Thread: 404 Error with the Translation Script

It appears my explanation was bad, very bad. When I said "It gets re-cached only if someone -- or something -- accesses it again", what I meant was "It gets re-cached only if someone -- or something -- accesses it again -- and if the cache file is older than the value of $translator_cache_days." Does that make more sense?

The largest site I'm translating is around 3k pages. If you're translating 20k pages, your servers IP will get blocked and unblocked for months while the cache gets built. The only method to get around this is technically complex and requires a dedicated server.

The pages I had removed only got reindexed when I changed their URL's. I implemented a subdomain model instead of a subdirectory model. It's complex, but it's working. Traffic to those translated pages is only 1/3rd of what it was. I'm not sure if I should assign the cause to the loss of inbound links due to the deindexing or if I should start thinking that subdomains don't transfer domain authority as well at subdirectories.

What is the default of $translator_cache_days

So what you are saying is the translate ban/unban will effectively regulate the speed the pages are created and thus the speed Google picks them up. Makes sense so not to worry there I think.

I'm also pretty sure Goggle treats subdomains as separate, new domains - So yeah, some trust/juice is going to be lost.

Originally Posted by davesnake

What is the default of $translator_cache_days

It's 30, but I may change it to 90.

Originally Posted by davesnake

So what you are saying is the translate ban/unban will effectively regulate the speed the pages are created and thus the speed Google picks them up. Makes sense so not to worry there I think.

Well, in your case... worry is senseless. You have so many pages that your server's IP is certain to be banned. But, the bans auto-expire after a day or two.

Originally Posted by davesnake

I'm also pretty sure Goggle treats subdomains as separate, new domains - So yeah, some trust/juice is going to be lost.

And $200/day in ad revenue.

Ouch, 6k a month is a tidy sum.

Been checking the wordpress blogs running N2H's mod.

Seems some get de-indexed and some don't.

I guess if Google de-indexes the lang folder ill just rename it.

48hrs and not a single translated page indexed, which is odd, seeing a Google normally picks up new pages sometimes within hours.

Not much I can do other than wait but something doesn't feel right.

For your other new pages, does your CMS use a rapid indexing technique such as pinging?

I wouldn't worry for at least a couple of weeks.

Yahoo has already indexed loads and is sending traffic, Google not 1 page.

I know something is up, it just doesn't feel right. It is not indexing the pages.

Think I got it!

New vbseo adds rel="canonical" to page headers

On the translated pages the lang/country/ folder is stripped from the rel="canonical" URL.

So all the translated pages are effectively saying we are the UK version, so are being ignored.

If you look at the function GoogleFixup() in translate.php, you will see some examples of how that data can be manipulated.

Unfortunately I have another error.

The /lang/en/ that is 301'd to the regular language, the 301 is not correct, header checkers show /lang/en/ is being 301'd to www.domain.com/translate/g/en/