Forum Moderators: open
or just: john.htm
?
[edited by: ciml at 10:07 am (utc) on Oct. 19, 2004]
[edit reason] Examplified [/edit]
5 good reasons to go with absolute URLs:www.searchengineworld.com/newsletter/1999/q3/absolute.htm
Come on now, that page is from 1999!
From the link noted:
Drop into Google some time and do some searches. Click on the cached option. Notice how links are presented. If you use Absolute addressing, the cached page points to fresh content on your site. If you use relative addressing, the link points to the stale Google cached page.
Not! I just tested by adding new text to a page then using the cached link(relative) and the new text it there.
If your site has good content on it that people will want to save and read at a later date, there is nothing worse than not having a buried absolute (easily identifiable) link that points to your site. (ie: if it is all relative addressing and they load it up off of disk later, they can't find your site because your links now point to their disk instead of your site). A major mistake in my humble opinion. Always have a prominent link pointing at your home page on every page near the top (a logo embedded href on every page is becoming standard).
Just tested in IE6. All relative URIs were auto converted to absolute! Plus, that's what the <base> element [w3.org] is for as well.
Your page is designed to be printed. Put that absolute url in there. When a user hands the printed document around the office, you *want/need* that url to stick out there. Some browsers will print the relative address instead of an absolute one!
Once again, I'm only testing with IE but I get full URI when printed.
If someone steals your content, and you are using full absolute addressing, they are forced to edit the page. Minor, but anything you can to do to slow down a page thieve is worth it. Its surprising how many don't know html. With relative addressing and they just dl it and the graphics, install on their server and go.
What! If it's all absolute addressing I simply download the html and all the graphics will be pulled from the originators site. Either way, you can't stop the thieves so this point is moot.
My feeling is that the absolute links make it into the link counting algos of a couple of (remaining nameless) search engines. Part of the algo has to do with different directories. If you have [full_url...] in directory A that points to directory B, then the link is counted in citation counting. Same is true for a link in dir B pointing at Dir A. So, you get two link credits for one site. Trivial, but if you have a few dozen directories, relevancy can be increased for your domain.
I definately do not agree here but this point is arguable and hard to prove. I'm sure there are many who will say they only use relative linking(me for one) and have good results.
It's an urban legend if you ask me. One reason to NOT use absolute is:
CODE BLOAT
Sorry for ranting but this topic keeps coming up and I believe strongly that relative is just fine.
Birdman
I guess it could be somebody performing some dumb ass requests to the server, but it doesn't look like that. it looks like a regular page being concatenated onto another page uri.
It was not an exact test since we launched two new sites, one with absolute and one with relative. Bottom line: IF there is a difference to the SE's anymore, we sure don't see it. Pages doing great (or not) either way, but comparable pages on those two sites have never shown any pattern at all of being influenced by this issue. We test a lot, so it was good enough for us. We go with relative for the code bloat reason. Conversion is important on our sites so we hate losing visitors because pages load too slowly.
However, a related word. What is far more important, in our view, is that you use one convention through your site (and also with external links, to the degree that you can influence those), and stick with it.
So, for example, always link back to the homepage the same way (".com/," or ".com," or "com/index.html," or "../" or whatever). Pick one; stick with it. A consistent approach is more important than whether or not the path is relative or not, at least as far as we've ever been able to see.
Still I am interesting what you think:
Is it better to use tables or layers for website or something other?
(Of course which gives better pr)
<a href="http;//www.mygoodolddomain.com/somefile1.php">
<a href="http;//www.mygoodolddomain.com/somefile2.php">
<a href="http;//www.mygoodolddomain.com/somefile3.php">
<a href="http;//www.mygoodolddomain.com/somefile4.php">
<a href="http;//www.mygoodolddomain.com/somefile5.php">
Better(IMO):
<a href="/somefile1.php">
<a href="/somefile2.php">
<a href="/somefile3.php">
<a href="/somefile4.php">
<a href="/somefile5.php">
Depending on the amount of links and images your pages have, it can really add up.
Layers vs divs:
Shouldn't matter as long as your HTML is valid(no errors). But try to stay away from alot of nested tables, as they will cause bloat too.
Birdman
One caveat that comes to mind is that if you have several domains serving the same content, absolute URIs may avoid some of the problems of unnecessary re-spidering and confusion over duplicate content.
But what will be the result by using only layers, not div (tables) at all?
Does somebody of you use only layers instead of tables? And which result he have with pr
I personally use layers and didn't have any problem, but I am not sure if it affects my current pr ... People often use tables, and I didn't see yet any website which uses layers (except mine) :-)
With relative links bots have to calculate where in the structure they are, which can introduce an error. :( With absolute links there is less chance of an crawl error occurring. :)
BTW, I emailed help at Google and asked them to add this topic to their FAQ. Hopefully they will :)
ht*p://www.abc.co.uk/123-wer.php
The page above has a relative link to another page like this:
567-uio.php
We found that Google started requesting URIs like this:
ht*p://www.abc.co.uk/123-wer.php/567-uio.php
What is worse, when I found this entry in our access log I found that our web server would actually serve a page due to some of Apache's special features (content negotiation, AcceptPathInfo etc). It should have given a 404.
Now this next part could be a coincidence, but this site was PR5 doing a few thousand UV/day, but it got hit with a PR0 and fell by about 80% in traffic, just about all Google traffic was wiped out. Wouldn't be surprised if its because Google basically thought the whole site was duplicated - once they picked up just one bad URI, then because all links are relative every other link would now be in the bad format giving what would appear to be thousands of duplicate pages.
I've fixed this now so we'll se if the site gets added back in.
So, if you're using relative links, make sure your servers setup properly as this could hurt you bad. Come to think of it, make sure your servers setup properly anyway or a competitor could just link to a dodgy URI and convince Google your whole site is duped.
The page above has a relative link to another page like this:567-uio.php
I use /567-uio.php
The wider issue has to do with whether there's a correlation between build quality and search performance.
By build quality I mean things like:
Pages that validate.
Pages that conform to accessibility standards.
Separation of content and presentation (using CSS).
Not using tables for layout.
Lean pages that load quickly.
JavaScript put in external files.
Many people argue there is a correlation, but I haven't personally observed much hard evidence so far. The case for build quality, of course, stands on its own.
It looks like a regular page being concatenated onto another page uri.
This occurs frequently with URI rewrites that are not set up properly, especially with IIS hosted sites. I know, I was the unfortunate recipient of a bad spidering using relative URIs. That was a couple of years ago, I've not used relative since. Although I will admit I was forced to in a recent build due to the setup of the shopping cart. But, I am confident we will not have the concatenated URI issues again.
I feel it is good practice to at least use absolute URIs for all main navigation links. And you know what, I believe there still is the issue of someone hosting your content on their domain in a completely relative environment. In fact, I know there is. I've had someone show me first hand how they could easily hijack a domain that was using a relative linking structure.
The question was:
"Does an Absolute Link Help in Google?"
The answer is NO. (IMO, of course)
I'd say certainly possible for SEs to get messed up using relative links. Take this example:ht*p://www.abc.co.uk/123-wer.php
The page above has a relative link to another page like this:
567-uio.php
We found that Google started requesting URIs like this:
ht*p://www.abc.co.uk/123-wer.php/567-uio.php
Robber, are you using mod_rewrite?
To answer the question directly though. I can see that there might be some benefit from keyword density in using absolute links, but it would only count as a standard bit of text, you could therefore structure your site to use relative links that still include these keywords if you really think its a big deal. Personally I'm not sure its worth it.
However, if its true that relative links can help Google (mistakenly) find a duplicate version of a site then that is certainly not going to help your rankings. But as mentioned previously, even if you use absolute links, if your server isn't right soeone else could just point to a dodgy URL and you'd get hit anyway.
Is it important (bad or good) if you have a lot of anchor text (linked text) on the home page and other pages of the website?
Is it important (bad or good) if you have a lot of anchor text (linked text) on the home page and other pages of the website?
Sure! I believe it adds some benefit to the pages being linked to.
<a href="/green_widgets.html">buy green widgets</a>
vs
to buy green widgets, <a href="/green_widgets">click here</a>
#1 is much better and should give the page some boost for those words. Of course, it's just an opinion :)
Sorry for the 'me too' post ... but agree that this is the one and only difference that we have seen with G.
I might add, however, that we *think* we've seen one or two cases where Y! had more trouble with relative links. The thing is, their coverage is so imcompete/spotty anyway it's hard to know. A couple of our sites (with absolute paths) do seem to be more fully indexed in Y, to the tune of about 20%-30%. Again though, could be coincidence.
[webmasterworld.com...]
the GoogleGuy post
A useful reminder.
Regarding Sim Spider, on occasions in the past (not today) I found that in the case of relative links to files (or other folders) in that folder it would add the link filename directly to linking page URL, omitting the forward slash and making the link unspiderable. I looked at a number of SEO sites and found some who added a forward slash in front of a relative link filename: /filename.htm so I started doing the same.
looked at a number of SEO sites and found some who added a forward slash in front of a relative link filename: /filename.htm so I started doing the same
However, if you had a subfolder "http://mydomain.com/folder/" and wanted to reference a page in the root, you could put "/mypage.htm" or "../mypage.htm".
Or if you wanted to reference a root document from 3 folders deep "http://mydomain.com/folder1/folder2/folder3/" you could use "../../../../mypage.htm" or more simply "/mypage.htm".
Richard
