Google and the Underscore
A while ago I posted my WordPress Underscore Plugin to modify the slug treatment of permalinks to use the ‘_’ character in place of the ‘-‘. Some cried foul play as the SEO gods wouldn’t understand their URI’s, but Philip Nilsson pointed out something I felt was too important to leave buried in the comments.
Philip pointed me to a Google search for ‘URI underscore dash’, and the results to me are conclusive. I have included a screen shot, with the important area highlighted.
So it would seem that Google is able to understand ‘_’ as a separating character. I’m sure that Movable Type people will be happy about that. I know I am, as it would only make sense. I never did understand the stigma of the underscore, and why a progressive, forward thinking company like Google would overlook such a common technical character.
As to why I prefer it to the dash, I believe it comes from my programming roots. Going all the way back to programming in C, variable were always named with underscores; at least in the classes I took. Even in the WordPress core, all variable and function names are separated with the underscore, so to me, it only makes sense. I’m happy to finally see some proof in this matter.
Update
Well it appears we have a dichotomy. While the displayed results (e.g. the screenshot above) would appear to use the underscore as a separating character, to my continued astonishment Google’s inurl: function apparently does not. Unfortunately my gut tells me that Google still does not understand the underscore as a space in URI’s, and that the bolding on the results page is simply the result of a regular-expression search and replace.
Either way, Google is misrepresenting its results, and I believe keeping good search material hidden if it is indeed ignoring such a popular separating character. If anyone has any inside information as to how Google’s inurl: function works, or how the keywords are highlighted on the results page, that would be helpful.
Post and Author Info.
Published May 22, 2005 by:
Commenting is currently off for this post.
So far there are 5 comments.
I was really excited when I read this at first, but then I realized that your search query wasn’t for the underscored words-
“wordpress underscore plugin”-meaning that you are not actually getting a hit as a result of the underscored words in your URL. (Plus, your page title has no underscores, and people are linking to you with “wordpress underscore plugin” so your ranking is elevated)If you actually isolate the URL as the search target, the underscores do work against you. Try this google search: http://www.google.com/search?q=inurl%3Awordpress+inurl%3Aunderscore+inurl%3Aplugin—you get 4 pages that include ”/wordpress-plugin-underscore/” in their URL, and none with underscores.
Unfortunately, it appears as if the underscore is actually NOT recognized as a space, but that the URL is (obviously) only a part of the algorithm.
May 22nd, 2005
I’m quite pleased you posted this Steve, as I was beginning to wonder the same thing this week.
Our website at works gets spidered all the time by Google and recently I’ve made some changes to a FAQ section of the site. In this FAQ section of the site, I’ve got a list of questions which link to other pages which the answer, while the linked page file name is something like “faq_some_meaningful_file_name.asp”.
The strange thing I couldn’t work out, which I was beginning to think I was going nutty, was that it wasn’t being spidered and indexed very well at all.
This has now confirmed why it wasn’t working, so I’ve changed the filenames to a dash, instead of an underscore to see how we go.
Al.
May 26th, 2005
Intriguing post. I too picked up a fondness for underscores during my early programming days. But I learned to avoid underscores in host names during my days as a DNS admin for a university. I’ve also been aware of the SEO issues involved with underscores in URIs.
Since then, I have developed a deep dislike of underscores for other reasons and avoid them altogether, even in variable names and file names. I now prefer the hyphen instead.
The simple reason: the hyphen doesn’t require shift key wrist acrobatics. I can code faster and I put less wear on my wrists.
Yes, the underscore is slightly more human-readable than the hyphen. But the human brain does amazingly well at connecting the dots and filtering out noise. I also realize this goes against popular coding style but I’m sticking to my guns.
Well, to each his/her own. Keep up the good writing.
Regards,
Lindsay McLennan Web Developer
June 13th, 2005
I have created my site with all underscores in url…. now they all have PR rank… what can I do now? convert them to hypen and lose out all the PR? or stay with it?
October 23rd, 2005
Because it highlighted “underscore” when the search was done, that doesn’t exactly tell me that it now values the ”_” as a word separator. More likely, the engine found the term term “underscore” within the string of: “wordpress_underscore_plugin”
Meaning, it may not have mattered if there were underscores or any other character. It probably would still have found ther term “underscore” in these examples too:
“wordpressunderscoreplugin” “wordpress8underscore9plugin” “wordpressunderscoreplugin” “wordpressUunderscorePplugin”
October 28th, 2005