Jump to content

Proposal talk:Offer Google translation services with cursor hover over word in webpage

From Strategic Planning
Latest comment: 15 years ago by Endo999 in topic Apertium is a better option

I think that this is a great idea for an opt-in option. I wouldn't want it on by default though. Gopher65talk 01:32, 23 August 2009 (UTC)Reply

Agreed. And it might impede the amount of people wanting to create articles in other languages. --Danlev 21:32, 22 September 2009 (UTC)Reply

I'm not sure we could do any better than browser integration of such services, e.g., gTranslate for Firefox. 99.60.1.164 02:22, 23 August 2009 (UTC)Reply

This isn't a strategy, it's a feature request. --Bodnotbod 14:38, 10 September 2009 (UTC)Reply
Then let's make it more strategic by defining some priorities:
  • translation tools such as Google's can assist multilingual editors in enhancing content pages (e.g. Wikipedia articles) by providing pre-translations from more complete and better referenced articles in other languages;
  • the English-language Wikipedia with its 3 million or more articles could benefit from this feature by providing a basis for expansion of existing articles or stubs which have more information in one or more of the non-English equivalents;
  • a statistical approach could be implemented on the basis of the quality and priority criteria listed on the talk pages in conjunction with the number of accesses to the article page;
  • a similar approach could be used to enhance non-English articles on the basis of well-researched, frequently accessed English-language articles.Ipigott 19:35, 1 October 2009 (UTC)Reply

Impact?

Some proposals will have massive impact on end-users, including non-editors. Some will have minimal impact. What will be the impact of this proposal on our end-users? -- Philippe 00:14, 3 September 2009 (UTC)Reply



Not open source

There is a serious problem with Google: It is a commercial company, and their MT software is their, and nothing for someone else to tamper with. We tried to used a non-open MT solution for Norwegian Nynorsk Wikipedia, but had to abandon that, due to a conflict of interest. My suggestion would be to go for some open-source MT. Trondtr 06:48, 16 September 2009 (UTC)Reply

Agree. I have a conflict of interest in that I work on free/open-source machine translation, but I think relying on a commercial vendor for any feature in Wikipedia would be a bad idea. Consider how many Wikipedians would feel if it were Microsoft in place of Google. - Francis Tyers 08:44, 16 September 2009 (UTC)Reply
It is either open source or it is not. When the source code is provided, when the license is a free one it IS open source. The fact that the code is provided / maintained by a commercial company is and should be irrelevant. GerardM 09:41, 16 September 2009 (UTC)Reply
Actually I completely agree here, if the source code and data are free software/open source then there is always the possibility to change vendor. That is not the case with Google Translate. - Francis Tyers 07:37, 17 September 2009 (UTC)Reply

GoogleTrans, the software proposed, currently uses Google translation services. This is just a software call. The javascript file can easily be changed to call any webbased translation service. The call to a translation service is a very small part of the code for the gadget in question and can be changed easily.

Google does not offer onerous conditions on its translation services. They just have to be free to the end user. They are doing something great here, and ideological considerations about the superiority of OPEN/SOURCE should not hinder us from using a very powerful and free to use service.

Endo999 18:59, 30 September 2009 (UTC)Reply

Madness. Open source is at the very heart of Wikimedia. -- Philippe 19:05, 30 September 2009 (UTC)Reply

Apertium is a better option

See the Apertium Wiki.

Apertium is a free/open-source machine translation platform. --ƒajro @ 19:47, 13 November 2009 (UTC)Reply

In order to have a popup with a translation appear in a wikipedia webpage, you need to

1) do a AJAX web lookup of the word at the Wikipedia site. This cannot be done for most translation packages, because they are not sited at Wikipedia, but on other URLs.
2) invoke the AJAX web lookup in a way that gets around the source url problem spoken of in 1). Google gets around this with an API that can make a call to Google from any website. I believe they hack the RSS feed to do this. If the translation service can perform this task, then the javascript spoken of (the gadget on the enwiki) can be quickly changed to make this call and get the translation from the Open Source service.

Currently, Google is the only service I am aware of that gets around this AJAX limitation in the web browsers. Any translation service that does, however, can be called by the gadget in question.

Endo999 02:58, 22 November 2009 (UTC)Reply