O'Reilly Radar Broadcast

Programming languages are exceptionally difficult to design. Most "new" programming languages end up looking a lot like the "old" programming languages but with a slightly different set of warts.

Programming language design is distinctly different from programming language implementation and distinctly different from program optimization. A programming language designer need to have experience and skills in implementation and in optimization. Optimization is important because it assists in understanding the underlying concepts that need to be supported by a programming language.

Brian Kernighan, Jon Bentley, and others have long advocated for designing and implementing little languages, small domain-specific languages to solve a limited class of problems using "natural" abstractions, operations, and data. The Unix pic language is a classic example. Read the literature about "little languages"; Google is your friend. See, for example, http://broadcast.oreilly.com/2009/04/an-interview-with-brian-kernig.html to learn about little languages.

What I would like to see is for you to learn about programming languages and their implementation before you rush off to design one. You might design, implement, and use a little language for a class of problems you encounter in a domain in which you work regularly. Use it to improve your productivity. Share the language with others who have the same problems to solve. Get their suggestions and feedback and incorporate them into…

As someone who worked for the city of Los Angeles I doubt the security part of your claim is true. People assume the government system is so backwards that it must be inferior in every way but that backwards-ness actually makes it more secure because a lot of their documents aren't on a publicly available intranet. Their system is so antiquated that most transmissions are done via modems which, if you think about it, is more secure.

I'm not suggesting LA stay backwards for the sake of security. But once they jump into the world of putting docs on servers that can be accessed by the public there's going to be a lot of new risk involved. Google's security track record isn't exactly great (http://broadcast.oreilly.com/2009/05/google-failings-say-little-about-cloud-computing.html) which makes me question the wisdom of trusting them simply because they're the cheapest option.

Put it this way. Next time you're bidding on a contract against a guy who will do anything to win how secure will you feel knowing all your private records (medical, divorce, business) are housed on Google Docs?

No, actually, it's called "the source is generally credible".

The paper cited is published by the Linux Foundation, which manages overall development of the Linux kernel, and employs the two lead devs, Linus Torvalds and Greg Kroah-Hartman, directly.

Greg's responsible for much of the device support himself.

I also noted that the claim isn't specifically referenced, but I'm generally inclined to believe it's factual.

Confirmation bias would be seeking out specific information to confirm a belief.

The claim dates back to at least 2008, where Greg KH states it was independently verified by Microsoft:

http://broadcast.oreilly.com/2008/10/how-linux-supports-more-device.html

I've seen quotes from you and I've even heard from you in person that you believe Linux, the kernel, supports more devices than any other operating system ever.

I can back it up by that's true, and it's been independently verified by somebody from Microsoft.

Greg's a developer, not a marketing type given to inflated statistics.

That said, I've asked if anyone's got sources.

Erlang's pattern matching, especially binary pattern matching is very powerful and really unique. Once you use it, you start to miss it in other languages.

Here is how to parse a TCP segment (from http://broadcast.oreilly.com/2009/03/the-power-of-erlang-bit-syntax.html):

decode(Segment) ->
case Segment of
<< SourcePort:16, DestinationPort:16,
SequenceNumber:32,
AckNumber:32,
DataOffset:4, _Reserved:4, Flags:8, WindowSize:16,
Checksum:16, UrgentPointer:16,
Payload/binary>> when DataOffset>4
->
OptSize = (DataOffset - 5)*32,
<< Options:OptSize, Message/binary >> = Payload,
<> = <>,
%% Can now process the Message according to the
%% Options (if any) and the flags CWR, ..., FIN.
binary_to_list(Message)
end.

Here is a link of how to parse an RTP packet:

https://forfunand.wordpress.com/2011/10/10/why-erlangs-binary-syntax-is-awesome/

"... Only podcast I'm regularly listening to is the Stack Overflow podcast ..."

SO is pretty good value. Try

- itconverstation (the best, varied) ~ http://itc.conversationsnetwork.org/

- guardian (light) ~ http://www.guardian.co.uk/technology/series/techweekly

- dan bricklan (quirky) ~ http://www.softwaregarden.com/podcast/dbsl.html

- dave winer (quirky) ~ http://www.morningcoffeenotes.com/

- oreilly week in review (light) ~ http://broadcast.oreilly.com/2008/12/the-best-of-oreilly---now-avai.html

- Microsoft Ch9 (software) ~ http://channel9.msdn.com/Media/Podcasts/

Apple also uses zip for Keynote, Pages and Numbers documents, again for very much the same reasons, and Java's .jar, .war, .ear, .whatevar files also use zip.

.zip seems to be the universal 'combine some stuff in a single file' format, probably because it is old enough to be guaranteed patent free (OTOH: http://broadcast.oreilly.com/2010/06/is-zip-in-the-public-domain-or.html and http://www.pkware.com/support/zip-app-note. http://www.pkware.com/documents/casestudies/APPNOTE.TXT only mentions encryption and patching as patented)