5 replies to this topic

[kology]

I'm in the mist of an SEO of a very large corporate site.

The site is built using WebObjects and has nearly a thousand pages. To say the site has not been optimized is an understatement, but here is the wired problem I'm encountering.

Virtually every page has been indexed by Google (not so with Yahoo and others, but I know why for that). However when I do a site:www.domain.com, to see all the pages, there are no page titles present just the URL is displayed.

I understand why there are no descriptions showing each page is missing the description meta tag and most pages are 100% graphics. Yet each page does have an unique title tag. The title appears when you few the physical page, it's there in the code, but not in Google.

I suspect the problem is caused by the site being very heavily parameter driven.

Most pages in the Google listing have urls similar to this:
www.domain.com/cgi-bin/WebObjects/WWW.woa/12/wo/Home.TMMC-FsPRI2lKWTFV1hy4RBb3AM/0.9?t160000e.html

Yet at the same time I get only the URL for a page listed as follows:
www.domain.com/gaspe/53162F.html

Any ideas on what's causing this?

I can fix the page design so there are words to be indexed, but if Google is currently ignoring the title tag, what's there to say it won't ignore other tags on the site and content as well.

All help and comments welcomed.

[Randy]

Alan this is all going to be kind of nebulous, as I've only worked with WebObjects once in my life and that was an IntrAnet application. But it may at least give you a starting point to work from. Hopefully my memory isn't too bad and someone who has more recent information will chime in.

Back in the day when I did the Corporate Consulting thing this one company had a major issue of new computers they had installed in their office being completely unable to call up any of their Intranet site. What was happening then was that they clicked on a link that was built into each of their internal systems (that worked via command line) to connect with different areas of their internal site.

But a browser was never launched on these new computers as it should have been via WebObjects.

The strange part initially was that if you launched a web browser and went to the internal URL, the page loaded just fine. And you could continue browsing as normal.

What was happening in that case was that WebObjects, which is CGI Adaptor, was unable to launch the browser because it was unable to detect the browser that was installed on these newer systems. It had not been upgraded, so it was looking for an older version of IE (3 I think?) and the new computers all had IE5 installed.

So it didn't work. Unless you had pre-fired your IE5 browser and went to the page with that.

Viewing the sample URL you put above, it looks to me like maybe the same sort of thing is happening. Since the Googlebot "browser" that isn't really a brower and is unrecognized is trying to connect, you're getting a command line entry instead the real URL. Only because it's a non-recognized browser as far as WebObjects is concerned.

There are about 4,000 assumptions and extrapolations based upon a fading memory in the above, so I may be totally off base. But it may also start you down the path to a solution. We ended up having to change something --sorry I don't remember what at this point-- something in the WebObjects config files in order to get it to recognize the newer machines.

[kology]

Thanks Randy,

That makes sense, I'll test your theory with a fake browser setting.

I too suffer from a fading memory. I guess that's why were suppose to write documentation and comment our code. Oh I knew I forgot something.

[Randy]

[Raphael]

Real programmers never comment their code.

If it was hard to write, it should be hard to read, dammit!

disclaimer: the above comment is intended for humor purposes only. Please have pity on the next poor sap that's supposed to modify the code and comment it

[kology]

I just noticed that I didn't post a follow-up message.

I did test the theory and yes, when the site was confronted with a browser it didn't know, the pages didn't display properly.