Support for multiple languages

It looks like the mod was moved to the graveyard...

One important thing that Dave fails to realize is that many of the sites that are using that wordpress plugin are small. How many blog posts are being translated vs thread pages? Probably a lot more thread pages, especially on a large forum. That wordpress plugin used 3 other tranlation services other than google which would make it even more unlikely that google would ban you for using that version.

Not to mention a google API key wasn't properly being used - someone else introduced a fix for that, but was never put into the actual mod.

All this equals more negative attention in google's eyes - not worth risking lowering in rankings and getting banned.

Dave - I think you need to lighten up and stop taking criticism so personally. I know this is something you made, but it clearly has issues.

Originally Posted by Brian Cummiskey

We have discussed this mod internally. At this point in time, we do not plan to support this mod. Further, we suggest not using this mod at all. We applaud the efforts and goal of the product, however, there are too many limitations and SEO consequences that led us to this position.

• The mod uses the parameter hl. hl is also used in search strings as "highlight". As you know, vbseo cookies this data parameter to achieve the highlight effect without keeping the parameter in the url. While it is highly unlikely for someone to search for 'en' or 'de' (if any results even come through being 2-chars), it will require some internal modifications to our base code.
• The value of the translation is minimal at best. The AI translations simply aren't 'there' yet. If you've ever translated a page manually, you will know that it is hard to understand-- if it makes sense at all. Further, should a foreign language member signup and post in his native tongue, you then are faced with your root language being translated as well on those posts.
• There are also charset issues to address with different languages. vb does not ship in utf-8 form, and neither is mysql set up for it out of the box (on most English systems). This may cause database errors.
• If you move to /xx/thread/ where xx is the lang code, you should also do the same for your base language (say, en). Thus, every link you currently have (assuming an English forum) without the language code should be moved to the lang directory as well, and thus, every page on your site will need to be re-indexed. We do not recommend changing url structure once you are indexed, and likewise, we do not recommend you add a lang code to your root language.
• Finally, as Mert points out above, there is the non-link consensus urls as a byproduct. This is one of the main goals that vbseo achieves and within lies one of its biggest values.

I read Brian's "critique" and, quite frankly, am dissapointed in the misinformation provided. So, to be fair, let me clarify:

It is certainly within the right of Brian and the vBSEO team to not "support" the mod; however, the post is not accurate and biased, in my opinion. I am not going to address the first bullet point, because that is trivial to fix.

However, the second point is completely an uninformed opinion Brian. I have a friend who is Japanese and has studied the Japanese translations by this mod. She says the translations are not easy to understand (she actually corrected many by hand and gets a good laugh out of any of them), but she says they are are very good for Japanese people because they can get find info they would otherwise not be able to find. While not perfect, it is certainly better than nothing, she says (and she is fluent in both Japanese and English) and is a valuable resource. For Brian to opine "the value of the translation is minimal at best" is to basically marginalize all machine language translation service. Everyone I know strongly disagrees with Brian's opinion, especially folks who are searching for information (which is the purpose of SEO!!)

The charset issue (UTF 8) mentioned is also misleading. vB supports multiple languages, so that is a red herring argument. This is simple to fix, if required.

Needless to say, the final two points are also trivial, technically speaking. The point about "people will post in their native language" is not our experience - especially if you have forum rules otherwise.

First hand experience, this mod is great for SEO, opens up "the world" for people, given the limitations of ML translations, and greatly increases search referrals. The proof is not in "theory" or "opinions" but in practice, and practice had shown that the mod under discussion:

• Provides people of other languages a means to locate information they may otherwise not receive. This is important. In fact, this is one of the goals of search and ML translations.
  
• Generates significant search engine referrals for sites and in turn,
  
• Increases Adsense revenue for Google, when sites are running Adsense, because more traffic to Adsense publishers means more revenue for Google. This is excellent for Google and the TOS is not meant to harm Google revenue (especially when they understand the positive impact on their revenue streams).

The bottom line is this:

If more people used a mod like this, the world would have more access to information. Even if the translation is not perfect, if relevant, then you can translate further.

The mod is great for Google revenue when used on sites that run Adsense (which is the vast majority of people). Google did not write their "no long term storage" TOS for purposes to hurt their own revenue and the profits of their stockholders.

Cheers and Best Regards.

Originally Posted by reborg

I read Brian's "critique" and, quite frankly, am dissapointed in the misinformation provided. So, to be fair, let me clarify:
...

What about the points I made about it violating google's TOS in 3 ways:

• Doesn't use google API key (in 'stock' form)
• Stores data in a database for more than 15 days
• Doesn't credit google as the originator of the translated data

I'm not arguing the fact that it is a very useful modification, but it's just in bad faith to go against google's TOS.

As pointed out above, storing the cache translation beyond 15 days is against Google's TOS. For that reason alone we will not support this mod nor recommend any of our customers to use it. Getting on Google's bad side could have long term detrimental effects to one's community. Ask anyone who's been in the sand box before.

I'm glad you have had good luck with it. It may be short lived. It may not. Regardless of the outcome, it is basically breaking the law of Google... and he's one bad man you DON'T want to mess with.

Originally Posted by reborg

However, the second point is completely an uninformed opinion Brian. I have a friend who is Japanese and has studied the Japanese translations by this mod. She says the translations are not easy to understand (she actually corrected many by hand and get a good laugh out of them), but she says they are are very good for Japanese people because they can get find info they would otherwise not be able to find. While not perfect, it is certainly better than nothing, she says (and she is fluent in both Japanese and English) and is a valuable resource. For Brian to opine "the value of the translation is minimal at best" is to basically marginalize all machine language translation service. Everyone I know strongly disagrees with Brian's opinion, especially folks who are searching for information (which is the purpose of SEO!!)

Turkish is my main language and when i check translations made by Google API , the only thing that the translated page gives me a huge smilie in my face. I tried with many other languages and it is always the same problem. You may find Google translation good but unfortunately it is not.

Originally Posted by reborg

The charset issue (UTF 8) mentioned is also misleading. vB supports multiple languages, so that is a red herring argument. This is simple to fix, if required.

Actually again you are not correct. If you want to run multiple language translations on your site , your charset should be defined as UTF8 . Any other charset will cause problems. Yes vBulletin has an option which allows you to define different languages but each language has an option to define charset that will display correct characters. But with this mod you'll not have that option.

Originally Posted by reborg

First hand experience, this mod is great for SEO, opens up "the world" for people, given the limitations of ML translations, and greatly increases search referrals. The proof is not in "theory" or "opinions" but in practice, and practice had shown that the mod under discussion:

• Provides people of other languages a means to locate information they may otherwise not receive. This is important. In fact, this is one of the goals of search and ML translations.
  
• Generates significant search engine referrals for sites and in turn,
  
• Increases Adsense revenue for Google, when sites are running Adsense, because more traffic to Adsense publishers means more revenue for Google. This is excellent for Google and the TOS is not meant to harm Google revenue (especially when they understand the positive impact on their revenue streams).

The bottom line is this:

If more people used a mod like this, the world would have more access to information. Even if the translation is not perfect, if relevant, then you can translate further.

The mod is great for Google revenue when used on sites that run Adsense (which is the vast majority of people). Google did not write their "no l

Actually , as i stated before , the idea of this mod is great however with the current structure , the mod makes more damage than offering better SEO. If you want to use translations you should make a correct setup for your site structure. Each language should be defined correctly and each translated page should have a unique id. Without this you may face a quick impact on your SE referrals but in long term you'll be open to face penalties. Because this mod breaks vBSEO's link consensus.

I have one question left for you.

Please explain me how you plan to explain the differences to SE's for following links and which link should SE's rank higher.

domain.com/thread1.html
domain.com/thread1.html?hl=en
domain.com/thread1.html?hl=tr
domain.com/thread1.html?hl=de
domain.com/thread1.html?hl=ru
domain.com/thread1.html?hl=ar

As it is also stated several times this mod is also violating Google Translate API's TOS.

Originally Posted by Mert Gökçeimam

As it is also stated several times this mod is also violating Google Translate API's TOS.

Actually, this is the bottom line perspective of vBSEO, not the technical issues.

If vBSEO believed the mod did not violate Google Translate API's TOS, then vBSEO would support the mod. The remainder of the discussion against the mod is a red herring, and is more a critique of machine language translation in general, which is off topic (and debatable outside the context of this discussion).

I have reviewed the API TOS many times, and I do not believe that using the translation API to drive more traffic to a site running Google Adsense is a violation of the intent of the TOS. Google wants to make money (actually it is their sole purpose as a corporate entity). Anyone who uses their API to make money for Google is a friend of Google.

I do agree that if you used the translations to drive traffic to sites that were not running Adsense (like a Google competitor), then you might make Google unhappy.

However, I am quite sure that if you are a Google Adsense customer and you are using the their translation API to increase Google's revenue, then you will be on Google's "good side" and certainly not their bad side.

Basically, the TOS was not intended to hurt Google's revenue. This is business, not theory or word-smithing. Google is in the business to make money. Not supporting the mod causes Google to lose money if the site is an Adsense publisher. That is an undisputed fact.

I think Google is happy to make money, in practice, and there is no practical reason for Google to be unhappy with anyone making information available to international users when it increases Google's revenue very significantly.

As I said, you guys at vBSEO are great guys. You have a great product. You are certainly entitled to your opinion. However, in reality, the issues are not nearly as "black and white" as you are presenting them.

You should support the mod, in my opinion; and you should also help forward the art-and-science of machine language translation (and not simply toss rocks at an evolving domain).

Cheers.

I have typed all I have to say and have nothing more to add to the discussion at this point. For some odd reason, this mod evokes too much human emotions with a number of people.

Originally Posted by reborg

If vBSEO believed the mod did not violate Google Translate API's TOS, then vBSEO would support the mod. The remainder of the discussion against the mod is a red herring, and is more a critique of machine language translation in general, which is off topic (and debatable outside the context of this discussion).

vbSEO said they would not support the mod way before I pointed out that it violated the tos - read the beginning of this thread

Originally Posted by reborg

I have reviewed the API TOS many times, and I do not believe that using the translation API to drive more traffic to a site running Google Adsense is a violation of the intent of the TOS. Google wants to make money (actually it is their sole purpose as a corporate entity). Anyone who uses their API to make money for Google is a friend of Google.

You were reviewing the wrong TOS... Support for multiple languages

Originally Posted by reborg

If vBSEO believed the mod did not violate Google Translate API's TOS, then vBSEO would support the mod.

No.

If you scroll up, you will notice that I made my annoucement BEFORE tavenger brought up the TOS issues that we didn't even know about before deciding on this stance. The TOS simply iced the cake.

We decided not to support it because it violates link consensus, and that is the foundation of the vbseo product.

Further, the product has been pulled from vb.org. There is 0 sense to supporting a dead product.

I think we've had enough discussion on this, and I will kindly close this thread down now.