The Great Debaters
Rogers Cadenhead picked up on some of the goofy questions asked of Ron Paul on the GOP debates last night. The one I found particularly galling was the question about whether he was willing to publicly repudiate supporters who subscribed to 9/11 conspiracy theories. Paul handled that one OK (particularly with the response "I'd like to respond to the same question the other candidates were"). Paul's questions regarding support for Israel vs. other Mideast countries were good ones (specifically with his question about why the US is giving money to both sides), and Huckabee and Giuliani immediately turned to stake themselves as staunch defenders of the Promised Land while completely ignoring the point that the US has a terrible habit of arming both sides of a conflict. Similarly, when Mc Cain tried to make a dumb quip about Paul tacitly supporting Al Qaeda, Paul shot back that the US Government already did plenty to arm terrorists by selling arms to Iraq and Afghan rebels. The major problem with Ron Paul is that in the end, everything comes down to money. "We can't afford it" isn't a good way to justify foreign policy, and very few people will vote for that kind of vision. Domestically, maybe, but not internationally.
Mc Cain came off the debates best, I think. He at least limited his chest-thumping to threatening to veto spending bills. Romney sounded fairly polished and rational. Huckabee sounded positively frightening talking about Iran. The candidates mostly said nothing interesting about Iran, and I wish that someone could reassure me that they recognized that the recent Iranian action against Navy ships is exactly the kind of harassment that North Korea's engaged in for 55 years now. Why should our policy be any different towards the PRK, particularly when Iran has at least a nominally democratically elected president, as opposed to dictatorship? Yet another elephant in a roomful of elephants. They can't even take a fair swing at the easy questions like the economy, let alone the really tough questions.
I swear, the GOP would vote for Reagan from the dead if they could raise him from the dead. The candidates were all furiously trying to channel Reagan him at least. It's no wonder there's a dearth of inspiration on the Republican side.
— Gordon Weakliem at permanent link
Who's in the Lead?
James Robertson, ranting on the professional media:
So... the hours of analysis over the meaning of the race is kind of silly - it's barely begun, and something like half the total delegates are at stake on February 5th. This is what passes for "professional" media, on a topic that is fairly important in the US.
What's funny is that for all the hand-wringing over Mit Romney's showing, he's 2 delegates behind Huckabee, and he'd be in the lead if the GOP restored the half of WY's delegates they took away for moving the WY primary up w/o GOP approval. It's entirely possible that Romney could take 2nd in every contest through Feb 5 and go into Super Tuesday in the lead.
Is it just me, or does CNN's primaries coverage really stink? Watching the CNN coverage last night made me think that the CNN anchors are simply talking and not listening to each other: a reporter in the field makes an observation, hands off to Lou Dobbs who makes a comment on a totally different subject, hands off to Anderson Cooper who immediately changes the subject again. It's no wonder they get no depth to their coverage; they have no continuity. The other day, I watched a panel on the NH results that included Jesse Jackson, who went on unabashedly shilling for Barack Obama. Jesse Jackson made a pretty big splash in 1988 - he took roughly 1/3 of the delegates in the convention and should have been given serious consideration as a VP candidate, but was basically shut out. He could have, should have, given some insight on what happened in '88, and the CNN interviewers should have tried to get that out of him. Meanwhile, over on Fox at roughly the same time, Juan Williams had some extremely insightful comments on the Obama campaign. CNN is making Fox's ironic "Fair and Balanced" mantra look pretty good.
— Gordon Weakliem at permanent link
The Fine Art of Computer Programming
In the meantime, think about how many computer science departments earned their reputation by writing an important piece of code: MIT's X Window, Athena, and Lisp Machine; CMU's Andrew File System, Mach, and Lycos; Berkeley's Unix; the University of Kansas' Lynx; Columbia's Kermit. Where are those today? What have the universities given us lately? What's the best college for a high school senior who really loves programming but isn't so excited about lambda calculus?
Undergraduate Programming
How about ITT Tech? No, seriously. This reminds me of a rant by a TA in my pre-calculus class in college, when a journalism student questioned why she needed to understand trig... she just wanted to go out and get a job at a newspaper. He responded by telling her that's the difference between a university and a trade school. He should have added that the world doesn't need another under-educated journalist.
More recently than anything on that list, Indiana University's Mosaic, and Google, which was outlined in The Anatomy of a Large-Scale Hypertextual Web Search Engine. But probably the reason Joel wonders what the universities have done lately is that professors tend to work on things that are interesting to them. The thing is, it's certainly possible to pick up a BFA in CS. My degree from UNT is a BA in CS with a minor in Music, which is probably equivalent to what Joel proposed - more like 1/3 music, 1/3 core requirements (English, Political Science, History, etc.), and 1/3 CS.
That said, I took a graduate course in Finite State Autonoma a couple years out of college and felt cheated that I didn't get that course as an undergrad.
— Gordon Weakliem at permanent link
All I Want For Christmas
I've been using Windows Live Writer happily for about 6 months now, so it's somewhat disappointing that the current version expires in 4 days and the latest version no longer works on Windows Server 2003. I've been running Server for work purposes (the theory is that it matches our production environment better), so now I'm left with either reloading the OS as XP Professional, running XP Pro in a VM, or switching to some other posting software entirely. It's pretty annoying that WLW worked just fine on Server 2003 up until now, the software is time-bombed and the next version has been intentionally disabled for my OS.
I'm working through a couple new books:
- Beginning Ruby on Rails E-Commerce: From Novice to Professional (Rails) - I purchased this used for myself after looking it over in a bookstore. The reviews generally savage this book and the prices of the used items would tend to back up that opinion, but so far that hasn't been my experience. So far, it seems to have good information.
- Transcending CSS: The Fine Art of Web Design (Voices That Matter) - this was a Christmas present, I haven't gotten very far into it. There's a bit of evangelizing in the early chapters, which I find a bit superfluous - the author's likely preaching to the choir. However, it's worth putting up with that. This is one of these books where I'm learning something on practically every page.
— Gordon Weakliem at permanent link
Back on the Rails
Some things I've tripped over setting up Rails on Ubuntu:
rake migrate says /tmp/mysql.sock doesn't exist
Also, mongrel_rails doesn't seem to work on Ubuntu, but script/server does exactly the same thing.
the package manager is really remarkable. I type
sha1
bash responds with
The program 'sha1' is currently not installed. You can install it by typing:
sudo apt-get install sleuthkit
Make sure you have the 'universe' component enabled
With all the gyrations you have to go through with Windows installers, the system doesn't tell you anything half as useful when you try to execute an unknown command.
— Gordon Weakliem at permanent link
Lead Testing & Politics
I got to meet Jared Polis today, at a lead testing event his campaign sponsored in Northglenn. He's doing the same thing tomorrow (Friday) in Boulder. I came in prepared to be overloaded with politicking and whatnot, but the event was not about that at all. Mr. Polis ran the testing himself and most of the discussion was over what toys were causing the detector to go off. We had a few small items that tested high, but fortunately, we won't be re-doing our Christmas shopping. The event was really lightly attended.
I don't live in the district Jared Polis is running in, so my opinion's sort of moot, but my main problem with him is his work with Colorado Amendment 41, which has been a mild legislative disaster. He's running chiefly on a platform of getting the Democratic party to act as a more forceful counterweight to Bush, which is fine, but again is sort of a moot point; Bush will be out of office when Mr. Polis would take office, if he is elected. His campaign website does discuss the disturbing (to me) trend of the use of mercenaries contractors in place of US soldiers in Iraq, which is an excellent point, though I'm not sure what Congress can do to convince the Defense Department that this is a bad policy. All that said, my personal impression of Jared Polis is that he's a class act.
The event was really lightly attended, which is surprising given the coverage around lead in toys. The testing is quite easy, the detector will work through cardboard, so you don't even have to unwrap toys that have already been wrapped for Christmas. The Polis campaign is doing a great public service here, and people really should take advantage of it.
— Gordon Weakliem at permanent link
I ran across some sample code with that was writing out a <script> block with something like this:
document.write('<script src="http://www.example.com/example.js"></scr' + 'ipt>')
I assumed that splitting the end tag was some sort of workaround for a literal string length (the actual example URL was much longer). Turns out that splitting the end tag is necessary to make the document.write() work correctly, and there's actually a functional reason for this. The short story is that the browser doesn't parse the contents of <script> tags, because there's no guarantee that it knows how to parse the language specified in the script, so it can't tell if it's a real end tag, embedded in a string literal, part of a comment, whatever.
Lesson learned: if you're outputting a <script> block in document.write(), you need to express the end tag in some way that hides it from the parser.
— Gordon Weakliem at permanent link
Calendar Girl
Given all the talk about the Enterprise Software Should Be Sexy-Meme, I can't believe nobody's brought up JWZ's Groupware Bad, with the classic quote:
Your "use case" should be, there's a 22 year old college student living in the dorms. How will this software get him laid?
Though I like what Nick Bradbury had to say:
Calling software "sexy" sounds weird to me.
Seriously. The way Scoble talks, next someone's going to come out with the software equivalent of the Snap-On calendar.
— Gordon Weakliem at permanent link
Optimization, SEO Style
I used to think of SEO as a dirty word until my current job. The truth is, search engine rank does matter a great deal to many online businesses. There are unethical ways to boost your rank, and there are some things that you could legitimately call optimization. Like optimizing code, sometimes these fly in the face of efficiency or elegance:
- If a page is indexed multiple times by a search engine, it dilutes the overall value of the page - essentially, the page's "importance" is divided among shadow copies.
- Relative links apparently cause MSN to index the page multiple times - that is, MSN views http://example.com/foo and /foo (contained in a document on example.com) as different resources. This penalizes sites that use relative links.
- Same for http:// vs. https:// - the same content exposed as HTTP vs. HTTPS is considered "different", diluting the value of the affected pages. As a result, resources that require HTTPS should be linked to with https:// and not rely on redirects to force the right behavior.
- Search engines may not account for redirects when building their index. If a crawler requests a document and is redirected, the original URL may be considered as the source URL, thereby diluting the value of the indexed content.
— Gordon Weakliem at permanent link
The Root of the Problem
I've seen the statistics that security on the web is pretty bad - systems get hacked all the time. I've seen some intimations that SSL is the culprit here; that SSL is partly to blame for the Web's security problems. I don't believe it. It's not like people are hacking SSL certs and breaking open sessions to hijack data. It's more like the key problem is weak passwords and social hacking. You don't have to study cryptography for long to realize this. A message with obvious content is not a secret, regardless of whether it's encrypted or not.