The Intel tech could be cool (thinking x86 code for the GPU) if an app written for larrabee can run on the CPU if the additional processing units aren’t there. This would be a scheduling issue, hopefully handled by the OS.

Muddled Bush Administration Analogy : The Fucking Point > 1:3 = Argument fail.

YAML is great not because the parser can be simpler, but because it’s simpler for *humans* to write. Ruby on Rails uses it extensively for its config files. What would you prefer?
development:
adapter: mysql
host: localhost
database: abc
username: foo
password: bar

or

mysql
localhost
abc
foo
bar

The YAML version is easier to read and easier to write. It looks similar to a .ini file but supports hierarchical data of various types. YAML is great for configuration files and for serializing hierarchical data. XML is just too verbose for those purposes.

As for Google Protocol Buffers, the main advantage is that it’s *simple*. ASN.1 is a full-fledged standard with a consortium around it. According to Wikipedia, ASN.1 includes a bunch of other standards. The first Google result returns http://asn1.elibel.tm.fr/, which looks like a website written in 1998. It doesn’t exactly encourage me to read it. None of the tutorials that I’ve seen so far look simple, and all of them look like as if they’re written a decade ago. Sorry, most of the time I just want to serialize some simple data. The learning curve of ASN.1 better be worth it if I am to use it.
Google Protocol Buffers on the other hand is well-documented. The website looks good, and so it invites me to actually read all that text. It’s simple and does only one thing: serialization of data. I don’t want an IPC mechanism, I don’t want remote objects, I just want to serialize some data. And since I didn’t already know ASN.1, and the data that I’m serializing is only used within the same application, there is no reason why I should choose ASN.1 over Google Protocol Buffers.