furiousBlog

This looks like fun!

Lie by Lie: Chronicle of a War Foretold: August 1990 to March 2003

The first drafts of history are fragmentary. Important revelations arrive late, and out of order. In this timeline, we’ve assembled the history of the Iraq War to create a resource we hope will help resolve open questions of the Bush era. What did our leaders know and when did they know it? And, perhaps just as important, what red flags did we miss, and how could we have missed them? This is the first installment in our Iraq War timeline project.

I haven't had a chance to go through it, but it should be interesting, depressing, and/or infuriating. Probably it'll be some of all three.

Clinton got impeached for lying about an affair with an intern. Yes, it was a lie and it was wrong. But, as far as I know, no one has died because of it and the only money wasted on it was the investigation. Bush's lies have caused thousands upon thousands of deaths and cost billions of dollars. Yet, due to the pure Republican congress, no impeachment. Not that I think an impeachment would do much to fix anything at this point, but at least it would cause him to focus his efforts on defending himself instead of on running the country into the ground.

Due to the uproar my previous post caused (one comment on this blog qualifies), here's a quick note on the database switch and the reasons for it.

A while ago, I started renting a dedicated server with the intention of starting a small web hosting company. The cost was $118/month for a pretty beefy server that could handle everything I was planning to throw at it.

At the time, I set it up with the intention of having many users. This colored the configuration significantly, since I had to secure it such that any of those users couldn't do any damage. This made the configuration a lot more complex.

One way I decided to minimize complexity and potential for harm was to use sqlite as the database for anything it could. Sqlite is a pretty cool little database system that resides all in one file. It's somewhat similar to an access database without being crappy and having a big, heavy, buggy graphical interfact slapped on top of it. Also, it's simple to setup and wicked fast. By using sqlite, each site's database would just be a file that resided in their directory, as opposed to accessing traditional database server system like Postgresql, MySql, MS SQL server (yuck), or Oracle (also yuck).

So, all the blogs that I setup (this one, Cyndi's, and a couple for some friends that have been abandoned) were setup to use sqlite. But, I still had a Postgresql database for the couple sites that didn't support sqlite and the email system.

But now, I've decided not to do the web hosting company and am in the process of switching to a virtual private server. This basically is identical to a dedicated server, except that instead of having a box all to myself, my virtual server is housed on a server with other virtual servers. I still have full root access and so can do whatever I want with it, but the specs are a lot closer to what I'll actually need. There are also a bunch of benefits to a virtual server--stuff like the hosting company can do a full system backup by backing up the one file that is my server. But, the main benefit is that instead of $118/month, the cost is $15/month. So, I'm saving $103 a month and don't have to compromise on functionality.

Anyway, while getting ready for this switch, and since I don't need to complex setup and super-secure site separation, I decided to consolidate everything onto one database system. I'm going with Postgresql since it's super powerful, all my sites, etc. will work with it, and it's centrally admined as opposed to the sqlite which is a bunch of separate files. Oh yeah, one other thing is that the upgrade process for Serendipity (the blog software I use) can't update sqlite databases with new database schema due to a less-than-full implentation of the sql standards.

Over the past couple days, I've migrated all of the blogs from sqlite to postgresql. It was a relatively easy process. You can dump a sqlite database to a file like this:

$> echo '.dump' | sqlite DATABASE.db > DUMPFILE

The resulting file is basically ready for importing into postgres. There were just a couple gotchas that I had to overcome. The first was that sqlite defines the length of fields with type 'int' while postgresql doesn't. So, I had to do a search for int(##) and replace it with just int. Another thing was that the dump made every percent sign (%) show up as %25. That caused a ton of problems and took me a while to realize what was causing it. So, I had to do a search and replace for that, too. The last thing was that it didn't set the sequences that controlled the auto-incrementing id fields for the various tables to the right value. This caused failures when trying to insert anything into the database (entries, comments, etc.). So, for each table, I pulled the highest id using this query (in postgres):

SELECT max(id) from TABLE;

Then, I updated the sequence to have that as the current value, so the next time it was called, it would give the next value:

SELECT setval('SEQUENCENAME', MAXID-FROM-LAST-QUERY);

Once I got it all figured out, it took about 5 minutes per blog to do the whole switch. While I was at it, I went ahead and upgraded all the blogs from Serendipity version 0.9 to 1.0.1.

Now that everything's on postgres, the migration to the new virtual server will be pretty easy. It would've been pretty easy beforehand, too, but now it's consistent throughout all of the sites.

I guess that wasn't the 'quick note' that I thought it'd be...

I've switched the database backend for furiousBlog. So, if you notice any weirdness, please let me know.

I've waited too long to post about the last couple days of the Lake Lure vacation. But, I feel like I should post about them just for the sake of completeness.

The last two days of the trip, August 4th and 5th were the most laid-back. My brother Brett and his wife Angela were leaving on the morning of the 4th. So, we just hung around the pool near our condos while they were getting packed up and ready to go. My neices Melinda and Emily enlisted the help of my sisters Janna and Debbie and my mom to try to dunk me. They'd hatch their schemes and 'sneak' up on me and spring their trap. Sadly for them, they were the ones who always ended up under the water. We spent about two hours there, at war the whole time.

Around noon, Brett and Angela were ready to go, so we got out of the pool and said our goodbyes. It's too bad they had to go right then, because after they left we headed out to a swimming hole up the river a little ways.

I've just posted two versions of a photo of a bee on a flower that I took at Biltmore. The first is converted to black and white, something I've never been very good at, although I quite like this one in b&w. I converted it using the GIMP's* channel mixer. I set it to monochrome and (I think) red 60, green 30, blue 10. A little sharpening, some burning in at the edges, and Bob's your uncle: a black and white photo.

I thought I was done after the b&w version, but for some reason as I went to close the GIMP, I started fiddling with the original color version of the photo instead. The process to the color version is actually a lot more interesting. OK, that's probably seems like a lie to everyone but Gary and I, but it's true. First, I imported the original from the RAW camera file twice; Gary mentioned this a couple days ago, so it must've been on my mind. The first import was set to darken the exposure about a half stop. The second was set to lighten it about a third stop. So, at this point, I had one image with two layers representing two different exposure settings.

I applied a layer mask to the lighter layer. A layer mask is a way to hide certain parts of a layer. You can paint on the mask to show or hide stuff. I made it so my mask only showed the bee and its flower show from the lighter layer. I tweaked this layer's levels, saturation and brightness to punch up the colors a bit.

On the darker layer--only showing the background due to the layer mask--I did a couple things to make it fall even further back and allow the foreground to really stand out. Basically, I burned in the whole layer somewhat, and then desaturated it about 30 or 40%. If you compare the full flower in the background on the two flowers, you can see that the color version's flower has very little detail in the center part. The burning in is what causes this. The two effects together are what cause the dramatic change in lighting between the foreground flower with the bee and the background flowers.

To finish up, I burned in the edges a bit. I think it was Ansel Adams who once said that every photo could use a little burning in at the edges. While I don't burn in the edges of every photo I take, I do it a lot.

I worked on these two photos for about an hour and a half. Mostly, that time was taken up by trying different things to bring out the image that I wanted to produce.

So, there you go. If you ever wondered what goes in to one of my photos, now you know. I'm not sure why I felt compelled to post about this since Gary, (and maybe Rob) is probably the only one who cares, but felt compelled I was. So, write I did. Looking back, I think this entry would've been better had I posted in-progress shots showing the effects of the various changes I made. Sadly, my compulsion has left me.

* The GIMP, or General Image Manipulation Program, is an Free and Open Source photo editing program.