“I can’t argue with him. I think GNOME lost its way when it decided to move from its excellent 2.x release series to a barely usable GNOME 3.x line in 2009. Like many Linux users, I loved GNOME 2.x and hated GNOME 3.x. I’m far from the only one who disliked GNOME 3.x that strongly. Linus Torvalds, Linux’s father, would like to see GNOME forked and the current GNOME 3.x buried…”
“It is available as installable Live media and contains standard Centos software plus some multimedia and desktop additions. I made it a rule not to overwrite Centos Base, so besides the changed artwork and naming what you get under the hood is basically Centos. Stella comes bundled with various 3rd party repos:
– nux-dextop (my own repo containing additional programs, mostly desktop oriented)
– nux-libreoffice (my own repo containing Libreoffice – backported from Fedora)
It was my intention to make it so that the bundled repos do NOT conflict. You do not need to mess around with yum configuration, priorities etc; it should all just work(tm)!…”
“Programmers tend to like round numbers, i.e. powers of two. So do hardware designers. Sadly, this shared value doesn’t always work to our advantage. One common issue is that of cache line aliasing induced by alignment.
Binary search suffers from a related ailment when executed on medium or large vectors of almost power-of-two size (in bytes), but it can be cured. Once that is done, searching a sorted vector can be as fast as searches with a well-tuned hash table, for a few realistic access patterns.
The task is interesting to me because I regularly work with static, or almost static, sets: sets for which there’s a majority of lookups, while updates are either rare or batchable. For such sets, the improved performance of explicit balanced search trees on insertions is rarely worth the slowdown on lookups, nor the additional space usage. Replacing binary search with slightly off-center binary or quaternary (four-way) searches only adds a bit more code to provide even quicker, more consistent lookup times…”
“Go scales quite well across multiple cores iff you decompose the problem in a way that’s amenable to Go’s strategy. Same with Erlang. No one is making “excuses”. It’s important to understand these problems. Not understanding concurrency, parallelism, their relationship, and Amdahl’s Law is what has Node.js in such trouble right now…”
“Threads and processes both require a context switch, but on posix systems the thread switch is considerably less expensive. Why? Mainly because the process switch involves changing the VM address space, which means all that hard-earned cache has to be fetched from DRAM again. You also pay a higher cost in synchronization: every message shared between processes requires crossing the kernel boundary. So not only do you have a higher memory use for shared structures and higher CPU costs for serialization, but more cache churn and context switching…”
“I think EmacsLisp is getting to be a great application base, a really good language and environment to write programs in, not just a fancy editor. A number of people seem to agree and are trying it out.
Here’s some tips and tricks distilled from my 15 years of using EmacsLisp to help budding Lisp hackers in Emacs…”
This tutorial is designed so that each lesson is self-contained. However, it is recommened you read all numbered items in the Elisp Basics section.
For new articles and updates, subscribe: Xah Emacs Blog…”