<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:wfw="http://wellformedweb.org/CommentAPI/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	>

<channel>
	<title>Mark's Blog</title>
	<atom:link href="http://mark.heily.com/blog/?feed=rss2" rel="self" type="application/rss+xml" />
	<link>http://mark.heily.com/blog</link>
	<description>Just another WordPress weblog</description>
	<pubDate>Sat, 13 Dec 2008 21:14:46 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.5.1</generator>
	<language>en</language>
			<item>
		<title>Lies, statistics, and anti-union propaganda</title>
		<link>http://mark.heily.com/blog/?p=6</link>
		<comments>http://mark.heily.com/blog/?p=6#comments</comments>
		<pubDate>Sat, 13 Dec 2008 21:14:46 +0000</pubDate>
		<dc:creator>mheily</dc:creator>
		
		<category><![CDATA[Politics]]></category>

		<guid isPermaLink="false">http://mark.heily.com/blog/?p=6</guid>
		<description><![CDATA[Do unionized auto workers make $75/hr on average, while their non-union counterparts only make $40/hr? Does this prove that unions are killing American manufacturing? Will the auto bailouts only help to keep the pampered UAW workers in a life of luxury?
Actually, no. The $75/hr statistic is misleading because it includes &#8220;legacy costs&#8221; &#8212; the cost [...]]]></description>
			<content:encoded><![CDATA[<p>Do unionized auto workers make $75/hr on average, while their non-union counterparts only make $40/hr? Does this prove that unions are killing American manufacturing? Will the auto bailouts only help to keep the pampered UAW workers in a life of luxury?<br />
Actually, no. The $75/hr statistic is misleading because it includes &#8220;legacy costs&#8221; &#8212; the cost of health care and pension paid to former employees. Current employees don&#8217;t see a dime of that money. There is a great article that explains how <a href="http://mediamatters.org/items/printable/200812060002">the myth that UAW auto workers make $150,000 per year on average</a>.  It shows how a lie has wings, and will travel all around the news media without being questioned. A <a href="http://www.latimes.com/business/la-fi-gopunions13-2008dec13,0,3433168.story">story</a> in the Business section of the L.A. Times explains that:</p>
<blockquote><p><em>But some Republicans framed the UAW as the villain, criticizing what they called lavish wages and benefits that they said had driven General Motors, Chrysler and, to a lesser extent, Ford to their knees.</em></p>
<p><em>&#8220;I&#8217;m sure that I&#8217;m going to be asked, &#8216;Congressman, I work at Honda&#8217; or &#8216;I work at Mercedes. I get $40 an hour. Why are you going to take my tax dollars and pay it to a company that&#8217;s paying their employees $75 an hour?&#8217; &#8221; Rep. Spencer Bachus (R-Ala.) said last month.</em></p>
<p><em>That wage figure &#8212; widely used by opponents of the auto industry bailout &#8212; is not in fact the wage paid to current workers. It is an approximation of the costs of salaries and benefits for current and retired workers. After wage concessions in recent contracts, the UAW says its workers at GM, Ford and Chrysler plants range from $33 an hour for skilled trades to $14 an hour for new hires.</em></p>
<p><em>Precise wages and extrapolated benefits costs for U.S. workers at nonunionized foreign companies, such as Honda and Toyota, are difficult to ascertain, but Block estimated salaries for current workers are approximately the same. </em></p></blockquote>
<p>Unions aren&#8217;t perfect, but they shouldn&#8217;t take the all the blame for the decline of the American automotive industry. There&#8217;s plenty of blame to go around.</p>
]]></content:encoded>
			<wfw:commentRss>http://mark.heily.com/blog/?feed=rss2&amp;p=6</wfw:commentRss>
		</item>
		<item>
		<title>The easy way out.. or is it?</title>
		<link>http://mark.heily.com/blog/?p=5</link>
		<comments>http://mark.heily.com/blog/?p=5#comments</comments>
		<pubDate>Wed, 10 Dec 2008 01:42:01 +0000</pubDate>
		<dc:creator>mheily</dc:creator>
		
		<category><![CDATA[Programming]]></category>

		<guid isPermaLink="false">http://mark.heily.com/blog/?p=5</guid>
		<description><![CDATA[Some problems are so hard, you are tempted to take the easy way out. Multiplexing I/O using poll(2) and non-blocking read(2) calls has a number of tricky corner cases. After writing and rewriting it several times, I was tempted by the dark side.. threads.
I thought that giving each session a dedicated thread and using fgets(3) [...]]]></description>
			<content:encoded><![CDATA[<p>Some problems are so hard, you are tempted to take the easy way out. Multiplexing I/O using poll(2) and non-blocking read(2) calls has a number of tricky corner cases. After writing and rewriting it several times, I was tempted by the dark side.. threads.</p>
<p>I thought that giving each session a dedicated thread and using fgets(3) to read lines would be easy. However, it is also very wasteful. The stack space required for 1000 threads (at 2MB per thread) would be 2GB. Plus, fgets(3) doesn&#8217;t tell you how many bytes were read, so you would have to call strlen(3) after each line of input was read.</p>
<p>The good news is, I wrote a pretty good thread pool implementation and will use thread pools for certain phases of the session that require blocking I/O, such as the fsync(3) and gethostbyname(3) calls. It&#8217;s also not too hard to revert the code changes and go back to an event-driven model.</p>
]]></content:encoded>
			<wfw:commentRss>http://mark.heily.com/blog/?feed=rss2&amp;p=5</wfw:commentRss>
		</item>
		<item>
		<title>OpenSolaris 2008.11</title>
		<link>http://mark.heily.com/blog/?p=4</link>
		<comments>http://mark.heily.com/blog/?p=4#comments</comments>
		<pubDate>Sat, 06 Dec 2008 20:54:25 +0000</pubDate>
		<dc:creator>mheily</dc:creator>
		
		<category><![CDATA[Uncategorized]]></category>

		<guid isPermaLink="false">http://mark.heily.com/blog/?p=4</guid>
		<description><![CDATA[I installed OpenSolaris 2008.11 on my home computer the other day, and took notes about what I liked and didn&#8217;t like.
It detected all of my hardware, and everything worked out of the box. The device driver utility is very nice, and gives you detailed information about which drivers are used to support each piece of [...]]]></description>
			<content:encoded><![CDATA[<p>I installed OpenSolaris 2008.11 on my home computer the other day, and took notes about what I liked and didn&#8217;t like.</p>
<p>It detected all of my hardware, and everything worked out of the box. The device driver utility is very nice, and gives you detailed information about which drivers are used to support each piece of hardware.</p>
<p>The installer doesn&#8217;t allow you to install into an extended partition; you must use a primary partition. I like to have multiple OS&#8217;s on my boot disk, so I would prefer to install it in an extended partition.</p>
<p>The installer allowed me to proceed without setting a root password or creating a normal user account. When I booted the system, the GUI wouldn&#8217;t let me login using an empty root password. I had to Alt-Control-Backspace to kill the GUI, then frantically login to the text console and set a password. Every 10 seconds, the GUI would automatically restart. I feel like the installer should require a root password.</p>
<p>The system feels very responsive, the graphics are crisp and quick, and the default theme is pleasing albeit a bit too much like Windows XP. There is a network monitor applet in the system tray that blinks when  a network packet is sent or received, which I found to be very distracting.</p>
<p>The package manager is good, but very slow. Unfortunately, there is no vpnc client available, and I was unable to build the program from source.<br />
By default, the system is installed into a ZFS root pool, but I don&#8217;t really like how it was setup. The name of the root pool is &#8216;rpool&#8217;, and the base of the filesystem is mounted at rpool/ROOT. This is aesthetically unpleasant.</p>
<p>For some reason, OpenSolaris still uses /export/home and the automounter, instead of just /home. Now that ZFS and NFS are integrated, I fail to see the need for an /export directory.. any dataset can be exported via NFS, why should it have to be mounted under a specific directory?</p>
<p>The splash screen covers up all the boot messages, and I couldn&#8217;t make it go away. The first time it booted, it took a long time (probably importing SMF services) but I couldn&#8217;t see what was happening.</p>
<p>The default install is very desktop oriented, and there are no developer tools or headers installed.</p>
<p>I tried to reboot from the GUI, and my X session crashed to display a blank white screen. I could type letters and they would appear on the screen, but there was no shell to interpret the commands.</p>
<p>The biggest problem I had was actually with GRUB. When you install OpenSolaris on a system that already has Linux installed, it wipes out your existing GRUB menu and will only allow you to boot into OpenSolaris. I had to jump through hoops in order to get my old GRUB menu back, so when I got it back, I said goodbye to OpenSolaris.</p>
<p>OpenSolaris is getting a lot better, but still has a way to go before I could even use it in a dual-boot setup.</p>
]]></content:encoded>
			<wfw:commentRss>http://mark.heily.com/blog/?feed=rss2&amp;p=4</wfw:commentRss>
		</item>
	</channel>
</rss>
