More on Blog Conversations

I just found Richi Jennings' href="http://richi.co.uk/blog/2005/02/blogs-as-conversations-yeah-right.html">post
about blogs as conversations in a Technorati search, and he's got me thinking
about this again. He writes:

Twaddle. What we have right now, is a messy prototype.
Comments are too prone to comment spam, most people find it hard to understand
the concept of trackbacks, and the whole thing is just too damned
difficult.

In thinking about it more, I realize that the
problem is that most blogging software is browser-based and it's really,
really hard to write web pages such that clicking on something (like an orange
RSS chiclet) pushes something useful upstream to the browser. In fact browsers
go to a lot of trouble to make sure this isn't possible for the most part,
because that "something useful" could well be a virus or something awful. So we
have abominations like bookmarklets that you're supposed to drag onto
your tool bar. Right. Great. Just because it works at all, people are slapping
each other on the back in congratulations. Not so fast, I say...

How do we allow people to write on specific parts of the web without (a)
spam, and (b) permissions problems from hell? This is the crux of the issue.

Sorry, Richi, I'm still thinking about it :)

The case against XML

If any of you happened to see the article in the Tech Monday section of the San Jose Mercury News about Bubbler, you may have seen me quoted as saying that XML is one of the most over-rated phenomena. I've gotten quite a few questions about that :)



Here's why I think XML is overrated.



XML is not really a technology. It's a syntax for tagging bits of data. It's not all that advanced in its evolution, harking back to SGML from some 20 years ago, which was actually better in many ways.



XML is syntax without any semantics. That's a big piece of what's wrong with it. Any program parsing an XML file still needs to know exactly what's in it in order to make sense of it (witness 5 different "flavors" of RSS).



But there are a lot of little things, too. To get an idea of what's wrong with XML, just open up any RSS file and take a look. The first thing you see is that XML does not nest worth a damn. You can't embed XML (or even HTML) within an XML file without escaping every last character, so it looks like &lt;tagname&gt; instead of <tagname> (it was hard enough to type this in and escape it correctly; I can't imagine what it looks like in an RSS feed :)



It doesn't stop there. The open and close delimiters are clunky, though I guess you can't blame XML for that as HTML has it too. The syntactic waste can double the file size easily. And parsing XML files requires reading the whole file before you can start to actually make sense of it. Audio and Video started this way, and guess what, it didn't work very well when transmitted over networks, so people worked really hard to coming up with streaming audio and video just to get around this shortcoming. Where's the streaming XML format? I rest my case.



XML is verbose, inefficient, suffers from whole-file syndrome, is difficult to parse correctly, has nesting/escaping problems to the nightmare degree ... and yet everybody is always raving about how great it is. I think it's overrated at best.