Silo-ing vs. Being as close to the TLD as Possible?

PhliipSEO, I disagree. Probably for different reason's than you think it's useless though.

It's only with the advent of CMS systems that it makes any sense at all to put everything at the root level, and even then it may not be the best in some situations.

I build static sites, usually by hand. I don't want 150 files in one directory. Using sub directories under the root serve two purposes. One it cuts the number of files down in one directory as I only have related files there, which makes maintenance easier. The second one is because I need subdirectories, I'm going to name them something useful.

Animals
Dogs (edit: dogs should be indented and beagles indented under dogs, the forum aligned them all to the left)
Beagles
Cats
Etc

It makes sense from a taxonomic standpoint and from a maintenance one. Everyone built silo's naturally when they built sites by hand. Otherwise after you got beyond a couple dozen pages, the insantity started to overtake you.

Also, the sub directories in the url can give the search engines a hint as to your subject matter.

Let's pretend for just a second that a person decided to be a contrarian. So to hell with the idea of getting an exact match domain name, building a tiny micro site on a popular cms leaving a nice footprint with your template and boilerplate and repeating that a few hundred times until you make $20 dollars a day that has become so common.

That's what the sheep are doing and sooner or later, sheep get slaughtered, so you decide maybe it's time to be different. So instead of all that you decide to go with a brandable domain. You pick bebibo.com or something similarly nonsensical. So your building out your site and it's to be a general info site you monetize primarily with adsense.

So at the root level you place an article about dogs and an article about needlepoint.

So is bebibo.com/dogs.html and bebibo.com/needlepoint.html going to help google and others understand what bebibo.com or your content is about better than if you were to use sub directories with names that relate to the topic of articles you'd find there such as bebibo.com/pets/dogs.html and bebibo.com/crafts/needlepoint.html

Your giving a search engine a little clue what kind of articles are going to be in the pets subdirectory and your giving your users a well laid out site.

In the grocery store you don't find deodorant next to the cantaloupe next to the charcoal starter fluid. I've no proof but I suspect users and search engines alike deem this useful as well.

One thing to avoid when using a cms is multiple categories and tags containing the same info. You would never do this on a static site, you'd need to go out of your way to break maintainability do do so. Yet wordpress and others make it easy to splatter duplicates of your "one" article in all manner of places, category, tag, author archive, data archive, etc. On top of that since tagging and categorizing is so easy, you go way overboard and have an article in three categories and eight tags. If you categorize with a cms be careful that you only allow one of them to be indexed, the one you want to show users, not all of them and let google pick one.

I see many people turning away from siloing because most cms's complicate the matter so much that it isn't worth the effort - because they fight you at every turn and in many cases require you to go into php template pages to really get what you want. For instance there is no reason on earth Wordpress should have ever gone with mysite.com/category/dogs over mystite.com/dogs. Sure you can kludge your way around that, but why should you have to. This is one of the main reasons I left database driven cms's behind for static sites that are easy to structure and build navigation exactly like I want without having to fight the cms and merge my changes into Wordpress and Wordpress theme upgrades, or count on plugins that performed poorly to get the results I wanted. So maybe the work required to silo a cms truly isn't worth the benefit, but that doesn't mean silo's have no benefit or that you can't benefit from one.

So in a long about way, no I don't think siloing is dead, in fact it's quite useful and has given me good results.