52 replies to this topic

[mcanerin]

Even if they can't read it, I would guess that the fact that it's hidden from them would tell them there's a reason you don't want them seeing it.

If I were a programmer at G, I'd be pretty suspicious of being locked out of a css file. I'm not sure how I'd handle it, though. If you punish a site for it, and it turns out to be an accident, then it would be unfair. Also, some sites have multiple css files and may want to restrict access in order not to confuse anyone.

maybe what I would do is load and index the site as if there was no CSS file. This would remove the ability to stuff it full of keywords (well, you still could, but it wouldn't matter).

The negative side to that would be that then you may find the css being used for hidden text.

Tough call. Maybe I'd load it and render the page, but not index or cache the CSS itself. After all, compliance with the norobots is voluntary, so it's up to the SE to either ignore it, ignore the css, or ignore your site...

Or maybe pretend that no one would ever misuse css to trick a SE. Not that I'd expect that option...


Ian

[domokun]

Not quite sure what you mean about CSS replacing tables. Are you referring to Layers perhaps?

i think SE's will like your page more!
your pages will be smaller 'cause all your text formatting is off-page in an external style sheet
you can use page divs to encompass your content (no more nested tables!) and then use css to position it on the page
you could also place content nearer the head of the page, therefore increasing its relevance

as far as im concerned, css is the way forward!

[qwerty]

I think we need to go back to the quote and consider the source. The problem is that the source isn't completely clear to me:

Of the Google toolbar, Greg Boser strongly says "uninstall it!" Even though the toolbar offers a basic gauge of PageRank (and therefore link popularity), it's simply not worth the obsession if rankings are holding fairly steady and traffic has not dropped off.

When evaluating the weight of a link, Google wants to know as much as it can about it. Complicated issues such as Javascript, bad coding, image map links, and CSS hurt the process because these make Googlebot work too hard, which can lead to mistakes while spidering webpages. Link text puts sites into context, said Dulitz, and that's critical for determining places of importance related to a subject, even if a link doesn't contain the exact phrases associated with the web page it points to.

Now, I'm looking at the abbreviated, free version of the article, so I don't know what's been edited out. I'm assuming that because it's in the same paragraph, the mention of CSS came from Dulitz, who is a Google software engineer. But maybe the article is continuing to paraphrase Boser (from WebGuerilla). I think that if this statement came from Google is carries much more weight.

Are there any SEW members who can shed some light on this?

[dimok]

I strongly believe that SEs do not take CSS files. Some reasons, imho:

1. require lots of resources
2. the impact of CSS info in relevance supposely extremely small, i.e. that info gives no more info

So, try to do everything nice for user , but don't play a lot with CSS.

[Randy]

I would have to respectfully disagree Dimok.

If nothing else the SE's should be looking for visability: hidden in CSS files. And maybe layers which are the same size and absolutely positioned in same screen space.

I trust that the SE's know everything about servers I do and more. From the server admin side of things there is so much you can do with those that it's not even funny. So I have to assume they're at least looking for the non-server manipulation stuff. If not it would be very, very simple to cloak as much as you wanted.

[Mel]

Agreed Randy, but don't forget display: none & Position:-200px & the combination of Z index and position to hide things

[jbelle]

The only reason I can think of that spiders might be slowed by CSS is that the processing of the page must take into account how various browsers actually implement the CSS spec (and CSS being more powerful in its formatting ability than HTML formatting, is thus more complex). For instance, in IE 3, any font specified in em units appeared at 1px high, so that could have been used to make text virtually disappear.

But it's just a guess!

[joker]

I think that there is a huge problem with giving SE bots access to your external CSS files. That problem is I highly doubt that SE bots will never be able to determine whether or not someone is spamming a site using CSS without incurring a tremendous amount of "false positives".

I think Mel and Randy make an excellent point when they cite such tricks as setting the visibility to "none" or moving items way off the screen, etc, but the problem is, that these methods are not necessarily spam, and when used correctly, vastly enhance the interactivity, usability, and navigability of sites.