WO2013163964A1

WO2013163964A1 - Website translation delivery and manipulation

Info

Publication number: WO2013163964A1
Application number: PCT/CY2012/000002
Authority: WO
Inventors: Daniel Andersson
Original assignee: Xcremeno Ltd
Priority date: 2012-05-02
Filing date: 2012-05-02
Publication date: 2013-11-07

Abstract

A method of delivering and manipulating content on a web page in near real time by a third party service that is compatible with any web page to deliver a translation. The web page may be displayed in the original language initially until the translation is available.

Description

WEBSITE TRANSLATION DELIVERY AN D MAN I PU LATI ON

Background of the invention [00] The present invention relates to website translation delivery. More

[0001] 1. Field of the Invention particularly it relates to how translations are stored, accessed and delivered from a third party while a user is accessing a web each part of a website, is the lack of page online. scalability and the increased complexity to produce a new function for a website in a

[00] 2. Description of Related Art new language. Because there's no standard for translations or content management

[00] Today there exist many types of

there can't be a system that fits the translation systems to store translation.

backend of every webpage due the custom These are custom systems tied into

nature of website functions and features. generally pre-defined content management

system that is operating on the host [00] The only current standardized computer of a website. These content webpage translation tools comes part of a management systems generally has prepackage deal for website backend systems, defined capabilities such as news content, where you have a preset amount of articles, shopping items tied in with text. features such as new and articles where the The translation system is then applied to translation system will support these each capability and is operating to translate functions and no other, unless extended these items. and customized by the end user creating new system functionality, if allowed by the

[00] Translation systems for web pages

license of the system used.

today allows for a website to be available in

multiple languages by having multiple Summary of the invention

copies of content available tied together in

a backend system. By example, a site might [00] The invention relates to an external have a database full of news articles, translation mechanism that will be instead of having it stored as an index such positioned as a third party service and host as News ID & Text, it's stored as a News ID translations on a separate host. Whenever a & English Text, German Text. There's then a visitor requests a webpage there will be a separate function that stores other parts of second request to an external server that the website, such as menu, about, footer will retrieve translations for a webpage. information, banners, etc in the same There will also be a system in-between the manner. When a website is served it calls visitor and the original host or a system up on these different custom functions to executed locally on the visitor's browser assemble a fully functional website in the that will manipulate the webpage code to language a user requested. insert any translated text in place of the original text. The end result will be a full

[00] By design current translation systems translation for any webpage independent of are always tied in to content, that will say to the underlying system or programming create a new function, such as a website language used to generate the web page evolving to create a news feed for it's HTML code by the original host.

visitors, it's then always required to also

create a custom function that can tie this [00] Because this system is offered as a news content into a translation system. third party service, and is executed after the Increasing the complexity of the backend fact of the website host and independent of system powering a website exponentially. the website host programming language and system functionalities it will work with

[00] The problem found in current systems, any website. It is capable of extracting the where the translation is tied together with text out of the web page the visitor requests, find any translation for said text, the visitor requested the web page from and then replace the text in the web page the original host, but fully translated. code, and forward it to the user in near real

time. [00] The second method, illustrated in Fig. 2 involves an external script being included

Brief description of the drawings ΐβη the webpage code. When a webpage is loaded, it will be sent to the visitor from the

[00] original host, and include an extra line of external code that will be executed in the

Detailed description of the invention

browser of the visitor. When the webpage

[00] The invented method of website is requested, sent and displayed to the translation stores every sentence and its visitor in its original language, this code will designated translation in a third party be executed on the visitor's browser to system that will offer near real time parse and find all text elements, which will translations as a service. The third party be sent to the third party service provider, system could simultaneously keep which will find all translations for these translations for millions of web pages of sentences, and send it back. The software thousands of different unique websites and code would then switch out text on the offer each sentence up as requested. webpage with the translation. What this second method provides is an easier

[00] By example, there are currently two integration for the website owner, in return different methods of delivering these the webpage will be displayed briefly, translations to a visitor. The first method , usually for less than 1 or 2 seconds, to the illustrated in Fig. 1 places an http proxy in- user in its original language and then the between the visitor and the website host. text will be switched out with the translated The website owner will have to arrange for counterpart in front of the eyes of the two things in order to setup the http proxy visitor on every web page.

for translation. A specific domain name,

such as de.example.com for German would [00] The background parsing element for a be forwarded via it's DNS settings to the website works in essentially the same way translation service providers IP, and the for both the first and second method. The owner would then specify where the source HTML code of a website, which is the website is. Which in this example would be standard of any website, can be broken an English website located on down to separate each text element from www.example.com. In effect, what happens any other part of the HTML code of a when a visitor visits webpage. Once this is done there's a list of de.example.com/example-pathl.htmljs sentences that is tied back to its position in that the visitor requests a webpage from the page. These sentences as an array is the translation service provider, the system sent to a translation delivery system would in turn contact the source at combined with the unique location on the www.example.com/example-pathl.html, web - by example

download the full web page, parse the code de.example.com/example-path.html. Once to split and find every translation. It would the translation storage server receives both then put the webpage together again with these data points, it can retrieve the the German text replacing the English text, translations for this path by the sentences it and then forwarded to the visitor just as if received and reply with a new array that contains. When the parser receives this and request an automatic translation from translation array, which maintains the any available service, such as Google pointers to where in the original HTML code Translate. Once the automatic translation is each sentence is located, can then go thru completed it can be mixed together with every sentence, remove the original the original translation and the request can sentence and put in the translated sentence be completed and sent to the parser and in in its place. The result would be the exact turn to the visitor. After this has been sent same web page but with every sentence the automatic translated text can be put in translated. to the translation storage database with the explicit tagged as automatic, allowing them

[00] The storage server of translations to be subsequently served for any future would maintain an index, which every single requests while also being tagged as webpage that would have a sub index of temporary and scheduled to be translated every sentence of that webpage. For by a real human translator.

immediate access, this full database will be

stored in RAM memory of the host server, [00] The third option can display an which will maintain a retrieval time of information box for the user that the milliseconds per sentence in the range of 5- current page is not yet translated in full, 50 milliseconds for the sentences of a full who can then choose to view the webpage webpage when accessed in parallel, to in it's original language, or as the first successfully match the speed expectations option, or as the second option.

by a web page visitor. When requested, it

can match the sentences sent with the

sentences stored and if every sentences

requested exists as a translation it can send

and complete the request. In case the

storage server finds that there are new

sentences sent by the parser which is not

stored, there are three different outcomes

that can be pre-defined by the website

owner on a per-website scope. What would

be constant every time a sentence is missed

is that it will be logged and alerted to a

translation index server which can then in

turn alert and request a translation for this

sentence, in effect automating detection of

any changes on a web page that can be

automatically sent out to be translated.

[00] The first option would send the full

website to the visitor which would translate

as many words as available, and then leave

the words not yet translated.

[00] The second option would create an

index of every sentence not yet translated

Claims

1. This invention is a unique way to provide for website translation delivery and manipulation.

2. It provides a solution for the lack of scalability found in current website translation methods.

3. Tied to the claim number 1 is that this invention is related to an external translation mechanism that will be positioned as a third party service and host translations on a separate host.

4. There will be a connection between the visitor and the original host that will manipulate the website code to insert any translated text in place of the original text.

5. This method is unique and independent. This is because it is offered as a third party service and will be executed after the website host and it will be independent from the website host programming language as well as any system functionalities. This means that it will work with any website.

6. In addition another way is by including an external script in the webpage code. When webpage is loaded, it will be sent to the visitor from the original host, and include an extra line of external code that will be executed in the browser of the visitor

7. It will extract the text out of the web page the visitor requests, find a translation for the said text and then replace the text in the webpage code, and forward it to the user in near real time.

8. The storage server of translations would maintain an index, with every single webpage and in turn a sub index if every sentence of that webpage. For immediate access, this full database will be stored in RAM memory of the host server, which will have a retrieval time of milliseconds per sentence.