Articles written as informal essays benefit from not being tied up with orthogonal settings or context. This includes site names and organisational structure, and also style and design whenever these are particularly intrusive.
Site names are bad for one of two main reasons: they may tie things together which often ought to be left unbundled; or they may lead you to create content for a site, instead of a site around existing content.
Tying things together which ought to be left unbundled makes a site name irrelevant, and orthogonal to the content that it's supposed bind. This is often true of weblogs. If you write an excellent article about the history of teapots, for example, what does it matter that you wrote it for a website called Yearnings and Learnings? If the thing that binds the content together is the fact that you wrote it, then explain that on the page.
Content for sites instead of sites for content stifles creativity when writing. Content often grows organically. You may speculate into new areas and be creative. When you have a site which is always evolving, always being creative, then giving a solid scheme to the site is a hindrance. Calling a site History of Art, for example, is no good if you start to concentrate on a particular author, and then concentrate on a particular technique used by that author, and so on. Why pick a name for the site at all?
Therefore:
Organisational structure is bad because taxonomies change. The relevance of an article can change not only as it's being updated, but due to external factors. There is no navigation worth having which is orthogonal to the content.
Therefore, avoid navigation bars. This might seem like strange advice on a web full of navigation bars. But the information at the top of any hierarchy is likely to incur the problems associated with site names as mentioned above. Information at the bottom of a hierarchy, on the other hand, is likely to be linkable from within the content of the page. So for example, if a page is about the cubist period of Picasso, you don't need to put a link to your site's Picasso section in a navigation bar. The article is presumably going to mention Picasso early on: you can link to it then.
And does your site even need a Picasso section? You should always plan for a sticky slope. In other words, don't make a section for something that doesn't have much material. Don't create the site first and then the content.
The idea of branding is to create some kind of unity of style which is based on print publications. If you're selling a magazine, you want people to keep coming back to your magazine, to keep buying it, and you can only do that if you have not just good content but, of course, a name and style which people are going to remember and recognise in the shop. In short, content and a brand.
This is relevent to the web to a large extent too. Content on the web isn't generally for sale, but people will only subscribe, for example, to a syndicated feed if they think there's going to be consistently good content in it. Sometimes they might stumble across feeds at random, but word of mouth often helps too. People talk about Slashdot, so people get led to Slashdot. People talk about the Times Archive Weblog, so people get led there. Having a site identify would seem, on the face of it, to be very useful in this capacity.
Of course, a site identity only helps if you stick to it and keep it consistent. This is the hardest thing to manage on the web. This is something which I'm especially bad at doing, and as a result I'm seeking to minimise branding altogether.
When I was updating my homepage recently, I decided that it would be a good idea to point people at a handful of articles that I'd put online. The idea was to showcase the best of what I'd written on the web, so I set about reading some of my old articles and decided which ones to include. When I was done, a surprising thing became evident: my interests were much more consistent than my site names, organisations, and designs.
Part of the problem had been caused by the fact that I've had so many soapboxes on the web, and each of them of course has a different name and style as fashion dictates. I decided to look at the articles and identify what characteristics of them exactly were problematic, and this led to me isolating problems based on setting, and problems based on the content itself.