Forum Moderators: open
How can Google Reduce it?
How about this. There has been talk of G reading CSS files and I have no doubt they have the technical ability to interpret the contents. Enough to know when H1 has been made to look like p text and divs are hidden etc etc.
I know that to check every page in the index might well melt their servers, so introduce random sampling. Check the odd page here and there and if the flag is raised, by finding naughtiness, then investigate further. Upon something being found a warning shot can be fired using the method of reducing the ranking for that page or site. No need for a warning email, the ranking will do the job.
I can almost guarantee the only people who might moan about this would be the people employing such tactics. There of course will be people who might say that unscrupulous SEO’s have screwed them over, well my opinion on that one is ignorance is no defence. Rather like acting on dodgy tax or legal advice quite often you yourself carry the can.
The upshot would be less people moaning about being beaten in the SERPs by hidden text etc. :)
Right…..back to work
How many ways could text be hidden, here's a CSS neophyte list
- overlapping divs/elements
- divs with same colour as text
- text.hidden!
- unreadably small font-size
- before i put thread off course, here's what i found [webmasterworld.com], looks like a list.
While these subtle differences could be misinterpreted as a SEO trick, I think Google is more concerned about "policing" factors that weigh more heavily in the algo, such as links and anchor text.
but grabbing font-sizes seems easy enough and useful, though finding hidden text sounds a bit (far!) more complex.
why is finding font-sizes easy?
hint: apply the same knowledge to your first statement as you did to the second and remember the CSS stands for "cascading" style sheets
PS ukG I'm sorry for following this off topic.... but no matter what you are good/bad at .. don't let it spoil your wints (no mispell it's a scottish word)
Suzy
PS: that list... well it depends on what perspective you look at it from. ... Do they know, want to know, couldn't care less, make excuses, ignore, think they know better?... delete as applicable...
A succession of stages, processes, operations, or units.? :)
>font-sizes
preg_match("'font\-size: ([^;]+)'",$css_element,$match);
Is there only one CSS "rule" for font sizes, i.e. font-size? Parse the HTML line by line, looking for references to the style sheet, keeping note of relative font sizes. Apply "prominence" relative to "1".
Seems easier than looking for hidden text, that's why I believe looking for hidden text would be more difficult.
preg_match("'font\-size: ([^;]+)'",$css_element,$match);
That's not CSS, ...well it might be be in some other programming language but you have to understand CSS first.. right?
>>Is there only one CSS "rule" for font sizes, i.e. font-size?
ermmmm. no!
lol, OK.
You're right, that's not CSS, it's regex to match some CSS.
So what other CSS commands would adjust the font-size, I skimmed through the a-z in topstyle and couldnt pick anything out, mostly ones that altered the emphasis of the font.
>understand CSS
I'm sure google would have that sort of knowledge at their disposal ;)
You're right, that's not CSS, it's regex to match some CSS.
OK that means you've already used some of your other knowledge .. I knew it was regular expressions too... but that would be another thread topic wouldn't it?..
apparently this is a discussion about CSS.. not about programming techniques/languages/etc...
So what other CSS commands would adjust the font-size, I skimmed through the a-z in topstyle and couldnt pick anything out, mostly ones that altered the emphasis of the font.
Stick with Topstyle they'll be right ;)
>understand CSSI'm sure google would have that sort of knowledge at their disposal
Sure they do, that's why they can work out when HTML text is small/same coloured/off the page etc... or perhaps they use Topstyle?
you know your stuff, Topstyle knows theirs , google knows theirs .....
Suzy
{oh and Google probably knows Topstyles...seeing as how it's a programme, i.e logical... algorithimic...etc...)
you know best, I'll get a refund on my CSS editor, since it doesnt seem to have as much CSS know it all as you ;).
Now what I'm saying in direct answer to your question is that yes you are right ~ font-size in the CSS is the logical way to read the size of your text...
but if that is (e.g.) percentages then what is it relative to.?
and by the time you are half way down (the cascade) a stylesheet are you still specifying your text in relation to the original(root) element or are you now using positioning so that it is relative to something (anything) else?
NO... I'm not trying to be difficult.. but I'm trying to say (or trying to make a feeble point ;)) that this is not as easy as reading HTML tags! ~ Which we all agree (I think) has not even been fully mastered yet!
Suzy
Display:none
font-size: <whatever small thing you want>
font-color: <matches something>
z-index: <behind something else>
I've probably messed up one or two of those, but that's just off the top of my head. There are probably another dozen ways to do it using legit CSS. Then if you throw in the same CSS in javascript, external CSS, cascades, etc... and you quickly realize that there isn't any percentage in Google trying to specially parse stuff for hidden text.
If you're Google, you hit the obvious stuff in the html you already have, then ignore the rest of it. With off-page factors being more important anyway, it doesn't really matter that much.
BOL
If I knew Google might be considering this sort of thing and I was in a mindset to give it a whirl with some domains I did not really care about there are few things I would do to make life as difficult as possible.
Multiple CSS sheets, mixing units and methods (em, px etc). I am sure with some thought it would be easy to come up with a few suggestions.
>> don't let it spoil your wints
What in the name of Brett Tabke are wints. I have some Scottish pals who might help but the fear of what it means stops me from asking. For all I know it could be a proposal of marriage :). Is it?
There's no need to parse through the CSS. Text has properties after the HTML/CSS code is interpreted by the browser. That's why we can write client-side JavaScript that can change text.hidden=false to text.hidden=true or if text.size>2 then text.size=1.
Of course, I'm not sure if Google would spend the time or processing power to interpet every page's code as loaded through a browser instead of just parsing text on the page. If it isn't very system intensive then it wouldn't be that hard to do. But if so, I think it would be a lost cause to parse through the text.
e.g. someone submits a spamreport => automatic processing of that page through IE (roughly said, of course there need to be some checks against abuse, DOS, etc).
It would be sufficient to process it through IE on demand.
If you change IE to browser, that is exactly what they were doing. They are no friend of microsoft, so they are not likely to get the source code from MS to modify. I would bet on mozilla, konqueror or gecko.
>> While these subtle differences could be misinterpreted as a SEO trick, I think Google is more concerned about "policing" factors that weigh more heavily in the algo, such as links and anchor text. <<
Google can tell the difference between this:
<h3> The Heading Here </h3>
<p> The long paragraph of content under the header goes in here </p>
and this:
<p> Some content stuffed with <h3> keyword keyword keyword </h3> in the middle </p>
The former is normal HTML content, and should rank well.
The latter should have some sort of penalty or should have the <h3> content ignored.
The styling of both is irrelevant.
