Nope, it's because Microsoft just don't care. Wikipedia has 2.5 million articles all in the same namespace and still manage to have nice URLs. It really doesn't take much effort to provide decent URLs for a very large corpus of information - but Microsoft clearly haven't even thought about the problem. Their URLs have always been disgracefully bad.

This is more a disagreement about the value of URLs than about whether Microsoft "cares" about their end users. With a keyword bookmark for Wikipedia, I'm about 50/50 for getting the content I want with simple URLs; I still wind up in the Wikipedia search bar all the time.

Nobody has a foolproof scheme.

For me it's not about finding the information by typing in a URL. It's more about scanning and having a clue what I'm going to see by the url I'm clicking on. A giant random number really has no meaning or value in this context.

For me, it's pretty irrelevant because I'm a developer using a development knowledge base (MSDN) and have been for over a decade.

By your argument, HN is a fail too, yet you still click on the links.

HN isn't a reference work like Wikipedia or the MSDN Knowledge Base, so it's less important here.

But the real reason that Microsoft's URLs suck is that the MSDN Knowledge Base is an older product than Google's, built back when the web was new and people didn't realize that their URLs sucked. And one thing that I respect Microsoft for is actually caring about reverse compatibility.

(Sidenote: This is why you should think about your URLs. They are how your website will be presented to the world.)

I didn't mean they don't care about their end users, I meant they don't care about their URLs.

I think the links earlier in this thread are a perfect example of why this is useful. You can tell what you'll get when you click on the Yahoo! and Google ones, but the Microsoft one is totally opaque.

What enterprise level CMS or KM are you aware of that generates semantic URLs?

That's kind of a loaded question since a reasonable definition of an enterprise CMS is "one that isn't very good". That said, Ellington (which I worked on a few years ago) is probably expensive enough that it can be considered "enterprise" and ships with nice URLs out of the box: http://www.ellingtoncms.com/

> definition of an enterprise CMS

Would Jonathan Schwartz's blog ( http://blogs.sun.com/jonathan/ ) count as an enterprise CMS? That has URLs which include the title. Does Wordpress ( http://www.wordpress.com/ ) count as an enterprise CMS? That also has URLs which include the title.

Jonathan's blog? Maybe. Wordpress? No. It's pretty clearly open-source and has the "started as a few lines of code" pedigree that makes it not enterprise.

Wikimedia?

Yahoo has lots of documentation organized in a single library. That's what the "developer.yahoo.com" bit is.

Yahoo's is just better organized.