Google PDF Anomalies

[T0PS3O]

As reported by Leo...

Originally Posted by leo

Since yesterday, doing the site:www.domain.xx - command for my site gives back only htm/html-pages, not a single pdf. What does that mean? Are pdfs no longer considered valid by G! or what?

... PDF SERPs have gone tits up.

Try search for 'instructions' (one would expect a million instruction manuals) and set it to PDF only.

http://www.google.com/search?num=10&...G=Search&meta=

Notice just 3 results but Results 1 - 3 of about 4,640,000 for instructions filetype: pdf.

Now if yuo mess around in the address bar, changing the results num ber (num) to 30, 50, 90 etc. you will get a few more. But still under 10 results....

The Gooooooogle NEXT >> buttons are gone too...

Very strange... Or is it just Leo, Jan and me seeing this?

[minstrel]

What happened to the rest of this thread? I'm seeing PDF in Google searches... I posted about 4 unselected links to samples... where'd they go?

<edit>okay... never mind... duplicate post by TOPS... see http://forums.digitalpoint.com/showt...8281#post48281

[Jan]

Well, nobody said ALL pdf's are gone. Just a lot of them, maybe most of them. For more than a day now.

E.g., a search on adobe filetype: pdf gives only 7 (!) results.

[T0PS3O]

I thought it deserved its own thread since it was off-topic in the other one. And it isn't so much ducplicate since I had a closer look and posted more details.

It is very strange indeed. Who wants to try and e-mail G?

Minstrel, your post is here http://forums.digitalpoint.com/showp...1&postcount=27 and you proof exactly our point.

See in your own links how just 3 to 7 results show up. No next page, no more results. But it does say 1-3 out of 4,000,000,whatever results.

[minstrel]

In my post in the other thread, those examples weren't an exhaustive list -- they were simply about 5 minutes work with a few test phrases, all of which produced some PDF files in the SERPs. That doesn't look unusual for me -- in the kinds of searches I do, I don't usually get a ton of PDFs returned. That was my point: it looks normal to me.

[T0PS3O]

But those pages I showed, and even those you linked to show exactly that it is far from normal.

Why isn't there the next page link at the bottom?

Why does it show 3 results only out of 4 million?

If that looks normal to you (OK G has been behaving strange a bit) I'd recommend you get your eyes checked.

[minstrel]

Thanks -- I did recently get my eyes checked. But I don't think it's necessary to be insulting simply because I'm not seeing the problem you seem to see.

No SERP contains a "next" link unless there are more than 10 (or whatever you have set as your default for the search) results to list.

Did you see this sample search from my other post? http://www.google.com/search?hl=en&l...filetype%3Apdf -- that one has a whole lot of "next" pages.

And that is, as I said earlier, just 4 quick searches, not exhaustive research. I'm sure it would not be difficult to find other examples if I wanted to invest more than a few minutes to do it.

I don't think this proves your point at all -- I think it demonstrates the opposite, actually. As for the "But it does say 1-3 out of 4,000,000,whatever results" part, isn't that just telling you there are 4,000,000 results of which 3 are PDF files? That number may be a little low for some searches (it would be for my example above) but it would be entirely normal for others.

In my experience, the majority of PDF files on the net contain larger manuals or research papers, etc., that don't easily fit into a manageable HTML page, or documents that require special formatting, such as an application form. I would not therefore expect the SERPs to contain very many of them in most cases.

That's what I meant by "it looks normal to me"...

[T0PS3O]

I didn't mean to be insulting at all. I appologize if it came across like that.

I was just surprised to see that you thought all was fine whilst the 2 links I clicked from your examples just confirmed my findings.

The ones limited to PDF files only seem to have this error in common.

[minstrel]

No, the example above is also a PDF-limited search: http://www.google.com/search?hl=en&l...filetype%3Apdf using the search term "psychopathy filetype: pdf" (minus the quote marks, space added after the ":" because the forum conderted ":" + "p" to a smiley ).

[Jan]

Let's try to summarize:

* Two persons in this forum (me and Leo) have reported losing all their pdf files from the G index;
* A number of the highest ranking pdf files on the web (from Adobe) also seem to have disappeared from the G index;
* Searches for "filetype: pdf" (without the space) or just "pdf" tend to return an unusually limited number of .pdf hits;
* For some pdf files that are still found, the toolbar shows they are not ranked or indexed;
* Some people (Minstrel c.s.) do not notice anything unusual;
* Just 4 persons have discussed this topic in the last 36 hours; this could be because others - like Minstrel - do not notice anything unusual, or they may not be very interested or experienced because they don't use pdf files; or whatever.

It seems to me a lot (or even most) pdf files have suddenly disappeared from the G index. What further tests could be done to clarify the situation?

----------

Added example

Google help page links to a pdf file. The pdf file is called The Anatomy of a Search Engine and is written by Sergey Brin and Lawrence Page. It is linked from a PR 10 page but not in the G index.

Help page:

http://www.google.com/help/features.html

Under File Types we find the referenced article: The Anatomy of a Search Engine. View as HTML option does not work.

URL of the pdf:

http://www-db.stanford.edu/pub/papers/google.pdf

Toolbar shows article is not indexed and pdf is not found in the G index...

[Jan]

For search complaints see:

http://groups.google.com/groups?dq=&...upport.general

It seems the search problem was first noticed on Monday 25 Oct.

[minstrel]

How odd. Especially since, even in the posts you reference (as well as at least one of the examples I posted), it doesn't seem to affect ALL pdf searches.

[leo]

Originally Posted by Jan

Let's try to summarize:
...
It seems to me a lot (or even most) pdf files have suddenly disappeared from the G index. What further tests could be done to clarify the situation?
...

I wouldn't say they have disappeared. I had a BL update recently (but AFTER having sent my very first mail regarding this matter) and the total of pages still indicated - that means, both normal html-pages as well as all my pdf-pages have been indexed by G!.

Still, not a single one of these pdf-pages is shown by the site:xxx-command, and if I manipulate the settings of G! to show pdf ONLY I get ZERO returns after having typed the same command. Rather strange indeed, isn't it.

Otherwise I completely agree with what was said by Jan.

[Jan]

I wouldn't say they have disappeared. I had a BL update recently (but AFTER having sent my very first mail regarding this matter) and the total of pages still indicated - that means, both normal html-pages as well as all my pdf-pages have been indexed by G!.

Hi Leo,

Could you elaborate a bit?

Most of us had a BL update very recently. I had another one a few days earlier which may have been caused by the loss of pdf files, as I reported then. But since I first noticed the loss of pdf files, nothing seems to have changed, and I do not have any indication that they may still be there somehow.

So, could you explain your observation in more detail? What search did you do, what result did you get, and what conclusion do you draw from it? If your conclusion is based on the number of backlinks reported, I would be very sceptical about it.

[minstrel]

Again, Jan, I do agree this is all very curious, but I wonder if you are focusing on the wrong questions here: Maybe it's not a PDF filetype issue at all but something else which is manifesting itself in PDF searches?

Any attempts to figure out what is going on have to explain not only the absences you are seeing but the fact that searches like the one I listed above -- http://www.google.com/search?hl=en&l...filetype%3Apdf -- show

Results 1 - 10 of about 3,760 for psychopathy filetype: pdf

The "display as HTML" feature does work on those listings. But what is especially curious about the output for that URL is that results 1-10 are displayed, along with the usual navigation information for pages 1-10.

Result Page: 1 2 3 4 5 6 7 8 9 10 Next

However, when you click on either "Next page" or any of the number 1 to 10, you get a second page with only 4 listings and the notice

In order to show you the most relevant results, we have omitted some entries very similar to the 14 already displayed.
If you like, you can repeat the search with the omitted results included.

and the navigation links now say

Result Page: Previous 1 2

There are no "view as HTML" options on that second page, by the way. The other odd thing is that although page 1 lists results 1-10, you get that second abbreviated page whether you click on "Next page" or any of the numbers in the Google string, even though clicking on the digit 5 for example adds "&start=40" to the search string.

Clicking on repeat the search with the omitted results included generates the same results -- and the second page again is abbreviated as above -- except this time the search string has "&filter=0" appended.

My preliminary conclusion is not that PDF files aren't in the index, but rather that something odd is happening with how those results are displayed...

[leo]

Originally Posted by Jan

Could you elaborate a bit? ...

Sorry about the confusion I may have caused. What I meant is the following:

Two days or so after discovering that the site:xxx-command doesn't yield any of my pdfs I had a BL update as seen in the backlink table of the Keyword Tracker, which at the same time gives the total number of pages (i.e. html and pdfs altogether) - and there was no change in the total number of pages at the time of the BL update.

From this I conclude that G! still indexes all my pages, i.e. html and pdf. Does that explain what you are asking?

[Jan]

Minstrel, interesting indeed. Looks like a bug to me; its nature is open for speculation.

Leo, do you know what query your Keyword Tracker uses to report the total number of pages? Or where can it be found?

In theory it is possible that the missing pdf files are still indexed in G, and the software simply does not show them. But so far I did not find a way to make them visible, or to prove their presence. They simply don't show up in any searches as they used to. The "site:" query does not show them. The toolbar - a separate mechanism - says they are not indexed. The "View as HTML" option does not work for them. Until further evidence I feel that my earlier statement is quite justified:

It seems to me a lot (or even most) pdf files have suddenly disappeared from the G index.

Can anyone demonstrate the presence (in the G index) of the Google pdf I mentioned before? - http://www-db.stanford.edu/pub/papers/google.pdf

[leo]

Originally Posted by Jan

...
Leo, do you know what query your Keyword Tracker uses to report the total number of pages? Or where can it be found?
...

No idea. I am using the Digital Point Keyword Tracker, of course, so most probably one of the KT gurus will know. If necessary I could supply the settings I have chosen.

[leo]

Good news: this morning all PDFs are back again!

[T0PS3O]

Good, it took them a while though!