McGehee - 15 May 2008 01:19 PM

But… but it’s not good manners to paste raw URLs onto a web page…!

Perhaps I should also add that my default CSS is to have links underlined only on mouseover, so that even a raw URL appearing on one of my sites would show underscores where they exist.

YMMV.

Sue Crocker - 15 May 2008 03:16 PM

You guys do realize the program is called ExpressionEngine, no spaces.

Sure, but as Andy Harris said, it is just a useful word to use in this instance. You can prove it by using any two words you like really.

Chances are, you all read this in a blog post by Matt Cutts, a software engineer at Google.

http://www.mattcutts.com/blog/dashes-vs-underscores/

I’ve always found I can get pretty good results using Google or any other search-engine by looking for content—where dashes or underscores generally aren’t a problem.

I had always used underscores, or even camelCaps, but recently started going with dashs.
They don’t require the shift key #1, and #2 I had a strange thing happen with underscores on my web host.
They started showing up as % 5 F had to add the spaces or it gets parsed to a _ here lol.
http://crssp.com/demo/internet % 5 F sharing/

I don’t think I’ve ever noticed any other time an underscore being converted to % 5 F.

I sent a support ticket to my host to see if they could enlighten me on why it would be happening, weird huh?!? Actually I’m having troubles recreating the error, if you’ld call it that.
All i know is I don’t have URL rewrite, and if every underscore were to be indexed on google as % 5F that would be a problem.

Jesse B. - 15 May 2008 09:08 PM

Chances are, you all read this in a blog post by Matt Cutts, a software engineer at Google.

http://www.mattcutts.com/blog/dashes-vs-underscores/

Thanks for pointing me back! It’s been a long time, I just couldn’t remember where it was I read it in the first place.

Well the article is, what, 3 years old? As McGehee said, good content is the main point.

Yep definitely good content and inbound links, they’re the two that will get you where you want to be.

Ingmar Greil - 16 May 2008 10:31 AM

Well the article is, what, 3 years old? As McGehee said, good content is the main point.

Yes, but the question was “Dashes or underscores in the URL?”, so to that end just doing the Google searches with a relevant two-worded search will prove it is still current practice to use hyphens in the URL over underscores.

Good content will, of course, keep the GoogleBot coming back for more to eat.

Those search examples with expression_engine vs. expression-engine are faulty, as that’s the wrong side of the coin.  The point isn’t about how Google treats dashes vs. underscores in search terms but in the URL.  So only the ‘expression engine’ search is relevant, and I think the search rankings of our site, EngineHosting, and thousands of other EE-powered sites using underscore separators show that what’s used in the URL is rather insignificant, if it matters at all.  It’s silly to assume that Google’s search algorithm both isn’t smart enough to handle both equally well, and that it hasn’t changed in three years.  So, I’m with Simon.  Semantics is king.

As stated many times before, your content, the relevance of your content in context to places that link to you, consistency in how you use and spell certain words in your content even. Google’s ranking is highly complex and something like the - vs _ in a URL is such a very very small fraction of the puzzle its not really worth fretting over, really.

Take a look at the URL of the indented entry under spot #1 after following the first link below.  You will notice that web and hosting are highlighted, and not the _ you will also notice further down the page there are sites with - in the URL between web and hosting did not help much, but you will get from the first example with the _ that its treated as a word separator as the words are highlighted separately.  Its the whole package, and how you present things on your site, and how others present links to you from their own articles, pages, etc.  Its really a cat and mouse game to be honest

Google Search: load balanced web hosting

Oh and for the folks that will say you can’t tell if Google is treating it as two words or a two word phrase even from this first link.  Try this:

Google Search #2: load balanced hosting web 

Google’s ranking is highly complex and something like the - vs _ in a URL is such a very very small fraction of the puzzle its not really worth fretting over, really.

I think you give Google too much credit. 

I don’t really know either way because nobody here really knows how Google works.  You can argue either way about URL’s but the bottom line is you probably aren’t going to get a first page ranking on a keyword which solely exists in your URL.

I would say the Matt Cutts blog post gives me the teensy weensy, itsy bitsy push I need to use dashes rather than underscores.  I don’t buy the argument on URL readability either.  URL readability went out the window a long time ago for me.  There are just too many situations where the URL is used to shuttle non human friendly information for me to even take the effort to raise my eyes enough to look.

Unless I see something clearly stated in the Google webmaster guidelines, then I would stick with the Matt Cutts article. 

Those search examples with expression_engine vs. expression-engine are faulty, as that’s the wrong side of the coin.  The point isn’t about how Google treats dashes vs. underscores in search terms but in the URL.

Sure, but what if Google treats them the same in the URL or in the content.  Maybe this sounds stupid and Google should be better but I don’t have as much faith in the intelligence of Google as everyone else does.  Google search is still just a crappy web app just like anything else out there and search is still in its infancy.

If you guys want to keep your searches strictly to URL’s then why not use the URL operator?

Try doing a search for this…

inurl:asdf-asdf

vs this…

inurl:asdf_asdf

Notice that for “inurl:asdf-asdf” you get the following…

asdf+asdf
asdf/asdf
asdfasdf
asdf-asdf

If you do a search for “inurl:asdf asdf” (with the space) then you also get results for asdf-asdf.

But if you do a search for inurl:asdf_asdf then all you get is the underscore versions.

My conclusion…

Google is stupid.  You all put too much faith into them.  IF (IF!!) you lived in a world where you could only rely on your URL’s for search results and you want to capture the widest audience then use dashes.

EDIT:  Though I do agree that the URL is not as important as the content.  As Nevin says, it is just another piece in the puzzle.

John, you’re still looking at search terms, and going down the same incorrect path.  It’s a leap of logic, or a “what if” scenario that amounts to nothing more than FUD.  Just because Google acts in a certain manner when it perceives that human is searching for an underscore has absolutely no bearing on whether or not it does or doesn’t apply higher or lower rankings to content with - vs. _ in the URLs.

Having read this lot, I think it’s really more of a personal preference thing as far as the URLs go.I’ll stick with the dash because it’s what I’m used to.

Has anyone ever seen a domain name with an underscore? I’ve seen plenty with dashes (jambor-ee.com and train-ee.com to name two).

In domain names, only lowercase a-z, dots (.), and hyphen (-) are allowed per rfc1738.

No, but that’s because letters, numbers and the dash are the only legal characters to be used in domain names. There are some local deviations of this rule (like, must not start with a number, or things like that), but underscores are generally not allowed.

Ah I was going to ask if they were even allowed but assumed they were as they feature elsewhere in the URL. Glad I posted this, have got a lot of useful info!

Actually I was looking at the scheme naming rules and not the host.  I thought plus looked funny in there, and indeed, it’s just letters, numbers, dash, dot, and the dash cannot begin or end your domain name.