Cafe Bucky

I wrote an article about PHP and Heredocs, which has just been published at ONLamp.com The gist of it is to use here documents to separate PHP classes, SQL statements and other resources, and HTML templates. The article evolved out of work I am doing on an upcoming web-based photo database. Enjoy!

..off my cuff… if it doesn’t make sense, then by all means let me know :-)

I’ll start off by quoting something I saw on Marc Canter’s site:

HHhmmmm – it seems like the world needs a standard for extending RSS with topics…. that are compatible with RDF and XTM. Bitzi could use it, Joi could proselytize it, we’ll build a front-end for it, Ross could analyze it’s social capital, and Doc could blog it. Amd Howard will report in from the Helsinki Burger King what the mob thinks. Another hot new meme. HHhmmmmmmma

This led me to check out TopicExchange.com, which has a page that makes a heroic attempt to corral Technology topics

I say “heroic” for a reason. I think we all benefit from seeing thinly sliced info-nuggets tucked away in categories. I also think there’s at least as many ways to define categories or topics as there are people. Indeed, over the course of a couple of minutes, I may come up with several different ways to think of the Sex Pistols:

• /music/punk/SexPistols
• /uk/england/1977/movements/punks/SexPistols
• /entrepreneurs/MalcolmMcLaren/SexPistols

and so on… For many people, the first example of /music/punk/SexPistols is the most obvious. But then, music is easy to categorize, compared to other topics with more dimensions (Film, for one…) Think of groups of people, and how they consume/peruse categories. As the group gets larger, and the number of categories increases, the potential for confusion (“oh, I wouldn’t have put English Cooking under Gourmet Cuisine”) skyrockets.

Which leads me back to what Marc said.

Categorization, topics, etc. where items are only assigned to one path are somewhat weakened in any context that extends further than a particular domain, clique, niche, or other form of “group”. So much depends on why someone is searching for something.. what is their context?

Context is everything.

Keywords are more flexible. Masses of keywords imply a multitude of relevant categories. General and Specific keywords give rise to the idea of self-organizing categories. Let’s go back to the Sex Pistols example. Let’s say I post a web page about the song “God Save The Queen”. I might use these keywords:

Queen, England, SexPistol, 1977, Punk, Music, MalcolmMcLaren

A number of paths spring out this grouping of keywords. Some of them even make sense! (1977/England/Queen/Punk would show SexPistol as a node)

To wind up, I am saying that getting large groups of people to agree on any fixed way of representing Topics or Categories is problematic. I am also saying that a method of taking groups of well-chosen keywords makes it much easier to find things by context. I would love to hear about a project that goes after the problem of “organizing category paths out of keywords”.

I blame it on dependencies.

I really do have a couple of blog entries in the works. They
depend on photos. Silly me, I installed Moveable Type without getting
ImageMagick sorted out. So
that’s my project this morning. Sort it out on the home box – make
sure it works – and then lather, rinse, and repeat on JavaJoint.com.

In other words, I got tired of looking at the “Do you
want to embed this photo, or make it a popup?” dialog.
I almost always want both…

Sometime in 2002, I mulled over the Spam problem and came up with an idea to thwart it. In a nutshell: you subscribe to an email-specific ISP, and every email you send through it gets a specific return address. You control when that return address expires, and who can respond to it. It’s not something you would use for all of your email; just for the ones that are outside of “the usual circle”…

This is the writeup I did on that. I don’t have any current plans
to pursue it, but thought it might be good to put it out there.

——————-
Private Mail – Daniel L. Smith – 2002

The Problem

Every time you email someone new, you widen the list
of people who have your permanent email address. Most
of the time, this is ok.

What about emailing someone you don’t know well, or a
company that may sell your email address?
(more…)

Some semi-random snippets about blogs I have observed, and
getting JavaJoint up to speed:

Small text… Lots of folks seem to like the idea of
cramming as much text on the screen as possible. I like using
a slightly larger point size. I suppose it’s a way
of being nice to folks whose eyes are strained enough
as it is. Hmmm, Squint or Scroll!?

Links Links Links.. There’s two things about links to other sites
that I am pondering… I don’t really want 50 of them on the main page,
and I am not sure how to categorize them. What do I mean?
With a lot of blogs, you’ll see a huge list of links to other
people. The problem is, no context.. The links on their own
don’t really say enough to make you want to click them.
I’m thinking I want loose categories that lead to another
web page/window. The other web page would have quick descriptions
of each link, or at least have them within a category that
makes sense.

…which brings me to my second bit about links.. categories…
I really balk at trying to categorize things sometimes. Many
links easily fit into a number of cubbyholes. What I will
most likely do is give it my best shot, but also look for
some sort of MT plugin / db app that lets you assign keywords
to each link, and then search on them. Can anyone point
me to an MT-run site that does keyword searches on external
links?

I try to keep myself a server-side developer most of the
time. Whenever I consider UI stuff, I always go through this
“clean, and functional – more layers, versus a little more cluttered
and a little more functional – one layer” internal debate.

An emerging theme I notice is this: Lots of everyday technology is
beneficial to the user, and disruptive to others. There’s a sort of
selfishness involved. Using technology with a sense of civility
or responsibilty is not common enough. Indeed, it seems
to be a worsening problem. Spam, mobile phones, dvd players,
subwoofers, telemarketing, and even GPS navigation systems
all weave into my examples…

(more…)

I’m in the midst of updating the look of this site. It’s nice to have a Linux box at home that I can subject to all manner of misguided CSS experiments…

The richer brown color that you see (at least on April 2nd, 2003…) comes directly from a photo of a coffee bean. Just wanted you to know that I strive for authenticity :-)

I was a little disappointed yesterday that there wasn’t much April Foolishness
on the web. I hope it’s just a “serious year 2003″ thing. One good one
though at oddtodd.com. Check it out.