Welcome to WebmasterWorld Guest from 126.96.36.199
I had not thought about this, but apparently blue_widgets.html is considered 1 word, where as blue-widgets.html is considered 2 words. So for multi-word titles used as static HTML names, it looks like using a hyphen is far better than using an underscore. A brief look at the SERPS does confirm this.
Since then, DMOZ changed the way their breadcrumb trail is coded, so now the keywords that come from the url are not connected by underbars. That change made this point harder to prove (and the DMOZ pages easier to find), except that we have Matt's confirmation so there no longer should be any debate,.
Avoid spaces as these get changed to %20 which makes things%20very%20hard%20to%20read%20indeed.
Hyphens are a good idea, but dots, commas, and colons work just as well. In fact, visually, I prefer dots, like: www.domain.com/folder/sub.folder/this.page.html.
When I asked on here a while ago it was suggested that:
wasn't as good as:
I bought the later, although I prefered the former based on their opinions.
(Both keywords described my site well so I didn't consider it spamming)
It seems like decisions on here are extrememly split on the item of using hypens in domains.
They are not seen as being separate words by search engines
then why are they bolded in the serps im looking at right now?
Google changed their parser some time ago. But Google is not the only search engine, and may not remain the dominant search engine forever (AltaVista was king just a few years ago, and where are they now?)
I thought hyphens in a domain name was a BIG no no?
a-long-keyword-stuffed-domain-name-with-many-hyphens.com would surely send up a flag if your site was a candidate for a hand review. Not very easy to type-in, either.
There is no single correct answer to these questions, but you can either try to stay in the safe zone by using hyphens rather than underscores in URLs because they have been handled correctly by all major search engines for a much longer time, and limiting your use of hyphens in domain names to one or two, or you can take risks by going outside of the lines. The answer is different depending on whether you have a branded domain name that you want to use forever, or a large collection of 'disposable' domain names, with no need for branding or long-term survival.
Here is a very relevant thread which Matt posted on his own blog at August 25, 2005, which you might find of interest.
[edited by: Brett_Tabke at 12:47 am (utc) on April 23, 2006]
[edit reason] please - no blog links and absolutly NO copied in content on this system. [/edit]
why are they bolded in the serps im looking at right now
I've looked hard at that question. In my view, the bolding on the SERPs page is a character matching routine, added as a final layer at page creation time to help users see their keywords in the results. But the fact that a character string is bold is not tied to the scoring algorithm itself.
There are also some purely user friendly reasons I don't use underbars in page names. For instance, if the link is underlined, the underbar character can be hard to notice. Also it can be nearly impossible to communicate over the phone. I like keeping things simple, and "dash" is simple.
Parallel issues can come with the tilde [webmasterworld.com].
But that does not change my view that it is easier to read a filename with underscores (because it looks more like spaces), and IMHO think the search engines should look at filenames and read them in a fashion that searchers do.
Unfortunately, common sense does not always go along with intelligence.
Back to watching
I get the default Wikipedia but as second result I get eclipse, which doesn't even have the word Main Page on the page nor in the source nor in the document in the cache.
then several other big sites obviously given a hand coded PR, none featuring Main Page
Only on page two I start having pages actually containing the words close together.
Serious hand PR delivery there...
On a search for Main_Page I get the expected list of Wikis.
Pretty clear demonstration of above.
Do a search for "googlebombing" -- that will neatly explain your results. You can get a page ranked highly for "whales," even though it's about albatrosses, simply by using the word "whales" in the link text of a lot of links that point to the albatross page. Those pages you found probably have a lot of links pointing to them with link text like "MTV's Main Page." (This subject is off-topic to the thread at hand, so you might want to start a new one on link text if you have questions).
> Upon searching using mykeyword_mykeyword I find that the page is ranking fine.
Assuming that you are searching for pages whose URLs contain mykeyword_mykeyword, that's no surprise in light of the info posted above.
> Should I change the underscore page names from mykeyword_mykeyword.html to mykeyword-mykeyword.html?
> And what about the external links using the underscore version?
Get as amny updated as you can, and...
> Is this where a 301 is used?
Yes. 301-redirect the old URLs to the new.
> Or should I just create another page, leave the old page on the server, and change the navigation to point at the new page? (probably not, since it would be duplicate content).
The old pages will become inaccessible once you 301 their URLs to the new pages, so take them down or leave them -- it won't matter to search engines or visitors.
In my view, the bolding on the SERPs page is a character matching routine, added as a final layer at page creation time to help users see their keywords in the results. But the fact that a character string is bold is not tied to the scoring algorithm itself.
Can't this process then be done on the way in, not just the way out? If they can character match for display purposes why can't they do so for ranking and scoring?
In other words if there's evidence that they can discern the use of underscores in text, why wouldn't they attempt to fold this into their ranking to reflect the way actual humans use it (i.e. interchangeable with other punctuation like dashes)?
I am debating on changing some of the filenames on my other site, but that would involve about 300 or so web pages that might do better if I rename them, replacing the underscores with periods. I do pretty well in my niche, and hate to risk loosing our current rankings.
So my first question is, how many 301 redirects in .htaccess is too many. All I have in it currently is the lines for the non-www to www redirects. This site(sells widgets for my brick & mortor store) gets about 5,000 uniques a day. (about 1600 referrals from G A day, 230 from Y & about 100 from MSN, most of the rest are from people coming in directly). How much additional load, or bandwidth would be used by approx 300 additional lines of 301 redirects.
Second question, is it even wise to rename pages that top the serps for the targeted keywords for those pages, or should I proceed in the future with the dotted naming convention as I create new pages with new products? I have always been a fan of the saying, "if it ain't broke, don't fix it"
Back to Watching
RedirectMatch permanent /useless.old.directory.still.in.searchengine(.*)$ http://www.example.com/
I also use a webserver to do the example.com and www.example.com redirect and all it does is.
RedirectMatch permanent ^(.*)$ www.example.com$1
I have none of what people here call, I think, canonical issues.
I have about convinced myself to leave well enough alone, and just create all new pages with the dotted naming convention as I move forward.
Back to watching
RedirectMatch is an Apache mod_alias directive. As such, there is no need to use mod_rewrite's RewriteEngine on directive with it.
And if you do use mod_rewrite directives, then a single RewriteEngine on at the top of the file will suffice.
Back on-topic, the use of keyword-in-URL is a small factor in ranking a site for those keywords. If your pages are already on top, and if you already have many on-topic links with those keywords in the link-text, don't bother to change existing pages.
I wonder if the dot technique any noticeable affect on ranking