Changelog

I like simple designs. Crowded pages are distracting, they move the user’s focus from content to presentation. This usually makes finding content ona site a chore rather than the natural function of the site.

I appreciate sites like NTK, craigslist, and Google becuase content is presented in a clean, straightforward fashion. They make good use of hyperlinking and try to help me find what I want. Material is neither hidden nor obfuscated. I wish that more sites would forget neat DHTML and Javascript tricks and simply present the user with information. Make the site easy to navigate, and provide hints about where a user might find additional information. If you want to keep a user at your site, give them something to see. Don’t make it a chore to find information or users will leave befor they really find what they want.

Visual design has its place. But I would rather have a easy to use site with no graphics and minimal layout then a beauftiful site with nothing to say.

Of course, I will accept that for some sites, design is their content. These sites may continure to ignore all of the above.

Plugin code can be found here to convert fully-qualified URLs into relative URLs. This will help anyone who wishes to setup a staging site or a mirror of their MT site. Two versions are available, including a filter to add a “link_type” tag attribute to your MT tags.

Mean Mr. Mustard says that the Chechen guerillas are the enemy. He’s right. We (the world) is at war with those who would use terror to further their aims. That these people think such actions will help them in any way is mind boggling. However, I think Instapundit says that anti-warbloggers won’t like what Mr. Mustard has to say. I can’t see why, Mr. Mustard had a point. Of course, my question for everyone is: Why is President Bush waisting our time in Iraq when there are truly useful things he could be doing?

So I took the GRE (general) on October 10th. Stated time for the official score report to arrive is 12 to 15 days. I want my report dammit! Ok, I admit, it’s a bit early to get mad about the whole thing, but I really want to know how I did. When I saw my preliminary report upon finishing the test I must have gone into shock, because I’m convinced those scores couldn’t have been real.

For the record, I think I got a 720 verbal (which is good, but I did better on the practice test) and 780 quantitative (this is about 150 points higher than I got in the practice). If these are my real scores, I’m going to be ecstatic. If I did that well on the analytical section (stupid essays) I may have a heart attack.

I can’t do it… I just can’t stop tweaking my site. I think I’ve created and discarded on the order of nine distinct designs and layouts in the past month. Every time I sit at my computer, I tweak something. Every time I browse the web, I tweak something. I can’t stop.

I see a site I like, everytime I see some neat new trick, I have to add it, I have to try it. What I really need to do is add themes or some kind of stylesheet switcher, so I can satisfy my hunger for tweaking.

• Auto Focus
• Punch Drunk Love
• Moonlight Mile
• Secretary
• Sprited Away

Must write grad school essay now… Cannot postpone inevitable any longer. Future hangs in the balance! Must write essay… Must write essay… Must wri… Oooh… Shiny things!

Scott Adams:

But once you start noticing that your historians are making stuff up and your figure-skating judges are rigging their events, and your priests are dating children and having better sex lives than rock stars, that’s what I call a bubble. That’s the cab driver giving you stock tips.

We’ve left the dotcom bubble and we’re left with the weasel bubble. That makes a bit of sense in my mind, but it doesn’t giev me much hope for the immediate future…

It seems that MT screws up lists… I hadn’t realised that MT is quite so enthusiastic about line breaks. I had imagined that is converted skipped lines to paragraphs, but hadn’t realised it inserted so many hard breaks. This made lists look really funny. To fix, I’m simply writing lists out all on one line. Ugly as hell in HTML, but looks right when rendered.

A quick recap of projects:

• Trackback and comments – One or the other, or both… Haven’t decided yet.
• Wiki Integration – 0xDECAFBAD style integration would rock (see also Steven Frank).
• Blosxom Hacks – Blosxom + email posting + category tree + good archives + trackback + pretty URLs + everything else I can find!

On of the neat things about the web is that it fosters distributed conversations. There are several forms of this idea floating around, including referer lists, pingback, and trackback. I should also include comments in that list as well. However, I’ve noted several neat people intentionally don’t allow comments.

Sites that stick to non-comment forms of post-linking seem to view the idea of distributed conversations in a very pure form. The web is supposed to be distributed, so why make you come to my site to make a comment? Make a comment on your site, and I’ll make note of it and let other people know about it as well. At the same time, this seems to remove something of the community spirit that grows around websites.

So should a site allow comments? Or stick to more indirect means of linking responses to posts? One could easily argue that forum-centered sites take care of the community thing, and so we need not bother. However, a community may be exactly what the author of a site desires. Come on, isn’t Slashdot just a big noisy blog? Anyways, I’ve once again managed to avoid any form on coherent argument, so let me summarize… Many forms of maintaining web conversations, both centrally and otherwise. Which you implement depends in large part on your goal with your site.

Two last thoughts:

• I like sites which integrate trackback and comments into a single commentary system. This is a nice compromise between the comments and trackback/referrers/pingback.
• Comments are nice when your freinds don’t have blogs on which to flame your ass.

I’ve frequently tried to come up with a way to make my computer match my thought process. In particular, a means to keep track of all the electronic cruft that builds up over time (movies, music, documents, email, bookmarks, etc.). I’ve tried coming up with some idea of how this information might be stored and how everything is structured without really finding a good mapping. Maybe the problem is that there isn’t one.

Part of the problem is that our mids are not heirarchichal. We understand such things, and we use them to construct meaningful relationships between things, but our minds aren’t constructed that way naturally (I think). I think this creeps through when we watch people trying to navigate deeply nested information structures; such things generally give people headaches rather than being useful.

So where do we go from here then? Well, the structure of the brain is most closely mapped to a network (see Neural Nets). We can map this to an information structure pretty easily (see The Brain, etc.). This allows for more complex relationships than a tree structure as it allows for nodes to be cross connected across heirarchies. In other words, we allow information to belong to multiple sets.

So how can we make a computer work this way? Well, I’ve yet to see a good implementation of a network structure, or at least a usable one. So here’s a simple architecture: information is stored in a database, keyed by topic. Each node in this database has a primary topic and can have an arbitrary number of sub topics. We view this by taking “slices” of the information set, presenting the closest matches, along with near matches. Of course, the next thing you ask is “Hey, that sounds like a search engine!”. Why yes it does! I think this is why Google is so popular. Anyways, so yes, basically I think the best way for us to archive information (and I thinks this goes for our hard drives as well) is to dro it all in a database, automatically (and implicitly) cross-link the hell out of it, then use Google to search it. I really think this would work better. I think…

Last thought: why can’t I ever make it to the bottom of a post without totally losing any semblance of logic or stucture? Maybe my attention span is too short…

Much has been made lately about the latest update to Google. A lot of people are whining about moving down the ranks, but beyond that there seems to be some evidence that Google doesn’t weight non DMOZ sites very well anymore.

If this is true, I hope they fix it. But I think I’ll give Google the benefit of the doubt for a few reasons:

• I can’t spell pallete (er, palette)
• Same easy to use, clean interface
• A history of Doing the Right Thing™

My misson (since I’ve chosen to accept it) is to take the next step in infogeekdom: integrating my weblog with a Wiki.

The scheme is simple. Following in Steven Frank’s footsteps, I shall install WikiTikiTavi and use Les Orchard’s MTWikiFormat MT plugin.

More references:

• Better living through regular expressions (Mark Pilgrim)

One last thought:

Why should all this have to be seperate? It seems to me that the aims of Wikis and ‘Blogs are hardly at odds. A weblog tool with Wiki like syntax would be a blast. Make linking shit even easier I say!

Ok, one more…

Afterwards, I’ll have to work on linking from the Wiki to the blog… Maybe an embedded MT-Search query or something?

One big question that seems to be floating around academic circles these days is whether tis’ nobler to teach students to understand foundations or to steep them in the latest research. My girlfriend has seen this debate in Biology circles, and I’ve seen it in Computer Science circles. I won’t pretend to answer the question, but it brings up an interesting (maybe false) trend in science: breadth versus specialization.

NOTE Rampant speculation void of supporting evidence follows… Specialization has been seen as an inevitable side-effect of our growth in knowledge. We have seen an explosion in -ologies of all shapes and sizes. The problem become instructing students in such a way that they get some breadth of knowledge about all these fields. Part of this is simply to broaden their minds, the second is to recruit new grad students. Soon there are so many -disciplines that teaching students some aspect of all of them because futile. Then what do we do?

Some would skip all that junk, and teach them about modern research. Let the basic get worked in to explain what is happening here and now in the broader field. Irrespective of who gets to decide what research is worth teaching, this approach bring up an interesting idea: has the trend toward specialization begun to reverse itself? We have entered into a period in our intellectual development wherein it seems that no field of study exists within a vacuum. Now, this may never have been the case, but it seems that some would have you think they did. A scientist or academic today does research that croses many disciplinary boundaries. Modern research seems to elude efforts to classify it into a single field. At the University of Colorado Dan Jurafsky recently won a MacArthur Genius grant. Prof. Jurafsky is a Linguist. He also happens to be listed under the faculty of the department of Computer Science.

Have we entered into a stage of intellectual development where disciplinary boundaries will cease to exist? Have we begun that great unification of fields at the end of which lies true enlightenment? Am I completely insane, and totally behind the times? I dont know…

I recently added a page to my site where I can keep a list of interesting articles and posts that I want to read. Mostly these come from weblogs and news sites, many of which are hopping on the RSS/RDF/XML/FOAF/etc. bandwagon. The script I wrote to track this bookmarks requires me to input any info I have about the link by hand. So far I enter the title (or make one up if there isn’t one), the author, and the URL. This gets parsed using Perl, and included in my page using SSI. Nothing special. Then it hit me. Most of these articles are about metadata and the handling of such. Why the hell do I have to put all this junk in by hand?! Most of the metadata I want is laying around in these sites already. Most have RSS/RDF feeds which specify who wrote an article, when they wrote it (posted it), where they posted it (permanent link), and the title (if it exists). So my next question is pretty basic: how the hell do I yank this stuff out and use it?