I hope that I can explain this correctly. I have been noticing, for some time, that Google is constantly picking up actual codes that are used on pages and inserting them as snippets under search results returned to my site. One of the biggest instances of this are search results to forum threads.
For instance, a recent google.com search for oxygenated water resulted in one of our forum threads being the number one result: http://www.gustrength.com/forum/t-262899/oxygenated-water
However, the snippet that google used, as often happens with forum results on google searches, looked like this:
Unfold Re: Oxygenated Water by EricT, 1283663096|%e %b %Y, %H:%M %Z| agohover. Fold. Oxygenated Water and …
I haven't seen google show this behavior with other forums and I can't figure why google is continually picking up the date code, etc. Is there something that can be changed to prevent this from happening? I would much rather actual relevant page content be returned as a snippet and I cannot put meta data on individual forum threads.
Another similar problem is that when html code blocks are used, google seems to see this as page content and returns the actual codes, sometimes, as snippets under search results. Doesn't seem like to me this should be happening but I do not know enough about anything to say why or how it could. Is this a "bug" or at least an issue that can be resolved in some way? Potential readers seening snippets like the above or some random code from a page is a bit embarrassing to me, to say the least. It certainly is not going to help my traffic.
Update. The problem with google showing snippets of html code may only be happening with the older method of using html on a page with iframes, not with the new html blocks. So that is easily solved. I have a lot of pages and haven't been able to change them all to html blocks, as needed.
This is not Google picking up HTML code, but rather Wikidot uses some constructs that embeds data inside HTML tags to be rendered on the browser. One of the examples is how we format dates.
E.g., we generate the following HTML:
Which is translated in the browser, using a JavaScript function into:
by: Ark94
5 Sep 2010, 04:02 GMT+02
However, Google does not run JavaScript and after stripping the HTML tags they get only:
by: Ark94 1283652141|%e %b %Y, %H:%M %Z|agohover
Agree, this is ugly. Let us think if there is any room for improvements.
Michał Frąckowiak @ Wikidot Inc.
Visit my blog at michalf.me
Thank you so much, Michael. To be fair, it is a pretty bad example of a snippet for Google to run. They are supposed to be better at generating snippets than that! So it makes no sense in the first place why such a thing would happen. As far as I know, they use one of three methods: meta data attached to the page, snippets taken from the text with the search words highlighted, or the DMOZ data (which is silly IMO). However, if there is any solution it would be great. This happens quite often for some reason with forum threads. In addition to the example that you described above:
by: Ark94 1283652141|%e %b %Y, %H:%M %Z|agohover
They also pick up the "unfold" and "fold" part so that you might get something like:
Unfold Re: Some Subject y: Ark94 1283652141|%e %b %Y, %H:%M %Z|agohover. Fold
For the HTML codes, I didn't not explain myself well. It was two different problems I was speaking of. One being the problem you explained above and a completely separate one being google picking up actual bits of html that are used on the pages. BUT this only happens with the old method of using a code block and iframe to render html on a page, not with the new HTML blocks..as far as I can tell..so it's not really anything that needs to be solved AND I have not seen this behavior happen for quite a while.
Thanks again.
Edit: If you want to easily see an example of this in action you can go to my site (http://www.gustrength.com) and use the google search in the top bar to search "oxygenated water". The exact same behavior happens in the custom search. It's not just the oxygenated water thread, of course, that is just the present example.
One thing we did this week was rewriting data encoding, so that the HTML code is much more cleaner. It was the easiest one to improve and should bring visible improvement. Now the dates in HTML look like this:
Which will be visible as simply 05 Sep 2010 02:02 to all bots and spiders that strip HTML tags.
What is also important, the dates are now visible to web browsers with disabled JavaScript, although they will be UTC (GMT) dates instead of local time, but still much better than blank date.
Hope this helps!
Michał Frąckowiak @ Wikidot Inc.
Visit my blog at michalf.me
That sounds great! Thanks, Michal. I'll let you know when I see the results of it.
Hey, Michal, forgot to let you know about the results of this. Now, from what I can see, the search results for forum threads are perfect. Here are some recent examples:
What do you bring to the gym? - Ground Up Strength
www.gustrength.com › Forum › Other / General Discussion
Mar 8, 2011 – Joe's post out here has inspired this thread.. So what do you guys carry with y'all to the gym? Here's my stash: Digital Camera; Chalk; Shoes …
Facial Muscles Video: Occipitofrontalis, Temporalis, Masseter …
www.gustrength.com/…/facial-muscles-video:occipitofrontalis-tempo…
Nov 12, 2010 – Responsible for the expression of "surprise". Temporalis … Orbicularis Oris Muscle*, Maxilla and mandible, Skin around the lips, forms the mouth in a circle to make the letter "O". Or the letter … A "frowning" or "pouting muscle".
As you can see, the date shows up perfectly. And what's more, google does not seem to be picking up the "fold" and "unfold" or any of that any more, but generating normal text snippets based on the query.